{
  "total": 532,
  "topics": [
    {
      "id": 327189,
      "title": "Welcome to Amex modeling challenge",
      "comment_count": 77,
      "views": 0,
      "votes": 105
    },
    {
      "id": 340114,
      "title": "Reminder: American Express is Hiring!",
      "comment_count": 30,
      "views": 0,
      "votes": 64
    },
    {
      "id": 348961,
      "title": "Leaderboard is Finalized - Congrats to our Winners, Recap",
      "comment_count": 10,
      "views": 0,
      "votes": 31
    },
    {
      "id": 328054,
      "title": "How To Reduce Data Size",
      "comment_count": 137,
      "views": 0,
      "votes": 521
    },
    {
      "id": 334670,
      "title": " DART algorithm explained",
      "comment_count": 17,
      "views": 0,
      "votes": 132
    },
    {
      "id": 347641,
      "title": "14th Place Gold – NN Transformer using LGBM Knowledge Distillation",
      "comment_count": 145,
      "views": 0,
      "votes": 272
    },
    {
      "id": 327534,
      "title": "Metric without DF",
      "comment_count": 4,
      "views": 0,
      "votes": 73
    },
    {
      "id": 349741,
      "title": "3rd solution--simple is the best",
      "comment_count": 23,
      "views": 0,
      "votes": 80
    },
    {
      "id": 471950,
      "title": "Data dicitonary for the american express- default prediction dataset",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 465760,
      "title": "Can we use the dataset for academic research purpose?",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 348111,
      "title": "1st solution(update github code)",
      "comment_count": 115,
      "views": 0,
      "votes": 303
    },
    {
      "id": 328514,
      "title": "Integer columns in the data - here you go!",
      "comment_count": 116,
      "views": 0,
      "votes": 834
    },
    {
      "id": 348014,
      "title": "13th Place Gold Solution",
      "comment_count": 20,
      "views": 0,
      "votes": 76
    },
    {
      "id": 327828,
      "title": "Kaggle Dataset for Transformers and RNNs",
      "comment_count": 28,
      "views": 0,
      "votes": 159
    },
    {
      "id": 327138,
      "title": "Parquet Format Dataset for Low Memory Use ",
      "comment_count": 70,
      "views": 0,
      "votes": 168
    },
    {
      "id": 328606,
      "title": "Speed Up XGB, CatBoost, and LGBM by 20x",
      "comment_count": 43,
      "views": 0,
      "votes": 211
    },
    {
      "id": 395565,
      "title": "Unique Customer IDs",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 347802,
      "title": "My first Kaggle competition: the learning is not over yet!",
      "comment_count": 4,
      "views": 0,
      "votes": 15
    },
    {
      "id": 335892,
      "title": "Tabular Classification - Tips and Tricks",
      "comment_count": 41,
      "views": 0,
      "votes": 358
    },
    {
      "id": 347637,
      "title": "2nd place solution - team JuneHomes (writeup)",
      "comment_count": 57,
      "views": 0,
      "votes": 180
    },
    {
      "id": 382586,
      "title": "It is possible to use those data for the development of Thesis research?",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 331565,
      "title": "LabelEncoder vs OrdinalEncoder?",
      "comment_count": 9,
      "views": 0,
      "votes": 16
    },
    {
      "id": 327132,
      "title": "Memethread :After seeing the dataset a MeME come to my mind.",
      "comment_count": 5,
      "views": 0,
      "votes": 20
    },
    {
      "id": 348789,
      "title": "I love Kaggle but decide to leave for a while (No More Kaggle, Goodbye)",
      "comment_count": 21,
      "views": 0,
      "votes": 36
    },
    {
      "id": 338635,
      "title": "GANs for tabular data?",
      "comment_count": 6,
      "views": 0,
      "votes": 5
    },
    {
      "id": 358845,
      "title": "Issues that could not be resolved",
      "comment_count": 0,
      "views": 0,
      "votes": 4
    },
    {
      "id": 350928,
      "title": "AmbrosM solution",
      "comment_count": 3,
      "views": 0,
      "votes": 7
    },
    {
      "id": 347786,
      "title": "11th Place Solution (LightGBM with meta features)",
      "comment_count": 44,
      "views": 0,
      "votes": 103
    },
    {
      "id": 327110,
      "title": "Handling large datasets with Dask",
      "comment_count": 3,
      "views": 0,
      "votes": 33
    },
    {
      "id": 357117,
      "title": "Table of Contents",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 348048,
      "title": "What competition will you do next?",
      "comment_count": 17,
      "views": 0,
      "votes": 5
    },
    {
      "id": 327333,
      "title": "How to reduce pandas memory while loading dataframe ?",
      "comment_count": 3,
      "views": 0,
      "votes": 17
    },
    {
      "id": 350471,
      "title": "American Express - Default Prediction | Understanding Top Solutions by CTDS",
      "comment_count": 8,
      "views": 0,
      "votes": 16
    },
    {
      "id": 333338,
      "title": "Understanding competition metric step by step",
      "comment_count": 26,
      "views": 0,
      "votes": 144
    },
    {
      "id": 327463,
      "title": "🔥🔥Marilia Prata Becomes Discussion GrandMaster",
      "comment_count": 196,
      "views": 0,
      "votes": 115
    },
    {
      "id": 350538,
      "title": "9th Place Solution ( XGBoost+LGBM+NN )",
      "comment_count": 8,
      "views": 0,
      "votes": 32
    },
    {
      "id": 349789,
      "title": "260th place solution",
      "comment_count": 2,
      "views": 0,
      "votes": 10
    },
    {
      "id": 347850,
      "title": "[Place 17th Solution]: Pseodo-label + FE.",
      "comment_count": 26,
      "views": 0,
      "votes": 48
    },
    {
      "id": 347688,
      "title": "Thank you all!",
      "comment_count": 34,
      "views": 0,
      "votes": 141
    },
    {
      "id": 331131,
      "title": "Which is the right feature importance?",
      "comment_count": 48,
      "views": 0,
      "votes": 210
    },
    {
      "id": 347996,
      "title": "135th place solution : +1483 shake-up",
      "comment_count": 6,
      "views": 0,
      "votes": 19
    },
    {
      "id": 347668,
      "title": "10th Place Solution: XGB with Autoregressive RNN features",
      "comment_count": 29,
      "views": 0,
      "votes": 93
    },
    {
      "id": 348058,
      "title": "12th Place Gold (2/2) - LGBM + XGBoost + Catboost",
      "comment_count": 13,
      "views": 0,
      "votes": 47
    },
    {
      "id": 347740,
      "title": "12th Place Gold (1/2) - lgbm+xgboost+FCN+Transformer",
      "comment_count": 20,
      "views": 0,
      "votes": 43
    },
    {
      "id": 348118,
      "title": "5th Place Solution - Team 💳VISA💳(Patrick's part)",
      "comment_count": 11,
      "views": 0,
      "votes": 51
    },
    {
      "id": 348097,
      "title": "5th Place Solution - Team 💳VISA💳(Summary&zakopuro's part)",
      "comment_count": 12,
      "views": 0,
      "votes": 61
    },
    {
      "id": 348304,
      "title": "How to choose between hundreds of machine learning algorithms?",
      "comment_count": 10,
      "views": 0,
      "votes": 2
    },
    {
      "id": 348530,
      "title": "21st Solution and Code Sharing",
      "comment_count": 18,
      "views": 0,
      "votes": 57
    },
    {
      "id": 347651,
      "title": "19th Place Solution",
      "comment_count": 19,
      "views": 0,
      "votes": 52
    },
    {
      "id": 348821,
      "title": "[Should be 12th solution before removal without any reasons]",
      "comment_count": 20,
      "views": 0,
      "votes": 6
    },
    {
      "id": 349250,
      "title": "18th Place Gold",
      "comment_count": 0,
      "views": 0,
      "votes": 18
    },
    {
      "id": 347966,
      "title": "45th place with XGBoost in first Kaggle competition",
      "comment_count": 4,
      "views": 0,
      "votes": 40
    },
    {
      "id": 348321,
      "title": "Remarkable Competition. Kaggle team, bring' on AMEX 2023!",
      "comment_count": 5,
      "views": 0,
      "votes": 23
    },
    {
      "id": 348764,
      "title": "A funny story ⛰️",
      "comment_count": 4,
      "views": 0,
      "votes": 14
    },
    {
      "id": 347880,
      "title": "25th place solution",
      "comment_count": 5,
      "views": 0,
      "votes": 24
    },
    {
      "id": 347763,
      "title": "(66th) My Reflections On This Competition",
      "comment_count": 2,
      "views": 0,
      "votes": 26
    },
    {
      "id": 347882,
      "title": "Silver (211 th) Long story short",
      "comment_count": 3,
      "views": 0,
      "votes": 20
    },
    {
      "id": 347863,
      "title": "72nd place solution(ensemble of LightGBM and Sequential NN)",
      "comment_count": 3,
      "views": 0,
      "votes": 20
    },
    {
      "id": 346975,
      "title": "I am addicted to this competition",
      "comment_count": 41,
      "views": 0,
      "votes": 34
    },
    {
      "id": 348366,
      "title": "Why ensemble model is more precise not the opposite?",
      "comment_count": 7,
      "views": 0,
      "votes": 6
    },
    {
      "id": 347540,
      "title": "Using high score notebook is risky?",
      "comment_count": 28,
      "views": 0,
      "votes": 6
    },
    {
      "id": 344695,
      "title": "Learning: Key learning or take away from this challenge.",
      "comment_count": 37,
      "views": 0,
      "votes": 22
    },
    {
      "id": 347700,
      "title": "Thank you competitors and host",
      "comment_count": 2,
      "views": 0,
      "votes": 13
    },
    {
      "id": 347709,
      "title": "My first solo silver",
      "comment_count": 3,
      "views": 0,
      "votes": 12
    },
    {
      "id": 348108,
      "title": "Remove?What happend",
      "comment_count": 12,
      "views": 0,
      "votes": 8
