{
  "id": 180424,
  "title": "Guidelines on creating model and improving accuracy",
  "url": "/competitions/osic-pulmonary-fibrosis-progression/discussion/180424",
  "author_name": "",
  "post_date": "2020-09-05T01:31:23.436564500Z",
  "votes": 4,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hi,<br>\nI created a model and has now a score of -7.86.<br>\nI am trying to improve the accuracy of the model.<br>\nI am using keras library.<br>\nI am using image, Week, Age, Sex, Smoking status, FVC as input parameters.<br>\nIs there any reference for this?</p>\n<p>Thanks,</p>",
  "messages": [
    {
      "id": "998702",
      "postDate": "09/05/2020 01:31:23",
      "content": "<p>Hi,<br>\nI created a model and has now a score of -7.86.<br>\nI am trying to improve the accuracy of the model.<br>\nI am using keras library.<br>\nI am using image, Week, Age, Sex, Smoking status, FVC as input parameters.<br>\nIs there any reference for this?</p>\n<p>Thanks,</p>",
      "rawMarkdown": "Hi,\nI created a model and has now a score of -7.86.\nI am trying to improve the accuracy of the model.\nI am using keras library.\nI am using image, Week, Age, Sex, Smoking status, FVC as input parameters.\nIs there any reference for this?\n\nThanks,",
      "votes": null
    },
    {
      "id": "999217",
      "postDate": "09/05/2020 13:38:44",
      "content": "<p>Hi! <a href=\"https://www.kaggle.com/rajpatel5521\" target=\"_blank\">@rajpatel5521</a> Here is my strategy.<br>\n0 Exploratory Data Analysis<br>\n1 Feature Engineering (Most Important)<br>\n2 Model Selection &amp; Construction (Try multiple models DT, NN, GLM)<br>\n3 Cross-Validation Strategy (Helps to have a robust model, Calibrate to LB)<br>\n4 Upload Model Predictions<br>\n5 Check Forum for New Ideas / Strategies / Read Papers</p>\n<p>1 Feature Engineering Again (Most Important)<br>\n2 Model Selection &amp; Construction (Try multiple models DT, NN, GLM)<br>\n3 Cross-Validation Strategy (Helps to have a robust model, Calibrate to LB)<br>\n4 Upload Model Predictions<br>\n5 Check Forum for New Ideas / Strategies / Read Papers</p>\n<p>6 Model Blend or Stacking</p>",
      "rawMarkdown": "Hi! @rajpatel5521 Here is my strategy.\n0 Exploratory Data Analysis\n1 Feature Engineering (Most Important)\n2 Model Selection & Construction (Try multiple models DT, NN, GLM)\n3 Cross-Validation Strategy (Helps to have a robust model, Calibrate to LB)\n4 Upload Model Predictions\n5 Check Forum for New Ideas / Strategies / Read Papers\n\n1 Feature Engineering Again (Most Important)\n2 Model Selection & Construction (Try multiple models DT, NN, GLM)\n3 Cross-Validation Strategy (Helps to have a robust model, Calibrate to LB)\n4 Upload Model Predictions\n5 Check Forum for New Ideas / Strategies / Read Papers\n\n6 Model Blend or Stacking",
      "votes": null
    },
    {
      "id": "1004650",
      "postDate": "09/09/2020 22:30:13",
      "content": "<p>I agree with <a href=\"https://www.kaggle.com/cv13j0\" target=\"_blank\">@cv13j0</a> a similar process. EDA, DATA CLEANING, MODEL ARCHITECTURE , TEST TRAIN, Ie. Research medical imaging dicom papers and experiments there are lots of approaches , models and variations.</p>",
      "rawMarkdown": "I agree with @cv13j0 a similar process. EDA, DATA CLEANING, MODEL ARCHITECTURE , TEST TRAIN, Ie. Research medical imaging dicom papers and experiments there are lots of approaches , models and variations.",
      "votes": null
    },
    {
      "id": "1004733",
      "postDate": "09/10/2020 01:17:25",
      "content": "<p>Thanks, for the confirmation <a href=\"https://www.kaggle.com/simo333\" target=\"_blank\">@simo333</a> </p>",
      "rawMarkdown": "Thanks, for the confirmation @simo333",
