This had been one of the Best Kaggle Competition for me , so far , although I didn't got a Medal still I think I learnt alot from this competition , Most Important skillset that I developed in this competition are .
\n1) Understanding of How to deal with Large Datasets .
\n2) Different Feature Engineering for Tabular Datasets/
\n3) Coming up with a Cross Validation Strategy for Time Based Competitions
\n4) Loop Feature Engineering .
\n5) Transformer Models.
\nI would like to thanks all the awesome Kernels author whose worked I used in this competition and they allow me to get a decent standings.
Thanks to @rohanrao for sharing how to deal with large datasets and his kernel on Online Learning Algorithm Follow the Regularized Leader (FTRL) .
\nThanks to @its717 for sharing his Validation Strategy which was very reliable and his Kernels on Loop Feature Engineering .
\nThanks to @hdkim for sharing Transformers Kernel , Also thanks to @manikanthhr5 and other authors who expanded upon Transformer SAKT Model , @ammarnassanalhajali Kernel on Combining Lightgbm with SAKT was also very useful.
\nThis competition was exciting because of unique format and also lot of sharing by top contributors .
", "messages": [ { "id": "1143732", "postDate": "01/08/2021 03:12:55", "content": "This had been one of the Best Kaggle Competition for me , so far , although I didn't got a Medal still I think I learnt alot from this competition , Most Important skillset that I developed in this competition are .
\n1) Understanding of How to deal with Large Datasets .
\n2) Different Feature Engineering for Tabular Datasets/
\n3) Coming up with a Cross Validation Strategy for Time Based Competitions
\n4) Loop Feature Engineering .
\n5) Transformer Models.
\nI would like to thanks all the awesome Kernels author whose worked I used in this competition and they allow me to get a decent standings.
Thanks to @rohanrao for sharing how to deal with large datasets and his kernel on Online Learning Algorithm Follow the Regularized Leader (FTRL) .
\nThanks to @its717 for sharing his Validation Strategy which was very reliable and his Kernels on Loop Feature Engineering .
\nThanks to @hdkim for sharing Transformers Kernel , Also thanks to @manikanthhr5 and other authors who expanded upon Transformer SAKT Model , @ammarnassanalhajali Kernel on Combining Lightgbm with SAKT was also very useful.
\nThis competition was exciting because of unique format and also lot of sharing by top contributors .
", "rawMarkdown": "This had been one of the Best Kaggle Competition for me , so far , although I didn't got a Medal still I think I learnt alot from this competition , Most Important skillset that I developed in this competition are .\n1) Understanding of How to deal with Large Datasets .\n2) Different Feature Engineering for Tabular Datasets/\n3) Coming up with a Cross Validation Strategy for Time Based Competitions \n4) Loop Feature Engineering .\n5) Transformer Models.\nI would like to thanks all the awesome Kernels author whose worked I used in this competition and they allow me to get a decent standings.\n\nThanks to @rohanrao for sharing how to deal with large datasets and his kernel on Online Learning Algorithm Follow the Regularized Leader (FTRL) . \n\nThanks to @its717 for sharing his Validation Strategy which was very reliable and his Kernels on Loop Feature Engineering .\n\nThanks to @hdkim for sharing Transformers Kernel , Also thanks to @manikanthhr5 and other authors who expanded upon Transformer SAKT Model , @ammarnassanalhajali Kernel on Combining Lightgbm with SAKT was also very useful.\n\nThis competition was exciting because of unique format and also lot of sharing by top contributors .", "votes": null }, { "id": "1144072", "postDate": "01/08/2021 08:11:12", "content": "+1 for this. This competition was lot of fun and learning. I’m thrilled that I could manage a transformer in our ensemble and couldn’t have done it without going through many forum posts and kernels on the topic.
\nThanks to the community!!
+100. I really appreciate it.
", "rawMarkdown": "100. I really appreciate it.", "votes": null } ], "comments": [ { "id": 1144072, "author_name": "rashmibanthia", "author_url": "", "post_date": "01/08/2021 08:11:12", "content": "+1 for this. This competition was lot of fun and learning. I’m thrilled that I could manage a transformer in our ensemble and couldn’t have done it without going through many forum posts and kernels on the topic.
\nThanks to the community!!
+100. I really appreciate it.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1143732": "This had been one of the Best Kaggle Competition for me , so far , although I didn't got a Medal still I think I learnt alot from this competition , Most Important skillset that I developed in this competition are .\n1) Understanding of How to deal with Large Datasets .\n2) Different Feature Engineering for Tabular Datasets/\n3) Coming up with a Cross Validation Strategy for Time Based Competitions \n4) Loop Feature Engineering .\n5) Transformer Models.\nI would like to thanks all the awesome Kernels author whose worked I used in this competition and they allow me to get a decent standings.\n\nThanks to @rohanrao for sharing how to deal with large datasets and his kernel on Online Learning Algorithm Follow the Regularized Leader (FTRL) . \n\nThanks to @its717 for sharing his Validation Strategy which was very reliable and his Kernels on Loop Feature Engineering .\n\nThanks to @hdkim for sharing Transformers Kernel , Also thanks to @manikanthhr5 and other authors who expanded upon Transformer SAKT Model , @ammarnassanalhajali Kernel on Combining Lightgbm with SAKT was also very useful.\n\nThis competition was exciting because of unique format and also lot of sharing by top contributors .", "1144072": "1 for this. This competition was lot of fun and learning. I’m thrilled that I could manage a transformer in our ensemble and couldn’t have done it without going through many forum posts and kernels on the topic. \nThanks to the community!!", "1145125": "100. I really appreciate it." }, "source": "meta" }