    },
    {
      "id": 348034,
      "title": "What is your strongest feature in this competition?",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 347643,
      "title": "Speculation time! Why did I jump from 402 to 49th?",
      "comment_count": 4,
      "views": 0,
      "votes": 8
    },
    {
      "id": 348306,
      "title": "American Express - My Learning Experience",
      "comment_count": 2,
      "views": 0,
      "votes": 6
    },
    {
      "id": 344545,
      "title": "A reminder for those of you who have used that 0.8 notebook just now.",
      "comment_count": 56,
      "views": 0,
      "votes": 49
    },
    {
      "id": 346775,
      "title": "Feature Selection: Random Bar Method⭐⭐⭐",
      "comment_count": 6,
      "views": 0,
      "votes": 17
    },
    {
      "id": 347805,
      "title": "Takeaways from the competition…",
      "comment_count": 4,
      "views": 0,
      "votes": 7
    },
    {
      "id": 347644,
      "title": "Thank you Raddar and Martin",
      "comment_count": 10,
      "views": 0,
      "votes": 108
    },
    {
      "id": 348093,
      "title": "Oh look there goes 4th place...Happy Kaggling is no more...",
      "comment_count": 9,
      "views": 0,
      "votes": 3
    },
    {
      "id": 347660,
      "title": "Share your shake up/down reasons",
      "comment_count": 27,
      "views": 0,
      "votes": 35
    },
    {
      "id": 348368,
      "title": "Key Takeaways: Understanding and categorizing the FE approaches of the top models",
      "comment_count": 2,
      "views": 0,
      "votes": 19
    },
    {
      "id": 347908,
      "title": "15th Place Solution Meta features ,FE, DART, CAT, XG , Tabnet , MLP , ensemble 😊",
      "comment_count": 5,
      "views": 0,
      "votes": 38
    },
    {
      "id": 347642,
      "title": "Tightest competition.",
      "comment_count": 1,
      "views": 0,
      "votes": 7
    },
    {
      "id": 347757,
      "title": "82-nd place (Silver) solution",
      "comment_count": 6,
      "views": 0,
      "votes": 20
    },
    {
      "id": 348352,
      "title": "Private Leaderboard 3442 - A Beginner's Simple Solution and Key Takeaways",
      "comment_count": 4,
      "views": 0,
      "votes": 2
    },
    {
      "id": 347745,
      "title": "2020 place overview",
      "comment_count": 9,
      "views": 0,
      "votes": 28
    },
    {
      "id": 347889,
      "title": "Share your best single model scores?",
      "comment_count": 28,
      "views": 0,
      "votes": 13
    },
    {
      "id": 344490,
      "title": "Let's stop publishing easy to fork high scoring notebooks.",
      "comment_count": 4,
      "views": 0,
      "votes": 23
    },
    {
      "id": 347640,
      "title": "Leaderboard Shakeup",
      "comment_count": 7,
      "views": 0,
      "votes": 19
    },
    {
      "id": 347722,
      "title": "27th place, +720 place shake up with NN model",
      "comment_count": 9,
      "views": 0,
      "votes": 35
    },
    {
      "id": 347858,
      "title": "[16th place solution] Features Diversity and Ensemble",
      "comment_count": 3,
      "views": 0,
      "votes": 27
    },
    {
      "id": 347809,
      "title": "What didn't work (and why ?)",
      "comment_count": 3,
      "views": 0,
      "votes": 19
    },
    {
      "id": 347647,
      "title": "Nine lines of feature engineering code to got 0.80727 private score --- alpha 191 factor base on Quant",
      "comment_count": 10,
      "views": 0,
      "votes": 31
    },
    {
      "id": 348243,
      "title": "Is there a better way than Grid Search to find the ideal parameters for XGBoost??",
      "comment_count": 6,
      "views": 0,
      "votes": 1
    },
    {
      "id": 347685,
      "title": "Bronze Medal Solution",
      "comment_count": 2,
      "views": 0,
      "votes": 33
    },
    {
      "id": 347181,
      "title": "Good luck!",
      "comment_count": 28,
      "views": 0,
      "votes": 36
    },
    {
      "id": 348273,
      "title": "Leakage from Out-Of-Fold features",
      "comment_count": 0,
      "views": 0,
      "votes": 7
    },
    {
      "id": 347992,
      "title": "Ultra Fast Adversarial Validation combined with SHAP Importance.",
      "comment_count": 1,
      "views": 0,
      "votes": 14
    },
    {
      "id": 347752,
      "title": "Silver Medal (105th Place) - Simple Solution (2x xgboost + NN).",
      "comment_count": 1,
      "views": 0,
      "votes": 24
    },
    {
      "id": 347956,
      "title": "showing more decimals",
      "comment_count": 2,
      "views": 0,
      "votes": 4
    },
    {
      "id": 347638,
      "title": "[Silver Medal] Our secret sauce (was using macro data)",
      "comment_count": 12,
      "views": 0,
      "votes": 18
    },
    {
      "id": 348035,
      "title": "when reality hits!!",
      "comment_count": 2,
      "views": 0,
      "votes": 8
    },
    {
      "id": 347632,
      "title": "Congratulations  for the winners",
      "comment_count": 7,
      "views": 0,
      "votes": 1
    },
    {
      "id": 347422,
      "title": "Meme Thread =))",
      "comment_count": 7,
      "views": 0,
      "votes": 19
    },
    {
      "id": 347652,
      "title": "Did you make the right model choice?",
      "comment_count": 11,
      "views": 0,
      "votes": 10
    },
    {
      "id": 346401,
      "title": "Ensemble competition",
      "comment_count": 9,
      "views": 0,
      "votes": 21
    },
    {
      "id": 347946,
      "title": "[To Competition Host]: Your Baseline Metric Score",
      "comment_count": 0,
      "views": 0,
      "votes": 6
    },
    {
      "id": 346145,
      "title": "[Rant] Why were we ever given the test data?",
      "comment_count": 18,
      "views": 0,
      "votes": 49
    },
    {
      "id": 347646,
      "title": "TFW you realize you put together ~5-6 bronze level submissions......",
      "comment_count": 2,
      "views": 0,
      "votes": 6
    },
    {
      "id": 346771,
      "title": "Interest to know how many time spend for a submissions",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 347656,
      "title": "Data first, Feature second, Algorithm third",
      "comment_count": 1,
      "views": 0,
      "votes": 5
    },
    {
      "id": 347397,
      "title": "How many decimal places will the private results be announced?",
      "comment_count": 4,
      "views": 0,
      "votes": 6
    },
    {
      "id": 347684,
      "title": "The private leaderboard is preliminary and will be finalized after the results are verified",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 347610,
      "title": "last insights",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 346897,
      "title": "Negative score !!!!!!",
      "comment_count": 6,
      "views": 0,
      "votes": 6
    },
    {
      "id": 344581,
      "title": "God, save LB from public works.",
      "comment_count": 11,
      "views": 0,
      "votes": 18
    },
    {
      "id": 343831,
      "title": "DO WE NEED TO SUBMIT FROM A NOTEBOOK?",
      "comment_count": 6,
      "views": 0,
      "votes": 2
    },
    {
      "id": 342757,
      "title": "Work after the end of the competition??",
      "comment_count": 9,
      "views": 0,
      "votes": 4
    },
    {
      "id": 347566,
      "title": "Is there any way we can understand more decimal on LB? ",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 347503,
      "title": "Hard to push the score  >.78 . Curious to know the trick to uplift it further?",
      "comment_count": 5,
      "views": 0,
      "votes": 2
    },
    {
      "id": 347705,
      "title": "Saving the best model doesn't seem to work",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 347626,
      "title": "Private/public scores",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 347635,
      "title": "Congratulation Winners ! ",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 347343,
      "title": "Clarification of rules",
      "comment_count": 4,
      "views": 0,
      "votes": 1
    },
    {
      "id": 347460,
      "title": "Engineering of additional features or management of missing data does not produce improvements to the score 0.799",
      "comment_count": 8,
      "views": 0,
      "votes": null
    },
    {
      "id": 347126,
      "title": "16 GB RAM for LGBM Instead of 13 GB",
      "comment_count": 0,
      "views": 0,
      "votes": 5
    },
    {
      "id": 347468,
      "title": "Problem with creating datasets",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 344038,
      "title": "Custom LGBM Obj: Weighted LogLoss Function",
      "comment_count": 17,
      "views": 0,
      "votes": 23
    },
    {
      "id": 347484,
      "title": "Help !!! .. test data exceeds allocated memory..",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 347636,
      "title": "Congratulations!!",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 347232,
      "title": "Failed Submission ",
      "comment_count": 12,
      "views": 0,
      "votes": null
    },
    {
      "id": 346024,
      "title": "Purpose of validation data",
      "comment_count": 6,
      "views": 0,
      "votes": 4
    },
    {
      "id": 346563,
      "title": "Help! Stacking OOF gone wrong?",
      "comment_count": 4,
      "views": 0,
      "votes": 5
    },
    {
      "id": 347467,
      "title": "Your notebook tried to allocate more memory than is available. It has restarted.",
      "comment_count": 4,
      "views": 0,
      "votes": 1
    },
    {
      "id": 346625,
      "title": "I should buy a 5995wx and 4090 GPU after this competition",
      "comment_count": 8,
      "views": 0,
      "votes": 5
    },
    {
      "id": 342679,
      "title": "t-SNE learns to separate data classess -- a short movie",
      "comment_count": 19,
      "views": 0,
      "votes": 28
    },
    {
      "id": 347399,
