{
  "id": 97008,
  "title": "Accuracy stuck at 72",
  "url": "/competitions/diabetic-retinopathy-detection/discussion/97008",
  "author_name": "",
  "post_date": "2019-06-25T07:40:15.945951700Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>which preprocessing techniques should be used for this type of data to approve our model accuracy. Need some suggestion for improving my accuracy.  I am just using CNN model without any preprocessing. </p>\n\n<p>Kaggle Notebook Link: <a href=\"https://www.kaggle.com/sohaibanwaar1203/drd-c-simple\">https://www.kaggle.com/sohaibanwaar1203/drd-c-simple</a></p>",
  "messages": [
    {
      "id": "560237",
      "postDate": "06/25/2019 07:40:15",
      "content": "<p>which preprocessing techniques should be used for this type of data to approve our model accuracy. Need some suggestion for improving my accuracy.  I am just using CNN model without any preprocessing. </p>\n\n<p>Kaggle Notebook Link: <a href=\"https://www.kaggle.com/sohaibanwaar1203/drd-c-simple\">https://www.kaggle.com/sohaibanwaar1203/drd-c-simple</a></p>",
      "rawMarkdown": "which preprocessing techniques should be used for this type of data to approve our model accuracy. Need some suggestion for improving my accuracy.  I am just using CNN model without any preprocessing. \n\nKaggle Notebook Link: https://www.kaggle.com/sohaibanwaar1203/drd-c-simple",
      "votes": null
    },
    {
      "id": "2926435",
      "postDate": "07/17/2024 18:04:49",
      "content": "<p>Hello <a href=\"https://www.kaggle.com/sohaibanwaar1203\" target=\"_blank\">@sohaibanwaar1203</a>,</p>\n<p>I would suggest plotting Y and X variables using lowess smoothing, looking at the association pattern and modifying the model according to the pattern. You may need to fit a non-linear pattern model or use transformations. </p>",
      "rawMarkdown": "Hello @sohaibanwaar1203,\n\nI would suggest plotting Y and X variables using lowess smoothing, looking at the association pattern and modifying the model according to the pattern. You may need to fit a non-linear pattern model or use transformations.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2926435,
      "author_name": "moinuddin",
      "author_url": "",
      "post_date": "07/17/2024 18:04:49",
      "content": "<p>Hello <a href=\"https://www.kaggle.com/sohaibanwaar1203\" target=\"_blank\">@sohaibanwaar1203</a>,</p>\n<p>I would suggest plotting Y and X variables using lowess smoothing, looking at the association pattern and modifying the model according to the pattern. You may need to fit a non-linear pattern model or use transformations. </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "560237": "which preprocessing techniques should be used for this type of data to approve our model accuracy. Need some suggestion for improving my accuracy.  I am just using CNN model without any preprocessing. \n\nKaggle Notebook Link: https://www.kaggle.com/sohaibanwaar1203/drd-c-simple",
    "2926435": "Hello @sohaibanwaar1203,\n\nI would suggest plotting Y and X variables using lowess smoothing, looking at the association pattern and modifying the model according to the pattern. You may need to fit a non-linear pattern model or use transformations."
  },
  "source": "meta"
}