{
  "fetched_at": "2026-10-19T17:10:01Z",
  "source_url": "https://www.kaggle.com/competitions/epf-montpellier-introduction-to-data-science/data",
  "pages": [
    {
      "name": "data-description",
      "content": "<p>We provide a small subset of auctions processed by our SSP during a week. Each auction is described by a set of features defining the context in which the auctions has run.</p>\n\n<p>The target variable we are trying to predict is a boolean defining wether the auction led to a sale or not</p>\n\n<h2>File descriptions</h2>\n<ul>\n<li><strong>train.csv</strong> - the training set</li>\n<li><strong>test.csv</strong> - the test set</li>\n<li><strong>testSubmissionFile.csv</strong> - a sample submission file in the correct format</li>\n</ul>\n<h2>Data fields</h2>\n<ul>\n<li><strong>auctionId</strong> - a random anonymous id unique to a given auction</li>\n<li><strong>timeStamp</strong> - the timestamp (in seconds) of the start of the auction</li>\n<li><strong>placementId</strong> - the id of a zone in the web page where the ad would be displayed</li>\n<li><strong>websiteId</strong> - the id of the corresponding website</li>\n<li><strong>refererDeepThree</strong> - the URL of the page where the ad would be displayed, truncated at its 3rd level and anonymized</li>\n<li><strong>country</strong> - the id of the user's country</li>\n<li><strong>operatingSystem</strong> - the user's operating system</li>\n<li><strong>browser</strong> - the user's browser</li>\n<li><strong>browserVersion</strong> - the user's browser version</li>\n<li><strong>device</strong> - the user's device type</li>\n<li><strong>environmentType</strong> - the type of the webpage environment</li>\n<li><strong>integrationType</strong> - the id of the webpage integration type</li>\n<li><strong>articleSafenessCategorization</strong> - the categorization of the content of the article</li>\n</ul>\n<h2>Target variable</h2>\n<ul>\n<li><strong>isSold</strong> - a boolean variable defining wether the auction led to a sale or not</li>\n</ul>",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "<p>In online advertising, real-time bidding refers to the buying and selling of online ad impressions through real-time auctions that occur in the time it takes a webpage to load. Those auctions are often facilitated by ad exchanges or supply-side platforms (SSP).</p>\n<p>As an ad impression loads in a user’s Web browser, information about the page it is on and the user viewing it is passed to an ad exchange, which auctions it off to the advertiser willing to pay the highest price for it. The winning bidder’s ad is then loaded into the webpage nearly instantly; the whole process takes just milliseconds to complete.</p>\n<p>Teads SSP processes billions of auctions per day. Only a part of these auctions would lead to an impression, as a part of the inventory is deemed not interesting by the buyers.</p>\n<p>The processing of these auctions has a significant computational cost. It involves some CPU-heavy computations, then the bid requests are sent to potential buyers, thus incurring some networking costs.\n</p>\n<h2>Project goal</h2>\n<p>The goal of this prediction model is to predict which auctions would lead to a sale and which auctions would not. By doing so, we can choose to not conduct auctions if we are confident that they would not lead to a sale. This can help us significantly decrease our infrastructure costs.</p>\n<p>Before each auction, we dispose of some information about the context. This information can be used as an input to a classification model. The output would be wether this auction would lead to a sale or not.</p>",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "<hr>\n<p>The evaluation metric for this competition is <a href=\"https://www.kaggle.com/wiki/MeanFScore\"> Mean F1-Score</a>. The F1 score, commonly used in information retrieval, measures accuracy using the statistics precision p and recall r. Precision is the ratio of true positives (tp) to all predicted positives (tp + fp). Recall is the ratio of true positives to all actual positives (tp + fn). The F1 score is given by:</p>\n<p>\\[ F1 = 2\\frac{p \\cdot r}{p+r}\\ \\ \\mathrm{where}\\ \\ p = \\frac{tp}{tp+fp},\\ \\ r = \\frac{tp}{tp+fn} \\]</p>\n<p>The F1 metric weights recall and precision equally, and a good retrieval algorithm will maximize both precision and recall simultaneously. Thus, moderately good performance on both will be favored over extremely good performance on one and poor performance on the other.</p>\n<h2>Submission Format</h2>\n<p><strong>For every auction in the dataset</strong>, submission files should contain two columns: auctionId and isSold. isSold should be a boolean. </p>\n<p>The file should contain a header and have the following format:</p>\n<pre>auctionId,isSold<br>00bce6c5-d97a-4c5c-a67d-621e21ec283a,false<br></pre>",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": " - Apply yourself!\n - Have fun!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}