      "title": "why the early_stop not run in lightGBM when i set the fobj and feval",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 347141,
      "title": "How is Amex Metric Score Working?",
      "comment_count": 6,
      "views": 0,
      "votes": 1
    },
    {
      "id": 342992,
      "title": "Is 0.801 a bottleneck?  ",
      "comment_count": 18,
      "views": 0,
      "votes": 9
    },
    {
      "id": 347309,
      "title": "After competition end, is still possible to send a submission and check the score?",
      "comment_count": 4,
      "views": 0,
      "votes": null
    },
    {
      "id": 339734,
      "title": "What score do you think is theoretically possible?",
      "comment_count": 24,
      "views": 0,
      "votes": 11
    },
    {
      "id": 347281,
      "title": "submission is dead",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 343966,
      "title": "[Paper Summary] Why do tree-based models still outperform deep learning on tabular data?",
      "comment_count": 14,
      "views": 0,
      "votes": 31
    },
    {
      "id": 347143,
      "title": "Why today still 2 days to go？ Is the deadline delay？",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 346994,
      "title": "Is after 8.24, I can't make my submissions?",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 346578,
      "title": "What is it that the top performers are juicing out of?",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 346521,
      "title": "Is RNN appropriate in this task?",
      "comment_count": 8,
      "views": 0,
      "votes": 1
    },
    {
      "id": 346348,
      "title": "How does ranking work on the public LB for same scores?",
      "comment_count": 5,
      "views": 0,
      "votes": 4
    },
    {
      "id": 344178,
      "title": "A faster implementation to create diff columns",
      "comment_count": 8,
      "views": 0,
      "votes": 19
    },
    {
      "id": 346016,
      "title": "Final Shakeup - How screwed are you?",
      "comment_count": 2,
      "views": 0,
      "votes": 11
    },
    {
      "id": 346251,
      "title": "Adversarial validation on (Aggregated Data)",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 346657,
      "title": "How much does NN model increase your public leader board??",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 346617,
      "title": "Amex Metric Score - 0.99145420 with test data but public leaderboard score - 0.57",
      "comment_count": 20,
      "views": 0,
      "votes": -5
    },
    {
      "id": 346581,
      "title": "Some question for the submission",
      "comment_count": 9,
      "views": 0,
      "votes": 1
    },
    {
      "id": 343732,
      "title": "Free the GPU Memory",
      "comment_count": 4,
      "views": 0,
      "votes": 14
    },
    {
      "id": 328020,
      "title": "10x fast metric (numpy)",
      "comment_count": 13,
      "views": 0,
      "votes": 111
    },
    {
      "id": 346961,
      "title": "Code submission",
      "comment_count": 0,
      "views": 0,
      "votes": -2
    },
    {
      "id": 346020,
      "title": "How to accelerate the training speed?",
      "comment_count": 4,
      "views": 0,
      "votes": 4
    },
    {
      "id": 343972,
      "title": "CV score 0.8016 but LB score 0.795, where is my problem?",
      "comment_count": 17,
      "views": 0,
      "votes": 3
    },
    {
      "id": 346286,
      "title": "Private Dataset evaluation",
      "comment_count": 4,
      "views": 0,
      "votes": 2
    },
    {
      "id": 346181,
      "title": "Lots of deleted accounts in silver zone!!!",
      "comment_count": 5,
      "views": 0,
      "votes": 3
    },
    {
      "id": 344578,
      "title": "Custom LGBM Obj Version 2: Weighted LogLoss Function",
      "comment_count": 4,
      "views": 0,
      "votes": 11
    },
    {
      "id": 344846,
      "title": "pandas noob",
      "comment_count": 3,
      "views": 0,
      "votes": 6
    },
    {
      "id": 343011,
      "title": "Find your perfect NN!!",
      "comment_count": 6,
      "views": 0,
      "votes": 24
    },
    {
      "id": 346277,
      "title": "Potential Rework of the Competitions points system",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 346143,
      "title": "Ensembling Neural Networks",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 344808,
      "title": "Some Questions about Private Evaluation (Reproducibility Specs)",
      "comment_count": 8,
      "views": 0,
      "votes": 3
    },
    {
      "id": 346372,
      "title": "Question about SEED fixing",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 346453,
      "title": "The kernel appears to have died. It will restart automatically.",
      "comment_count": 4,
      "views": 0,
      "votes": 1
    },
    {
      "id": 344704,
      "title": "Features importance and hyper parameters optimization - which fisrt, which after?",
      "comment_count": 5,
      "views": 0,
      "votes": 1
    },
    {
      "id": 343949,
      "title": "Any advices to improve the LB to the 0.799??",
      "comment_count": 6,
      "views": 0,
      "votes": 1
    },
    {
      "id": 346443,
      "title": "train test split vs CV",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 342798,
      "title": "🤓 🤓 Feature Engineering Data 🚀🚀",
      "comment_count": 9,
      "views": 0,
      "votes": 15
    },
    {
      "id": 344592,
      "title": "How much time are serious contenders spending on this competition?  ",
      "comment_count": 8,
      "views": 0,
      "votes": 1
    },
    {
      "id": 346400,
      "title": "Not able to submit submission.csv from console",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 344263,
      "title": "How do I submit my code",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 346126,
      "title": "Unable to upload dataset to notebook",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 346282,
      "title": "where is 0.800 notebook? I just want to take a look",
      "comment_count": 2,
      "views": 0,
      "votes": -16
    },
    {
      "id": 344577,
      "title": "Top scores on leaderboard, Stacking/Ensembling or single model?",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 329787,
      "title": "Can you find the best seed?",
      "comment_count": 37,
      "views": 0,
      "votes": 127
    },
    {
      "id": 344612,
      "title": "Hidden column statistics when using .isnull().sum()",
      "comment_count": 8,
      "views": 0,
      "votes": null
    },
    {
      "id": 344619,
      "title": "how does rank ensemble work?",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 336557,
      "title": "Basic Feature Engineering - 1500 features",
      "comment_count": 20,
      "views": 0,
      "votes": 42
    },
    {
      "id": 344195,
      "title": "Do your model have low D and high G score same as mine? ",
      "comment_count": 6,
      "views": 0,
      "votes": 1
    },
    {
      "id": 346304,
      "title": "Great Topic",
      "comment_count": 1,
      "views": 0,
      "votes": -16
    },
    {
      "id": 344470,
      "title": "My lb score is lower than cv score?Can anyone explain?",
      "comment_count": 4,
      "views": 0,
      "votes": null
    },
    {
      "id": 344056,
      "title": "Test Data set size",
      "comment_count": 4,
      "views": 0,
      "votes": 0
    },
    {
      "id": 344825,
      "title": "One solution to the Out of Memory message",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 342622,
      "title": "Suddenly... I cannot read plain old test file",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 344706,
      "title": "which seeds is more importance, is cv seed or  model seed ？",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 338906,
      "title": "CV & LB of CatBoost",
      "comment_count": 24,
      "views": 0,
      "votes": 14
    },
    {
      "id": 344541,
      "title": "May i use the data to write my dissertation",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 341348,
      "title": "Searching an easy way to implement RNN...someone can help me?",
      "comment_count": 14,
      "views": 0,
      "votes": 3
    },
    {
      "id": 327116,
      "title": "Normalized Gini Coefficient (G). Default Rate (D).",
      "comment_count": 21,
      "views": 0,
      "votes": 94
    },
    {
      "id": 343191,
      "title": "🥢🪝Selected Features Dataset🪝🥢",
      "comment_count": 2,
      "views": 0,
      "votes": 7
    },
    {
      "id": 339726,
      "title": "Better models make better abstract art",
      "comment_count": 29,
      "views": 0,
      "votes": 51
    },
    {
      "id": 343004,
      "title": "## Using Encoding Target method",
      "comment_count": 10,
      "views": 0,
      "votes": 1
    },
    {
      "id": 342895,
      "title": "unable to enable GPU",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 342547,
      "title": "RAM out of memory in colab ",
      "comment_count": 7,
      "views": 0,
      "votes": 3
    },
    {
      "id": 339426,
      "title": "Stratification by P_2_LST can significantly reduce your CV variance",
      "comment_count": 14,
      "views": 0,
      "votes": 43
    },
    {
      "id": 343267,
      "title": "The difference between .apply(func) and .func()",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 343783,
      "title": "why my cv score +0.01 but lb score -0.01?",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 342967,
      "title": "Correlation analysis matters.",
      "comment_count": 0,
      "views": 0,
      "votes": 5
    },
    {
      "id": 342175,
      "title": "why do I have high variance in my cv?",
      "comment_count": 6,
      "views": 0,
      "votes": 14
    },
    {
      "id": 342817,
      "title": "Sharing The Details of CV Metrics",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 341538,
