{
  "id": 173628,
  "title": "can any one explain to me how we can apply  over-sampling for this problem ???",
  "url": "/competitions/siim-isic-melanoma-classification/discussion/173628",
  "author_name": "",
  "post_date": "2020-08-10T04:29:43.584057300Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>```</p>\n\n<h1>Class count</h1>\n\n<p>count_class_0, count_class_1 = train.target.value_counts()</p>\n\n<h1>Divide by class</h1>\n\n<p>df_0 = train[train['target'] == 0]\ndf_1= train[train['target'] == 1]\n```</p>\n\n<p>```\ndf_class_1_over = df_1.sample(count_class_0, replace=True)\ndf = pd.concat([df_0 , df_class_1_over], axis=0)</p>\n\n<p>print('Random over-sampling:')\nprint(df.target.value_counts())</p>\n\n<p>df.target.value_counts().plot(kind='bar', title='Count (target)');\n```</p>",
  "messages": [
    {
      "id": "964647",
      "postDate": "08/10/2020 04:29:43",
      "content": "<p>```</p>\n\n<h1>Class count</h1>\n\n<p>count_class_0, count_class_1 = train.target.value_counts()</p>\n\n<h1>Divide by class</h1>\n\n<p>df_0 = train[train['target'] == 0]\ndf_1= train[train['target'] == 1]\n```</p>\n\n<p>```\ndf_class_1_over = df_1.sample(count_class_0, replace=True)\ndf = pd.concat([df_0 , df_class_1_over], axis=0)</p>\n\n<p>print('Random over-sampling:')\nprint(df.target.value_counts())</p>\n\n<p>df.target.value_counts().plot(kind='bar', title='Count (target)');\n```</p>",
      "rawMarkdown": "```\n# Class count\ncount_class_0, count_class_1 = train.target.value_counts()\n\n# Divide by class\ndf_0 = train[train['target'] == 0]\ndf_1= train[train['target'] == 1]\n```\n\n\n```\ndf_class_1_over = df_1.sample(count_class_0, replace=True)\ndf = pd.concat([df_0 , df_class_1_over], axis=0)\n\nprint('Random over-sampling:')\nprint(df.target.value_counts())\n\ndf.target.value_counts().plot(kind='bar', title='Count (target)');\n```",
      "votes": null
    },
    {
      "id": "964821",
      "postDate": "08/10/2020 07:43:27",
      "content": "<p><a href=\"/cdeotte\">@cdeotte</a> has provided a way <a href=\"https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/169139\">here</a></p>",
      "rawMarkdown": "cdeotte has provided a way [here](https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/169139)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 964821,
      "author_name": "vicioussong",
      "author_url": "",
      "post_date": "08/10/2020 07:43:27",
      "content": "<p><a href=\"/cdeotte\">@cdeotte</a> has provided a way <a href=\"https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/169139\">here</a></p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "964647": "```\n# Class count\ncount_class_0, count_class_1 = train.target.value_counts()\n\n# Divide by class\ndf_0 = train[train['target'] == 0]\ndf_1= train[train['target'] == 1]\n```\n\n\n```\ndf_class_1_over = df_1.sample(count_class_0, replace=True)\ndf = pd.concat([df_0 , df_class_1_over], axis=0)\n\nprint('Random over-sampling:')\nprint(df.target.value_counts())\n\ndf.target.value_counts().plot(kind='bar', title='Count (target)');\n```",
    "964821": "cdeotte has provided a way [here](https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/169139)"
  },
  "source": "meta"
}