{
  "id": 318928,
  "title": "Clustering and feature importance notebook",
  "url": "/competitions/h-and-m-personalized-fashion-recommendations/discussion/318928",
  "author_name": "Ben Lebovitz",
  "post_date": "2022-04-14T15:33:36.874000",
  "votes": 1,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Might be useful to someone… I did the following:</p>\n<ol>\n<li>Used k-means to group articles of clothing </li>\n<li>Grouped number of sales for total sales for each item</li>\n<li>Used a random forest to predict number of sales for each item</li>\n</ol>\n<p>The point being here to be able to cluster items in groups via k-means, to be used in a later model for recommending to users who bought items from the same clusters. And also to be able to see what features are most influential for each item being sold, and thus allowing to narrow down feature selection in future models.</p>\n<p>Check it out here: <br>\n<a href=\"https://www.kaggle.com/code/beezus666/k-means-and-feature-importance-for-articles\" target=\"_blank\">https://www.kaggle.com/code/beezus666/k-means-and-feature-importance-for-articles</a></p>",
  "messages": [
    {
      "id": 1755429,
      "postDate": "2022-04-14T15:33:36.873Z",
      "content": "<p>Might be useful to someone… I did the following:</p>\n<ol>\n<li>Used k-means to group articles of clothing </li>\n<li>Grouped number of sales for total sales for each item</li>\n<li>Used a random forest to predict number of sales for each item</li>\n</ol>\n<p>The point being here to be able to cluster items in groups via k-means, to be used in a later model for recommending to users who bought items from the same clusters. And also to be able to see what features are most influential for each item being sold, and thus allowing to narrow down feature selection in future models.</p>\n<p>Check it out here: <br>\n<a href=\"https://www.kaggle.com/code/beezus666/k-means-and-feature-importance-for-articles\" target=\"_blank\">https://www.kaggle.com/code/beezus666/k-means-and-feature-importance-for-articles</a></p>",
      "rawMarkdown": "Might be useful to someone... I did the following:\n1. Used k-means to group articles of clothing \n2. Grouped number of sales for total sales for each item\n3. Used a random forest to predict number of sales for each item\n\nThe point being here to be able to cluster items in groups via k-means, to be used in a later model for recommending to users who bought items from the same clusters. And also to be able to see what features are most influential for each item being sold, and thus allowing to narrow down feature selection in future models.\n\nCheck it out here: \nhttps://www.kaggle.com/code/beezus666/k-means-and-feature-importance-for-articles",
      "votes": 1
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1755429": "Might be useful to someone... I did the following:\n1. Used k-means to group articles of clothing \n2. Grouped number of sales for total sales for each item\n3. Used a random forest to predict number of sales for each item\n\nThe point being here to be able to cluster items in groups via k-means, to be used in a later model for recommending to users who bought items from the same clusters. And also to be able to see what features are most influential for each item being sold, and thus allowing to narrow down feature selection in future models.\n\nCheck it out here: \nhttps://www.kaggle.com/code/beezus666/k-means-and-feature-importance-for-articles"
  }
}