      "title": "How To Compress Humongous Dataset",
      "comment_count": 7,
      "views": 0,
      "votes": 2
    },
    {
      "id": 342803,
      "title": "What is the perfect score?",
      "comment_count": 4,
      "views": 0,
      "votes": 2
    },
    {
      "id": 342038,
      "title": "issue downloading parquet file from Google Cloud notebook",
      "comment_count": 3,
      "views": 0,
      "votes": 3
    },
    {
      "id": 332729,
      "title": "The \"Kaggle Ensembling Guide\"",
      "comment_count": 14,
      "views": 0,
      "votes": 55
    },
    {
      "id": 343177,
      "title": "what does merger deadline mean?",
      "comment_count": 2,
      "views": 0,
      "votes": -1
    },
    {
      "id": 342023,
      "title": "CatBoost + GPU + custom_metric = ERROR",
      "comment_count": 5,
      "views": 0,
      "votes": 4
    },
    {
      "id": 332575,
      "title": "DART, LGBM, Saving Best Models  [Callbacks Code Snippet]",
      "comment_count": 21,
      "views": 0,
      "votes": 89
    },
    {
      "id": 342847,
      "title": "submission file not created properly",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 341983,
      "title": "Can we use google data studio?",
      "comment_count": 4,
      "views": 0,
      "votes": 5
    },
    {
      "id": 342024,
      "title": "Look for team member and share my model",
      "comment_count": 5,
      "views": 0,
      "votes": 5
    },
    {
      "id": 341886,
      "title": "Length of training labels vs length of training data",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 332018,
      "title": "Tabular Deep Learning - A Tutorial",
      "comment_count": 11,
      "views": 0,
      "votes": 66
    },
    {
      "id": 342659,
      "title": "Looking for a teammate!",
      "comment_count": 1,
      "views": 0,
      "votes": -4
    },
    {
      "id": 335698,
      "title": "An Evil Thought Experiment",
      "comment_count": 22,
      "views": 0,
      "votes": 22
    },
    {
      "id": 341797,
      "title": "the problem of \"NN\" and \"GBDT\"",
      "comment_count": 4,
      "views": 0,
      "votes": 2
    },
    {
      "id": 341553,
      "title": "competition metric running on the GPU [pytorch]",
      "comment_count": 0,
      "views": 0,
      "votes": 8
    },
    {
      "id": 338712,
      "title": "For those looking for early_stopping for LGBM (dart)",
      "comment_count": 4,
      "views": 0,
      "votes": 5
    },
    {
      "id": 339278,
      "title": "How good are the predictions? See for yourself!",
      "comment_count": 19,
      "views": 0,
      "votes": 43
    },
    {
      "id": 331840,
      "title": "Summary: Basic pipeline for tabular competition for beginner!",
      "comment_count": 6,
      "views": 0,
      "votes": 58
    },
    {
      "id": 340932,
      "title": "Trying to get to .6",
      "comment_count": 9,
      "views": 0,
      "votes": 11
    },
    {
      "id": 341227,
      "title": "Ensemble OOF 101",
      "comment_count": 0,
      "views": 0,
      "votes": 11
    },
    {
      "id": 341724,
      "title": "What should i do, if the train feature distribution shape is sloped from the test feature distribution",
      "comment_count": 5,
      "views": 0,
      "votes": 2
    },
    {
      "id": 335689,
      "title": "Overfitting vs overfitting",
      "comment_count": 8,
      "views": 0,
      "votes": 20
    },
    {
      "id": 341256,
      "title": "How to use df.loc in cudf???",
      "comment_count": 6,
      "views": 0,
      "votes": null
    },
    {
      "id": 341061,
      "title": "TPUs are popular right now. You are #142 in the queue. You can wait, try connecting again later, or use another accelerator",
      "comment_count": 3,
      "views": 0,
      "votes": -9
    },
    {
      "id": 339451,
      "title": "See how LightGBM learns as it grows trees",
      "comment_count": 12,
      "views": 0,
      "votes": 31
    },
    {
      "id": 329425,
      "title": "Lessons learnt from Professional experience in Finance @ AmEx",
      "comment_count": 8,
      "views": 0,
      "votes": 15
    },
    {
      "id": 341248,
      "title": "Strategy for dealing with prediction files that are too big to hold in computer RAM",
      "comment_count": 1,
      "views": 0,
      "votes": 4
    },
    {
      "id": 341561,
      "title": "Are we guessing blindly here ??",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 341222,
      "title": "Better Ensembling Methods",
      "comment_count": 7,
      "views": 0,
      "votes": -1
    },
    {
      "id": 341633,
      "title": "how to load data and process data",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 341962,
      "title": "Importing the dataset takes much time ! Any workaround?",
      "comment_count": 1,
      "views": 0,
      "votes": -2
    },
    {
      "id": 341842,
      "title": "TypeError: No matching signature found",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 341827,
      "title": "Which hyperparameters should be adjusted in dart model?",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 341077,
      "title": "Pytorch LSTM not using GPU",
      "comment_count": 10,
      "views": 0,
      "votes": 2
    },
    {
      "id": 341166,
      "title": "How to convert \"csv-file\" to \"parquet-file\"",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 341174,
      "title": "My First Competition (Need Guidance) ",
      "comment_count": 4,
      "views": 0,
      "votes": 1
    },
    {
      "id": 339200,
      "title": "Observation time vs performance window",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 341125,
      "title": "I can't submit correctly because of chunking !!",
      "comment_count": 7,
      "views": 0,
      "votes": 1
    },
    {
      "id": 339541,
      "title": "how to use cudf on Google Colab?",
      "comment_count": 4,
      "views": 0,
      "votes": 4
    },
    {
      "id": 340542,
      "title": "Private Leaderboard and \"Ensemble\" notebook",
      "comment_count": 9,
      "views": 0,
      "votes": 4
    },
    {
      "id": 340945,
      "title": "difference between result on given data and test data",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 327143,
      "title": "⚡ 9x Data Compression achieved with Feather🕊️",
      "comment_count": 36,
      "views": 0,
      "votes": 157
    },
    {
      "id": 340714,
      "title": "Imputing values on training and testing sets does not work",
      "comment_count": 9,
      "views": 0,
      "votes": 2
    },
    {
      "id": 336229,
      "title": "How to make the model better?",
      "comment_count": 16,
      "views": 0,
      "votes": 13
    },
    {
      "id": 341033,
      "title": "machine resources",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 341171,
      "title": "The Default Rate Captured at 4%",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 337329,
      "title": "Feature Engineering - Inverse relationship between number of samples at one timestamp versus the overall default rate",
      "comment_count": 5,
      "views": 0,
      "votes": 8
    },
    {
      "id": 340793,
      "title": "Feather data is not available",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 328565,
      "title": "Let's catchup with all the learnings so far",
      "comment_count": 11,
      "views": 0,
      "votes": 143
    },
    {
      "id": 340002,
      "title": "Is Taking tail values is really worth it?",
      "comment_count": 11,
      "views": 0,
      "votes": 10
    },
    {
      "id": 339071,
      "title": "Feature Selection approach : Base->Blowup->Base",
      "comment_count": 14,
      "views": 0,
      "votes": 17
    },
    {
      "id": 338905,
      "title": "Dimensionality Reduction?",
      "comment_count": 12,
      "views": 0,
      "votes": 5
    },
    {
      "id": 330347,
      "title": "A List of EDA tricks when memory is limited",
      "comment_count": 14,
      "views": 0,
      "votes": 56
    },
    {
      "id": 338140,
      "title": "How overfit is the leaderboard?",
      "comment_count": 17,
      "views": 0,
      "votes": 49
    },
    {
      "id": 339389,
      "title": "AMEX REFERENCES",
      "comment_count": 4,
      "views": 0,
      "votes": 19
    },
    {
      "id": 339112,
      "title": "Can anything beat lightgbm and xgboost?",
      "comment_count": 7,
      "views": 0,
      "votes": 9
    },
    {
      "id": 338569,
      "title": "Are features correlated and does it matter?",
      "comment_count": 5,
      "views": 0,
      "votes": 1
    },
    {
      "id": 334046,
      "title": "How do you deal with missing features in the competition?",
      "comment_count": 12,
      "views": 0,
      "votes": 12
    },
    {
      "id": 337836,
      "title": "“problem description” I have a problem",
      "comment_count": 5,
      "views": 0,
      "votes": 1
    },
    {
      "id": 330444,
      "title": "Experiences from the field: A few things to remember about credit risk models in practice",
      "comment_count": 12,
      "views": 0,
      "votes": 40
    },
    {
      "id": 340082,
      "title": "Baseline Model Question",
      "comment_count": 8,
      "views": 0,
      "votes": 0
    },
    {
      "id": 338268,
      "title": "The curse of exotic scoring",
      "comment_count": 25,
      "views": 0,
      "votes": 39
    },
    {
      "id": 339625,
      "title": "How can I interpret the features?",
      "comment_count": 5,
      "views": 0,
      "votes": 3
    },
    {
      "id": 339583,
      "title": "Feature Selection approaches",
      "comment_count": 0,
      "views": 0,
      "votes": 14
    },
    {
      "id": 332645,
      "title": "Data Size Reducing🏋 + feather 🪶format data",
      "comment_count": 3,
      "views": 0,
      "votes": 7
    },
