{
  "id": 78402,
  "title": "What's the best way to deal with outliers in  purchase_amount ?",
  "url": "/competitions/elo-merchant-category-recommendation/discussion/78402",
  "author_name": "",
  "post_date": "2019-01-23T09:47:52.961135700Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>what's the best as you think to deal with outliers in purchase amount ? </p>",
  "messages": [
    {
      "id": "460254",
      "postDate": "01/23/2019 09:47:52",
      "content": "<p>what's the best as you think to deal with outliers in purchase amount ? </p>",
      "rawMarkdown": "what's the best as you think to deal with outliers in purchase amount ?",
      "votes": null
    },
    {
      "id": "462112",
      "postDate": "01/27/2019 17:55:51",
      "content": "<p>Hi Haithem\nActually the best way to deal with outliers is imputing values\nhow ?\nFor example you have a numerical data points range from 1 to 5 but near 5 there are only 2 data points and mostly data points are from 1 to 4 Now what you will do you have to impute value 4 in place of data points at 5..\nand if you have a  outliers in categorical variable first you have to find out the frequency of categories impute high freq category in place of outliers\nand\nif you column contain mostly null values you should have to drop it. </p>",
      "rawMarkdown": "Hi Haithem\nActually the best way to deal with outliers is imputing values\nhow ?\nFor example you have a numerical data points range from 1 to 5 but near 5 there are only 2 data points and mostly data points are from 1 to 4 Now what you will do you have to impute value 4 in place of data points at 5..\nand if you have a  outliers in categorical variable first you have to find out the frequency of categories impute high freq category in place of outliers\nand\nif you column contain mostly null values you should have to drop it.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 462112,
      "author_name": "tahirmuslimscientist",
      "author_url": "",
      "post_date": "01/27/2019 17:55:51",
      "content": "<p>Hi Haithem\nActually the best way to deal with outliers is imputing values\nhow ?\nFor example you have a numerical data points range from 1 to 5 but near 5 there are only 2 data points and mostly data points are from 1 to 4 Now what you will do you have to impute value 4 in place of data points at 5..\nand if you have a  outliers in categorical variable first you have to find out the frequency of categories impute high freq category in place of outliers\nand\nif you column contain mostly null values you should have to drop it. </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "460254": "what's the best as you think to deal with outliers in purchase amount ?",
    "462112": "Hi Haithem\nActually the best way to deal with outliers is imputing values\nhow ?\nFor example you have a numerical data points range from 1 to 5 but near 5 there are only 2 data points and mostly data points are from 1 to 4 Now what you will do you have to impute value 4 in place of data points at 5..\nand if you have a  outliers in categorical variable first you have to find out the frequency of categories impute high freq category in place of outliers\nand\nif you column contain mostly null values you should have to drop it."
  },
  "source": "meta"
}