      "votes": null
    },
    {
      "id": "1015044",
      "postDate": "09/17/2020 22:19:47",
      "content": "<p>I analyzed the model, read some documents and improved the accuracy from 7.9 to 7.25.</p>",
      "rawMarkdown": "I analyzed the model, read some documents and improved the accuracy from 7.9 to 7.25.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 999217,
      "author_name": "cv13j0",
      "author_url": "",
      "post_date": "09/05/2020 13:38:44",
      "content": "<p>Hi! <a href=\"https://www.kaggle.com/rajpatel5521\" target=\"_blank\">@rajpatel5521</a> Here is my strategy.<br>\n0 Exploratory Data Analysis<br>\n1 Feature Engineering (Most Important)<br>\n2 Model Selection &amp; Construction (Try multiple models DT, NN, GLM)<br>\n3 Cross-Validation Strategy (Helps to have a robust model, Calibrate to LB)<br>\n4 Upload Model Predictions<br>\n5 Check Forum for New Ideas / Strategies / Read Papers</p>\n<p>1 Feature Engineering Again (Most Important)<br>\n2 Model Selection &amp; Construction (Try multiple models DT, NN, GLM)<br>\n3 Cross-Validation Strategy (Helps to have a robust model, Calibrate to LB)<br>\n4 Upload Model Predictions<br>\n5 Check Forum for New Ideas / Strategies / Read Papers</p>\n<p>6 Model Blend or Stacking</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1004650,
      "author_name": "simo333",
      "author_url": "",
      "post_date": "09/09/2020 22:30:13",
      "content": "<p>I agree with <a href=\"https://www.kaggle.com/cv13j0\" target=\"_blank\">@cv13j0</a> a similar process. EDA, DATA CLEANING, MODEL ARCHITECTURE , TEST TRAIN, Ie. Research medical imaging dicom papers and experiments there are lots of approaches , models and variations.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1004733,
          "author_name": "cv13j0",
          "author_url": "",
          "post_date": "09/10/2020 01:17:25",
          "content": "<p>Thanks, for the confirmation <a href=\"https://www.kaggle.com/simo333\" target=\"_blank\">@simo333</a> </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1015044,
      "author_name": "rajpatel5521",
      "author_url": "",
      "post_date": "09/17/2020 22:19:47",
      "content": "<p>I analyzed the model, read some documents and improved the accuracy from 7.9 to 7.25.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "998702": "Hi,\nI created a model and has now a score of -7.86.\nI am trying to improve the accuracy of the model.\nI am using keras library.\nI am using image, Week, Age, Sex, Smoking status, FVC as input parameters.\nIs there any reference for this?\n\nThanks,",
    "999217": "Hi! @rajpatel5521 Here is my strategy.\n0 Exploratory Data Analysis\n1 Feature Engineering (Most Important)\n2 Model Selection & Construction (Try multiple models DT, NN, GLM)\n3 Cross-Validation Strategy (Helps to have a robust model, Calibrate to LB)\n4 Upload Model Predictions\n5 Check Forum for New Ideas / Strategies / Read Papers\n\n1 Feature Engineering Again (Most Important)\n2 Model Selection & Construction (Try multiple models DT, NN, GLM)\n3 Cross-Validation Strategy (Helps to have a robust model, Calibrate to LB)\n4 Upload Model Predictions\n5 Check Forum for New Ideas / Strategies / Read Papers\n\n6 Model Blend or Stacking",
    "1004650": "I agree with @cv13j0 a similar process. EDA, DATA CLEANING, MODEL ARCHITECTURE , TEST TRAIN, Ie. Research medical imaging dicom papers and experiments there are lots of approaches , models and variations.",
    "1004733": "Thanks, for the confirmation @simo333",
    "1015044": "I analyzed the model, read some documents and improved the accuracy from 7.9 to 7.25."
  },
  "source": "meta"
}