    {
      "id": 339059,
      "title": "Choice of Scalers for Neural Nets",
      "comment_count": 5,
      "views": 0,
      "votes": 6
    },
    {
      "id": 338809,
      "title": "Use CSI and KS may avoid overfitting",
      "comment_count": 3,
      "views": 0,
      "votes": 8
    },
    {
      "id": 338752,
      "title": "Boosting XGBoost model score - without slowness of DART",
      "comment_count": 7,
      "views": 0,
      "votes": 30
    },
    {
      "id": 339261,
      "title": "Why model scores are so sensitive to seed?",
      "comment_count": 6,
      "views": 0,
      "votes": 7
    },
    {
      "id": 339573,
      "title": "Bagging or full data training",
      "comment_count": 0,
      "views": 0,
      "votes": 8
    },
    {
      "id": 339238,
      "title": "Blending by rank",
      "comment_count": 2,
      "views": 0,
      "votes": 17
    },
    {
      "id": 339195,
      "title": "What is your best XGBoost score?",
      "comment_count": 5,
      "views": 0,
      "votes": 11
    },
    {
      "id": 338877,
      "title": "For Colab Users, I have a question about memory overflow problem.",
      "comment_count": 8,
      "views": 0,
      "votes": 1
    },
    {
      "id": 326904,
      "title": "New to Kaggle or Machine Learning? Come Say Hi!",
      "comment_count": 52,
      "views": 0,
      "votes": 51
    },
    {
      "id": 340296,
      "title": "Does any one see that the LB don't improve or change even with different predictions?",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 339191,
      "title": "Performance of models build with categorical columns only",
      "comment_count": 5,
      "views": 0,
      "votes": 4
    },
    {
      "id": 336546,
      "title": "Sharing my ablations studies",
      "comment_count": 27,
      "views": 0,
      "votes": 63
    },
    {
      "id": 338634,
      "title": "Dropping features affect on cv and LB?",
      "comment_count": 6,
      "views": 0,
      "votes": 6
    },
    {
      "id": 334391,
      "title": "Best Single NN Score",
      "comment_count": 8,
      "views": 0,
      "votes": 25
    },
    {
      "id": 331165,
      "title": "A checklist of ideas for Improving GBM Baselines",
      "comment_count": 9,
      "views": 0,
      "votes": 69
    },
    {
      "id": 340289,
      "title": "Does adding count occurrence column helps?",
      "comment_count": 0,
      "views": 0,
      "votes": -4
    },
    {
      "id": 337525,
      "title": "Optuna to find optimal weights",
      "comment_count": 14,
      "views": 0,
      "votes": 19
    },
    {
      "id": 327926,
      "title": "How to identify Public and Private",
      "comment_count": 8,
      "views": 0,
      "votes": 89
    },
    {
      "id": 331214,
      "title": "What do we know so far? - Insights from the top discussions ",
      "comment_count": 1,
      "views": 0,
      "votes": 19
    },
    {
      "id": 339817,
      "title": "GPU Notebook Queued",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 339787,
      "title": "How to Aggregate the test data?",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 339791,
      "title": "What submission require ? is it classification problem.",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 328651,
      "title": "Pivot datasets - One row for customer_ID",
      "comment_count": 5,
      "views": 0,
      "votes": 11
    },
    {
      "id": 338313,
      "title": "Data Preprocessing for Neural Networks",
      "comment_count": 4,
      "views": 0,
      "votes": 5
    },
    {
      "id": 339205,
      "title": "How to do a ranking ensemble?",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 339458,
      "title": "What's in there with 'D_63' and 'D_64' columns?",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 338366,
      "title": "Data Understanding ",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 337610,
      "title": "What models ensemble well?",
      "comment_count": 15,
      "views": 0,
      "votes": 46
    },
    {
      "id": 327464,
      "title": "Graphical explanation of the competition metric",
      "comment_count": 31,
      "views": 0,
      "votes": 254
    },
    {
      "id": 339033,
      "title": "\"Fun fact\":All competitions on kaggle exist using future data.",
      "comment_count": 5,
      "views": 0,
      "votes": 2
    },
    {
      "id": 337609,
      "title": "Which score should we trust more, CV or LB?",
      "comment_count": 17,
      "views": 0,
      "votes": 18
    },
    {
      "id": 339162,
      "title": "Is reproducibility important?",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 337920,
      "title": "Is the dataset normalized?",
      "comment_count": 14,
      "views": 0,
      "votes": 1
    },
    {
      "id": 337962,
      "title": "Strange behaviour of the Leaderboard",
      "comment_count": 7,
      "views": 0,
      "votes": 8
    },
    {
      "id": 337819,
      "title": "gc.collect() vs _ = gc.collect()",
      "comment_count": 6,
      "views": 0,
      "votes": 6
    },
    {
      "id": 332574,
      "title": "Towards data de-anonymization",
      "comment_count": 11,
      "views": 0,
      "votes": 71
    },
    {
      "id": 338121,
      "title": "📆 Last statement dates (S_2): default rate seasonality in train set March 2018",
      "comment_count": 2,
      "views": 0,
      "votes": 8
    },
    {
      "id": 327205,
      "title": "Tutorial on reading large datasets by Rohan",
      "comment_count": 12,
      "views": 0,
      "votes": 48
    },
    {
      "id": 337259,
      "title": "Feature selection techniques commonly used on kaggle",
      "comment_count": 3,
      "views": 0,
      "votes": 7
    },
    {
      "id": 338620,
      "title": "train labels",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 337813,
      "title": "When are models different enough to complement each other?",
      "comment_count": 1,
      "views": 0,
      "votes": 21
    },
    {
      "id": 338283,
      "title": "Experimenting with different Early Stopping strategies",
      "comment_count": 1,
      "views": 0,
      "votes": 6
    },
    {
      "id": 338001,
      "title": "Avg validation CV or overall validation CV?",
      "comment_count": 4,
      "views": 0,
      "votes": 6
    },
    {
      "id": 336957,
      "title": "Unstable random seed behaviour on public LB",
      "comment_count": 13,
      "views": 0,
      "votes": 28
    },
    {
      "id": 338318,
      "title": "Usability of F1 score for AmEx competition",
      "comment_count": 5,
      "views": 0,
      "votes": 1
    },
    {
      "id": 327106,
      "title": "Training data starter",
      "comment_count": 7,
      "views": 0,
      "votes": 35
    },
    {
      "id": 326905,
      "title": "Looking for a Team Megathread",
      "comment_count": 224,
      "views": 0,
      "votes": 25
    },
    {
      "id": 329467,
      "title": "Feature 'S_2' (Monthly Statement date): initial analysis and engineering",
      "comment_count": 1,
      "views": 0,
      "votes": 15
    },
    {
      "id": 337846,
      "title": "Out of memory reading test file (parquet, 3.3GiB). Any suggestions?",
      "comment_count": 9,
      "views": 0,
      "votes": 6
    },
    {
      "id": 338363,
      "title": "Customer segmentation",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 338076,
      "title": "Does incremental training of LGBM negatively affect performance?",
      "comment_count": 1,
      "views": 0,
      "votes": 4
    },
    {
      "id": 337188,
      "title": "Model ensembling",
      "comment_count": 6,
      "views": 0,
      "votes": 27
    },
    {
      "id": 338161,
      "title": "Kaggle Kernel Restarts Randomly",
      "comment_count": 4,
      "views": 0,
      "votes": 3
    },
    {
      "id": 336475,
      "title": "4th decimal sign for submission scoring",
      "comment_count": 12,
      "views": 0,
      "votes": 19
    },
    {
      "id": 338361,
      "title": "How does the Leaderboard work?",
      "comment_count": 4,
      "views": 0,
      "votes": null
    },
    {
      "id": 338572,
      "title": "How do you guys load the data?",
      "comment_count": 1,
      "views": 0,
      "votes": -1
    },
    {
      "id": 337891,
      "title": "Difference between DART algorithm and Neural Network dropout?",
      "comment_count": 4,
      "views": 0,
      "votes": 1
    },
    {
      "id": 333953,
      "title": "What All we can try to improve our model performance?",
      "comment_count": 16,
      "views": 0,
      "votes": 34
    },
    {
      "id": 338289,
      "title": "Duplicates in test dataset ",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 337274,
      "title": "New way to monitor and experiment-tracking",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 338188,
      "title": "Clarification on subsampling needed",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 333000,
      "title": "Best Single Model",
      "comment_count": 32,
      "views": 0,
      "votes": 54
    },
    {
      "id": 337736,
      "title": "What does the weighting in the scoring metric mean ? ",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 330259,
      "title": "Compress data size notebook.",
      "comment_count": 2,
      "views": 0,
      "votes": -2
    },
    {
      "id": 328756,
      "title": "The distribution of missing values over time",
      "comment_count": 9,
      "views": 0,
      "votes": 100
    },
    {
      "id": 337160,
      "title": "Tree correlation vs highly correlated features",
      "comment_count": 11,
      "views": 0,
      "votes": 6
    },
    {
      "id": 336911,
      "title": "Label Encoding is NOT essential for the LightGBM",
      "comment_count": 6,
      "views": 0,
      "votes": 14
    },
    {
      "id": 336625,
      "title": "Scoring, How Does it Work?",
      "comment_count": 6,
      "views": 0,
      "votes": 4
    },
    {
      "id": 337488,
      "title": "Example request: LightGBM w/ Early Stopping",
      "comment_count": 3,
      "views": 0,
      "votes": 3
    },
    {
      "id": 335986,
      "title": "Two notebook medals for the same code",
      "comment_count": 27,
      "views": 0,
      "votes": 34
    },
    {
      "id": 337530,
      "title": "AmEx Competition 2022 Individual Customer Data",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 337460,
      "title": "AMEX Datasets as Pickles - 918 Features",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 338007,
      "title": "What is the competition metric? Accuracy, precision, recall, or F1 score?",
      "comment_count": 1,
      "views": 0,
      "votes": -9
    },
    {
      "id": 337731,
      "title": "Missing Value Imputation",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 327158,
      "title": "Default Definition",
      "comment_count": 3,
      "views": 0,
      "votes": 14
    },
    {
      "id": 337729,
      "title": "Slow notebook",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 337789,
      "title": "Any tips anyone is willing to share",
      "comment_count": 1,
      "views": 0,
      "votes": -4
    },
    {
      "id": 330486,
      "title": "Some useful information I found about Features and Feature groups ",
      "comment_count": 10,
      "views": 0,
      "votes": 56
    },
    {
      "id": 336145,
      "title": "Feature Selection",
      "comment_count": 10,
      "views": 0,
      "votes": 31
    },
    {
      "id": 337005,
      "title": "Too time-consuming to make agg-like features",
      "comment_count": 11,
      "views": 0,
      "votes": 3
    },
    {
      "id": 336193,
      "title": "Customers with fewer statements have a higher probability of default",
      "comment_count": 6,
      "views": 0,
      "votes": 22
    },
    {
      "id": 337055,
      "title": "Loading the data ",
      "comment_count": 3,
      "views": 0,
      "votes": 3
    },
    {
      "id": 336146,
      "title": "Ranking approaches ?",
      "comment_count": 8,
      "views": 0,
      "votes": 13
    },
    {
      "id": 327148,
      "title": "Insights from a previous default prediction competition",
      "comment_count": 12,
      "views": 0,
      "votes": 95
    },
    {
      "id": 336926,
      "title": "My notebook memory is just exploding to max 16 GB and restarts",
      "comment_count": 6,
      "views": 0,
      "votes": 2
    },
    {
      "id": 336967,
      "title": "AMEX HDF5 - Last Statement - Train Only",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 329738,
      "title": "Big \"positional\" shakeup",
      "comment_count": 13,
      "views": 0,
      "votes": 47
    },
    {
      "id": 336704,
      "title": "Need Hints on Feature Engineering",
      "comment_count": 5,
      "views": 0,
      "votes": 3
    },
    {
      "id": 335944,
      "title": "Permutation importance XGB",
      "comment_count": 7,
      "views": 0,
      "votes": 18
    },
    {
      "id": 336349,
      "title": "Features to detect outliers?",
      "comment_count": 11,
      "views": 0,
      "votes": 6
    },
    {
      "id": 335524,
      "title": "Lag Features Are All You Need",
      "comment_count": 6,
      "views": 0,
      "votes": 55
    },
    {
      "id": 329436,
      "title": "What can we still do when we have no information about the feaures",
      "comment_count": 7,
      "views": 0,
      "votes": 61
    },
    {
      "id": 336714,
      "title": "Variable Explanation",
      "comment_count": 2,
      "views": 0,
      "votes": 3
    },
    {
      "id": 336143,
      "title": "Feature engineering ideas - Blanket approach",
      "comment_count": 2,
      "views": 0,
      "votes": 19
    },
    {
      "id": 329094,
      "title": "Bypass the features anonymization with LightGBM's feature interactions",
      "comment_count": 5,
      "views": 0,
      "votes": 63
    },
    {
      "id": 337080,
      "title": "Overview、Data Desciption, Rulesの日本語訳（Japanese translation）",
      "comment_count": 0,
      "views": 0,
      "votes": -13
    },
    {
      "id": 336064,
      "title": "Nice to meet you!",
      "comment_count": 2,
      "views": 0,
      "votes": -1
    },
    {
      "id": 330931,
      "title": "How To Select Features?",
      "comment_count": 6,
      "views": 0,
      "votes": 30
    },
    {
      "id": 334603,
      "title": "Correlation KOs",
      "comment_count": 7,
      "views": 0,
      "votes": 10
    },
    {
      "id": 327441,
      "title": "🗜️10x Compression of whole dataset using Feather",
      "comment_count": 3,
      "views": 0,
      "votes": 6
    },
    {
      "id": 336149,
      "title": "What does it mean to balance variables in the context of credit risk?",
      "comment_count": 2,
      "views": 0,
      "votes": 3
    },
    {
      "id": 327162,
      "title": "Amex metric using pd.Series",
      "comment_count": 3,
      "views": 0,
      "votes": 9
    },
    {
      "id": 327761,
      "title": "Time Series EDA and GRU Starter - LB 0.790",
      "comment_count": 14,
      "views": 0,
      "votes": 85
    },
    {
      "id": 335423,
      "title": "How to make your Model More robust ?",
      "comment_count": 2,
      "views": 0,
      "votes": 13
    },
    {
      "id": 335054,
      "title": "Large Datasets from Kaggle to Google CoLab",
      "comment_count": 16,
      "views": 0,
      "votes": 21
    },
    {
      "id": 327160,
      "title": "Minimize memory load",
      "comment_count": 2,
      "views": 0,
      "votes": 5
    },
    {
      "id": 335918,
      "title": "Why not 4 digits precision in the public leaderboard?",
      "comment_count": 4,
      "views": 0,
      "votes": 1
    },
    {
      "id": 329447,
      "title": "Ensembling models focusing on parts of the competition metric",
      "comment_count": 4,
      "views": 0,
      "votes": 11
    },
    {
      "id": 327094,
      "title": "Last month per customer",
      "comment_count": 6,
      "views": 0,
      "votes": 64
    },
    {
      "id": 335398,
      "title": "Does Adversarial Validation Always help in Improving scores on unseen test data ?",
      "comment_count": 2,
      "views": 0,
      "votes": 6
    },
    {
      "id": 335158,
      "title": "Doubt on 18 months window",
      "comment_count": 2,
      "views": 0,
      "votes": 17
    },
    {
      "id": 329103,
      "title": "Ensembling probabilities with log-odds",
      "comment_count": 14,
      "views": 0,
      "votes": 50
    },
    {
      "id": 335540,
      "title": "最頻値について(About mode)",
      "comment_count": 7,
      "views": 0,
      "votes": null
    },
    {
      "id": 336133,
      "title": "Customers with 13 rows?",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 334139,
      "title": "Should one tune hyperparameters before feature engineering?",
      "comment_count": 12,
      "views": 0,
      "votes": 12
    },
    {
      "id": 332880,
      "title": "Help Needed!",
      "comment_count": 10,
      "views": 0,
      "votes": 6
    },
    {
      "id": 334886,
      "title": "Reading the CSV file",
      "comment_count": 6,
      "views": 0,
      "votes": 3
    },
    {
      "id": 335587,
      "title": "Curious about the model",
      "comment_count": 1,
      "views": 0,
      "votes": 4
    },
    {
      "id": 334157,
      "title": "G and D learning curves",
      "comment_count": 6,
      "views": 0,
      "votes": 30
    },
    {
      "id": 331725,
      "title": "Understanding NA in the dataset",
      "comment_count": 14,
      "views": 0,
      "votes": 66
    },
    {
      "id": 336014,
      "title": "Loading Issues",
      "comment_count": 1,
      "views": 0,
      "votes": -2
    },
    {
      "id": 335705,
      "title": "Does the target encoding work?",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 335222,
      "title": "how to load big file in Rapid  ?",
      "comment_count": 5,
      "views": 0,
      "votes": 3
    },
    {
      "id": 335998,
      "title": "No customer falls into default within the 13-month period",
      "comment_count": 2,
      "views": 0,
      "votes": -5
    },
    {
      "id": 334956,
      "title": "Reading Gigabytes size files in Notebook",
      "comment_count": 6,
      "views": 0,
      "votes": 6
    },
    {
      "id": 335629,
      "title": "Reading data through parquet is also memory constraints?",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 335209,
      "title": "How to use different notebooks for same dataset?",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 334100,
      "title": "Ensemble with 3 decimals",
      "comment_count": 8,
      "views": 0,
      "votes": 11
    },
    {
      "id": 335362,
      "title": "Strange overflow error (df.write_parquet()",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 335322,
      "title": "Data Preprocessing ",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 333940,
      "title": "RAPIDS Feature Engineering",
      "comment_count": 5,
      "views": 0,
      "votes": 23
    },
    {
      "id": 333824,
      "title": "Collaborate with my son",
      "comment_count": 5,
      "views": 0,
      "votes": 9
    },
    {
      "id": 334113,
      "title": "The label distribution of train set will strongly impact model metric",
      "comment_count": 3,
      "views": 0,
      "votes": 5
    },
    {
      "id": 332286,
      "title": "The Random Kaggle estimator",
      "comment_count": 23,
      "views": 0,
      "votes": 43
    },
    {
      "id": 333668,
      "title": "Strange plot(D86_mean)",
      "comment_count": 5,
      "views": 0,
      "votes": 7
    },
    {
      "id": 333507,
      "title": "Tips needed for getting good performance in Collab with GPU for cudf,cupy,XGB and LGMB ",
      "comment_count": 7,
      "views": 0,
      "votes": 9
    },
    {
      "id": 327400,
      "title": "⚡[FAST LOADING] only 1.4GB Training Data using Feather",
      "comment_count": 1,
      "views": 0,
      "votes": 33
    },
    {
      "id": 333351,
      "title": "Time Series Pattern of B_2",
      "comment_count": 2,
      "views": 0,
      "votes": 24
    },
    {
      "id": 332218,
      "title": "FYI: Nvidia blog - How American Express Uses Deep Learning for Better Decision Making",
      "comment_count": 5,
      "views": 0,
      "votes": 42
    },
    {
      "id": 333118,
      "title": "what does \"Note that the negative class has been subsampled for this dataset at 5%, and thus receives a 20x weighting in the scoring metric.\" mean?",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 332930,
      "title": "Tabular Data Augmentations",
      "comment_count": 4,
      "views": 0,
      "votes": 25
    },
    {
      "id": 327295,
      "title": "CV vs LB Scores",
      "comment_count": 14,
      "views": 0,
      "votes": 29
    },
    {
      "id": 333864,
      "title": "Data (train, test)for best performing model? ",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 331634,
      "title": "Cohorts explained",
      "comment_count": 1,
      "views": 0,
      "votes": 14
    },
    {
      "id": 333575,
      "title": "\"the probability of a future payment default (target = 1)\" means?",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 332221,
      "title": "Correlation to target?",
      "comment_count": 1,
      "views": 0,
      "votes": 6
    },
    {
      "id": 333734,
      "title": "running out of memory for loading data ",
      "comment_count": 6,
      "views": 0,
      "votes": 3
    },
    {
      "id": 333917,
      "title": "Taking too long to turn test.csv into a dataframe",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 331148,
      "title": "How to Measure K-Fold Performance?",
      "comment_count": 8,
      "views": 0,
      "votes": 12
    },
    {
      "id": 333434,
      "title": "Time gap between Public / Private LB ",
      "comment_count": 2,
      "views": 0,
      "votes": 4
    },
    {
      "id": 333514,
      "title": "cudaErrorMemoryAllocation out of memory",
      "comment_count": 2,
      "views": 0,
      "votes": 3
    },
    {
      "id": 332461,
      "title": "Pearson's and Spearman's correlation for high-cardinality features",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 333342,
      "title": "Can't implement Standardization or PCA due to memory limitation",
      "comment_count": 3,
      "views": 0,
      "votes": 3
    },
    {
      "id": 332923,
      "title": "Bad day at Kaggle ?",
      "comment_count": 3,
      "views": 0,
      "votes": 8
    },
    {
      "id": 333333,
      "title": "Column Documentation?",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 332558,
      "title": "An explanation of the Boruta-Shap Feature Selection Technique",
      "comment_count": 0,
      "views": 0,
      "votes": 19
    },
    {
      "id": 333313,
      "title": "Can't Load Dataset",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 327651,
      "title": "Strange Histograms",
      "comment_count": 14,
      "views": 0,
      "votes": 73
    },
    {
      "id": 333302,
      "title": "Anybody wanna team up?",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 332797,
      "title": "The best way to learn is to practice",
      "comment_count": 5,
      "views": 0,
      "votes": 5
    },
    {
      "id": 332446,
      "title": "Understanding Features with Quantiles",
      "comment_count": 4,
      "views": 0,
      "votes": 5
    },
    {
      "id": 331454,
      "title": "How to move from beyond 0.795? ",
      "comment_count": 10,
      "views": 0,
      "votes": 28
    },
    {
      "id": 332038,
      "title": "Is it ethical to refuse service to a customer based on predicted default?",
      "comment_count": 8,
      "views": 0,
      "votes": 23
    },
    {
      "id": 328303,
      "title": "The large dataset must not make you give up reading it !",
      "comment_count": 2,
      "views": 0,
      "votes": 5
    },
    {
      "id": 333281,
      "title": "does the system you want, is going predict depending on more then 1 record(row) or just 1 record as inputs?",
      "comment_count": 0,
      "views": 0,
      "votes": -1
    },
    {
      "id": 327602,
      "title": "Final/Private dataset evaluation on October 2019 statement",
      "comment_count": 5,
      "views": 0,
      "votes": 42
    },
    {
      "id": 327649,
      "title": "The data has uniform random noise injected",
      "comment_count": 10,
      "views": 0,
      "votes": 82
    },
    {
      "id": 332627,
      "title": "Prediction Submission",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 332009,
      "title": "My Approach with Pyspark",
      "comment_count": 3,
      "views": 0,
      "votes": 3
    },
    {
      "id": 331887,
      "title": "Memory efficient ways to save predictions of various models",
      "comment_count": 6,
      "views": 0,
      "votes": 10
    },
    {
      "id": 332505,
      "title": "Can the competition use the Internet?",
      "comment_count": 5,
      "views": 0,
      "votes": 1
    },
    {
      "id": 327361,
      "title": "Another way to keep only the last month without groupby",
      "comment_count": 5,
      "views": 0,
      "votes": 34
    },
    {
      "id": 331388,
      "title": "Caution of using \"publicly shared codes\"",
      "comment_count": 4,
      "views": 0,
      "votes": 26
    },
    {
      "id": 331886,
      "title": "Understanding competition metric ",
      "comment_count": 0,
      "views": 0,
      "votes": 14
    },
    {
      "id": 332417,
      "title": "Looking for people to create a team (python coders)",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 331389,
      "title": "Interesting paper: SCARF: SELF-SUPERVISED CONTRASTIVE LEARNING USING RANDOM FEATURE CORRUPTION",
      "comment_count": 2,
      "views": 0,
      "votes": 20
    },
    {
      "id": 332579,
      "title": "Access to Parquet/Feather data",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 332559,
      "title": "Is sorting in order important ???",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 327258,
      "title": "what do the terms mean",
      "comment_count": 4,
      "views": 0,
      "votes": 8
    },
    {
      "id": 330987,
      "title": "Temporal features in the dataset",
      "comment_count": 2,
      "views": 0,
      "votes": 41
    },
    {
      "id": 330432,
      "title": "Memory friendly dataset in Parquet",
      "comment_count": 5,
      "views": 0,
      "votes": 4
    },
    {
      "id": 331320,
      "title": "Revealing time patterns of features",
      "comment_count": 0,
      "views": 0,
      "votes": 14
    },
    {
      "id": 332094,
      "title": " How to read the csv files , without getting memory issues. ",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 331866,
      "title": "How to decrease the size of data files?",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 331778,
      "title": "Compressed Data Set for Amex - Default Prediction",
      "comment_count": 4,
      "views": 0,
      "votes": 1
    },
    {
      "id": 331007,
      "title": "AMEX : Missing Value or Outlier Treatment ? ",
      "comment_count": 5,
      "views": 0,
      "votes": 11
    },
    {
      "id": 331172,
      "title": "Papers: Machine Learning + Credit default prediction",
      "comment_count": 1,
      "views": 0,
      "votes": 10
    },
    {
      "id": 330949,
      "title": "Question regarding Default rate at 4% ",
      "comment_count": 4,
      "views": 0,
      "votes": 1
    },
    {
      "id": 330754,
      "title": "Your notebook tried to allocate more memory than is available. It has restarted Error.",
      "comment_count": 11,
      "views": 0,
      "votes": 2
    },
    {
      "id": 328057,
      "title": "Analysis of Information loss during conversion from float64 to float16",
      "comment_count": 3,
      "views": 0,
      "votes": 28
    },
    {
      "id": 331180,
      "title": "difference between 'empty', '0.0', '-1.0', '-1' && category vs label?",
      "comment_count": 0,
      "views": 0,
      "votes": 4
    },
    {
      "id": 331059,
      "title": "RAPIDS cuDF aggregation operation numeric stability problem",
      "comment_count": 3,
      "views": 0,
      "votes": 4
    },
    {
      "id": 330130,
      "title": "Maybe release the fourth decimal place in the leaderboard?",
      "comment_count": 8,
      "views": 0,
      "votes": 29
    },
    {
      "id": 330952,
      "title": "Dataset Dictionary",
      "comment_count": 1,
      "views": 0,
      "votes": 4
    },
    {
      "id": 330714,
      "title": "No Matching Signature Found Error When Using ffill()",
      "comment_count": 5,
      "views": 0,
      "votes": 1
    },
    {
      "id": 330449,
      "title": "Could someone explain me what is magic feature?",
      "comment_count": 6,
      "views": 0,
      "votes": 6
    },
    {
      "id": 329088,
      "title": "Why is there so much test data?",
      "comment_count": 12,
      "views": 0,
      "votes": 40
    },
    {
      "id": 330381,
      "title": "How to do feature engineering  and adjust parameters?",
      "comment_count": 6,
      "views": 0,
      "votes": 3
    },
    {
      "id": 328462,
      "title": "So many GrandMasters！！",
      "comment_count": 10,
      "views": 0,
      "votes": 23
    },
    {
      "id": 330990,
      "title": "Call For a Team ",
      "comment_count": 0,
      "views": 0,
      "votes": -2
    },
    {
      "id": 329647,
      "title": "Need help: How to run in Colab  🙏",
      "comment_count": 10,
      "views": 0,
      "votes": 6
    },
    {
      "id": 330245,
      "title": "How to download parquet file from website? ",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 329879,
      "title": "American Express - Default Prediction",
      "comment_count": 4,
      "views": 0,
      "votes": 6
    },
    {
      "id": 330710,
      "title": "EDA on Anonymous Features",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 329736,
      "title": "Is there any chance of shake up?",
      "comment_count": 6,
      "views": 0,
      "votes": 8
    },
    {
      "id": 329695,
      "title": "CV and LB do not seem to correlate well",
      "comment_count": 3,
      "views": 0,
      "votes": 8
    },
    {
      "id": 329685,
      "title": "methods to load test data set",
      "comment_count": 2,
      "views": 0,
      "votes": 9
    },
    {
      "id": 327595,
      "title": "Mutiple vintages for test data - April 2019 and Oct 2019",
      "comment_count": 5,
      "views": 0,
      "votes": 6
    },
    {
      "id": 329275,
      "title": "Be aware of the counts per time! ",
      "comment_count": 7,
      "views": 0,
      "votes": 16
    },
    {
      "id": 329706,
      "title": "what is the feature meaning, higher is worse?",
      "comment_count": 0,
      "views": 0,
      "votes": 7
    },
    {
      "id": 329808,
      "title": "Too big to fit in memory",
      "comment_count": 3,
      "views": 0,
      "votes": 3
    },
    {
      "id": 329607,
      "title": "Competition evaluation Metrics formula",
      "comment_count": 2,
      "views": 0,
      "votes": 4
    },
    {
      "id": 327142,
      "title": "Understanding the Input data using Pandas",
      "comment_count": 1,
      "views": 0,
      "votes": 22
    },
    {
      "id": 328846,
      "title": "🥇 Model Performance Ranking",
      "comment_count": 8,
      "views": 0,
      "votes": 35
    },
    {
      "id": 329057,
      "title": "question from newbie",
      "comment_count": 4,
      "views": 0,
      "votes": 10
    },
    {
      "id": 329534,
      "title": "Duplicate Customer IDs in test data and hence in submission",
      "comment_count": 1,
      "views": 0,
      "votes": 5
    },
    {
      "id": 329756,
      "title": "Should test predictions be rounded?",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 329465,
      "title": "\"Unsupported type_id conversion to cudf\" while reading from feather",
      "comment_count": 2,
      "views": 0,
      "votes": 3
    },
    {
      "id": 330232,
      "title": "Add me in your team",
      "comment_count": 2,
      "views": 0,
      "votes": -9
    },
    {
      "id": 328084,
      "title": "Advanced EDA - UMAP/Hdbscan",
      "comment_count": 5,
      "views": 0,
      "votes": 13
    },
    {
      "id": 329160,
      "title": "Beware of wrong merges",
      "comment_count": 0,
      "views": 0,
      "votes": 9
    },
    {
      "id": 328801,
      "title": "Hopular with GBDT ....so XGBoost is all you need =)))",
      "comment_count": 1,
      "views": 0,
      "votes": 21
    },
    {
      "id": 327629,
      "title": "Cross-validation",
      "comment_count": 9,
      "views": 0,
      "votes": 10
    },
    {
      "id": 329428,
      "title": "Fast Data Processing",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 329689,
      "title": "American Express Credit Prediction-Traine LightGBM Model",
      "comment_count": 2,
      "views": 0,
      "votes": -5
    },
    {
      "id": 328138,
      "title": "Risk of overflow with Float 16 conversion",
      "comment_count": 4,
      "views": 0,
      "votes": 15
    },
    {
      "id": 328885,
      "title": "Predictive Features",
      "comment_count": 0,
      "views": 0,
      "votes": 17
    },
    {
      "id": 328230,
      "title": "Discrepancy between train\\val metrics and leaderboard metrics",
      "comment_count": 4,
      "views": 0,
      "votes": 6
    },
    {
      "id": 328859,
      "title": "It seems as if deep learning models are not performing well...",
      "comment_count": 2,
      "views": 0,
      "votes": 5
    },
    {
      "id": 327597,
      "title": "Minority Report",
      "comment_count": 12,
      "views": 0,
      "votes": 25
    },
    {
      "id": 328890,
      "title": "R version of the metric?",
      "comment_count": 2,
      "views": 0,
      "votes": 4
    },
    {
      "id": 328686,
      "title": "Amex Metric in Pytorch",
      "comment_count": 4,
      "views": 0,
      "votes": 3
    },
    {
      "id": 328343,
      "title": "Are we in for a Surprise in private LB?",
      "comment_count": 3,
      "views": 0,
      "votes": 11
    },
    {
      "id": 329003,
      "title": "How to submit? Unique Customer ID in test data is lesser than sample submission.",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 327765,
      "title": "GBDT or NN,which is the winner of this competition",
      "comment_count": 9,
      "views": 0,
      "votes": 25
    },
    {
      "id": 327135,
      "title": "Articles, Research Papers and Methodologies ",
      "comment_count": 2,
      "views": 0,
      "votes": 59
    },
    {
      "id": 328819,
      "title": "AE -Customer Default Prediction Research",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 327084,
      "title": "Is this the largest tabular dataset on Kaggle?",
      "comment_count": 15,
      "views": 0,
      "votes": 32
    },
    {
      "id": 327984,
      "title": "Evaluation Metric in Datatable (6x speedup over Pandas) 🔥",
      "comment_count": 2,
      "views": 0,
      "votes": 13
    },
    {
      "id": 327759,
      "title": "Good Correlation!",
      "comment_count": 7,
      "views": 0,
      "votes": 11
    },
    {
      "id": 328245,
      "title": "Is there a shorter version?",
      "comment_count": 4,
      "views": 0,
      "votes": 4
    },
    {
      "id": 327161,
      "title": "Unique number of categorical features",
      "comment_count": 1,
      "views": 0,
      "votes": 29
    },
    {
      "id": 328083,
      "title": "Nan while calculating mean",
      "comment_count": 2,
      "views": 0,
      "votes": 3
    },
    {
      "id": 327908,
      "title": "American Express Default Prediction Snappy Parquet Files (50 GB -> 10 GB)",
      "comment_count": 2,
      "views": 0,
      "votes": 9
    },
    {
      "id": 327924,
      "title": "Collection of discussions for Amex Competition",
      "comment_count": 0,
      "views": 0,
      "votes": 9
    },
    {
      "id": 328508,
      "title": "Does Features Like this Make Sense?",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 327922,
      "title": "An obvious pointer?",
      "comment_count": 2,
      "views": 0,
      "votes": 5
    },
    {
      "id": 327612,
      "title": "Which format of data best ! Used in large amount of data?",
      "comment_count": 10,
      "views": 0,
      "votes": 4
    },
    {
      "id": 327696,
      "title": "What I found !",
      "comment_count": 3,
      "views": 0,
      "votes": 12
    },
    {
      "id": 327513,
      "title": "Does anyone have any ideas about this competition？",
      "comment_count": 6,
      "views": 0,
      "votes": 3
    },
    {
      "id": 327609,
      "title": "Custom metric without pandas DataFrame, and accelerated by numba",
      "comment_count": 3,
      "views": 0,
      "votes": 9
    },
    {
      "id": 327558,
      "title": "Add 2 decimal places to the LB",
      "comment_count": 1,
      "views": 0,
      "votes": 15
    },
    {
      "id": 328034,
      "title": "Distributions - American Express - Default Prediction",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 327901,
      "title": "Similar Compeition: Home Credit Default Risk Solutions & Code Summary (🥇Gold Medal)",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 327195,
      "title": "Reading & Working with Large Dataset",
      "comment_count": 0,
      "views": 0,
      "votes": 30
    },
    {
      "id": 327228,
      "title": "6.53GB Dataset + Kaggle Notebook training and Inference Pipeline",
      "comment_count": 1,
      "views": 0,
      "votes": 21
    },
    {
      "id": 327268,
      "title": "Compressed Dataset with targets (~10x compression) & Notebook",
      "comment_count": 0,
      "views": 0,
      "votes": 20
    },
    {
      "id": 327265,
      "title": "Maximum runtime?",
      "comment_count": 5,
      "views": 0,
      "votes": 6
    },
    {
      "id": 327263,
      "title": "What is the benchmark entry in the leaderboard?",
      "comment_count": 2,
      "views": 0,
      "votes": 9
    },
    {
      "id": 327164,
      "title": "R snippet to estimate Normalized Gini Coefficient.",
      "comment_count": 0,
      "views": 0,
      "votes": 16
    },
    {
      "id": 327965,
      "title": "Segmentation Fault 😱 with PyArrow & Dask",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 327367,
      "title": "About the nature of the target",
      "comment_count": 2,
      "views": 0,
      "votes": 6
    },
    {
      "id": 327104,
      "title": "The tabular data is soooooo huge that reading is a problem! Sad but exciting.",
      "comment_count": 5,
      "views": 0,
      "votes": 6
    },
    {
      "id": 327906,
      "title": "Similar Compeition: G-Research Crypto Forecasting Solution & Code Summary (🥇Gold Medal)",
      "comment_count": 0,
      "views": 0,
      "votes": -9
    },
    {
      "id": 327101,
      "title": "Have fun and good luck :)",
      "comment_count": 0,
      "views": 0,
      "votes": 8
    },
    {
      "id": 327180,
      "title": "What a familiar taste---Tabular Data",
      "comment_count": 0,
      "views": 0,
      "votes": 5
    },
    {
      "id": 327124,
      "title": "Wow! This is a very large tabular data!",
      "comment_count": 0,
      "views": 0,
      "votes": 5
    }
  ],
  "errors": []
}