{
  "id": 51176,
  "title": "Welcome from TalkingData!",
  "url": "/competitions/talkingdata-adtracking-fraud-detection/discussion/51176",
  "author_name": "",
  "post_date": "2018-03-06T07:42:33.305992900Z",
  "votes": 12,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Hi Kagglers,</p>\n\n<p>Welcome to the TalkingData AdTracking Fraud Detection Challenge! We are excited about our second competition with Kaggle, and we hope that you will enjoy working on this problem. Our data science team will be on this forum throughout the competition to answer your questions.</p>\n\n<p>We hope you will gain insight into the Chinese mobile market through looking at our dataset. Since click fraud happens at such a large volume, the problem you are solving is very important to us. We are looking forward to your innovative solutions. Thank you for participating and good luck! 祝你好运！</p>\n\n<p>Best Wishes,</p>\n\n<p>The TalkingData Team</p>",
  "messages": [
    {
      "id": "291473",
      "postDate": "03/06/2018 07:42:33",
      "content": "<p>Hi Kagglers,</p>\n\n<p>Welcome to the TalkingData AdTracking Fraud Detection Challenge! We are excited about our second competition with Kaggle, and we hope that you will enjoy working on this problem. Our data science team will be on this forum throughout the competition to answer your questions.</p>\n\n<p>We hope you will gain insight into the Chinese mobile market through looking at our dataset. Since click fraud happens at such a large volume, the problem you are solving is very important to us. We are looking forward to your innovative solutions. Thank you for participating and good luck! 祝你好运！</p>\n\n<p>Best Wishes,</p>\n\n<p>The TalkingData Team</p>",
      "rawMarkdown": "Hi Kagglers,\n\nWelcome to the TalkingData AdTracking Fraud Detection Challenge! We are excited about our second competition with Kaggle, and we hope that you will enjoy working on this problem. Our data science team will be on this forum throughout the competition to answer your questions.\n\nWe hope you will gain insight into the Chinese mobile market through looking at our dataset. Since click fraud happens at such a large volume, the problem you are solving is very important to us. We are looking forward to your innovative solutions. Thank you for participating and good luck! 祝你好运！\n\nBest Wishes,\n\nThe TalkingData Team",
      "votes": null
    },
    {
      "id": "295626",
      "postDate": "03/13/2018 23:51:11",
      "content": "<p>Hi Jack, thank you for hosting this competition, I have a question, the end goal is to detect fraudulent clicks, but the goal of the competition seems to predicting whether an app will be downloaded or not. How does it differentiate between fraudulent clickers v.s people who are just browsing apps but don't end up downloading ?</p>",
      "rawMarkdown": "Hi Jack, thank you for hosting this competition, I have a question, the end goal is to detect fraudulent clicks, but the goal of the competition seems to predicting whether an app will be downloaded or not. How does it differentiate between fraudulent clickers v.s people who are just browsing apps but don't end up downloading ?",
      "votes": null
    },
    {
      "id": "296024",
      "postDate": "03/14/2018 15:55:25",
      "content": "<p>I think the only goal iis to predict who will download the app.</p>\n\n<p>The explanation about fraudulent clickers was just to give you some business context.</p>",
      "rawMarkdown": "I think the only goal iis to predict who will download the app.\n\nThe explanation about fraudulent clickers was just to give you some business context.",
      "votes": null
    },
    {
      "id": "302498",
      "postDate": "03/24/2018 07:11:19",
      "content": "<p>Hi! I am new to Kaggle and data science. I have read the description carefully but didn't get it. The submission should contain 0s and 1s or the probabilities of a download after clicking.</p>\n\n<p>谢谢！</p>",
      "rawMarkdown": "Hi! I am new to Kaggle and data science. I have read the description carefully but didn't get it. The submission should contain 0s and 1s or the probabilities of a download after clicking.\n\n谢谢！",
      "votes": null
    },
    {
      "id": "302718",
      "postDate": "03/24/2018 16:23:56",
      "content": "<p>It should contain probabilities. (But note that the evaluation criterion depends only on order, so any order-preserving transformation of the probabilities will not affect the score.)</p>",
      "rawMarkdown": "It should contain probabilities. (But note that the evaluation criterion depends only on order, so any order-preserving transformation of the probabilities will not affect the score.)",
      "votes": null
    },
    {
      "id": "310321",
      "postDate": "04/07/2018 05:17:16",
      "content": "<p>i am working on this dataset but i m not able to read my file in jupyter notebook. Is there any platform  or environment  where i can read and do EDA on this dataset. Also help me working on encoding the variables .</p>",
      "rawMarkdown": "i am working on this dataset but i m not able to read my file in jupyter notebook. Is there any platform  or environment  where i can read and do EDA on this dataset. Also help me working on encoding the variables .",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 295626,
      "author_name": "rteja1113",
      "author_url": "",
      "post_date": "03/13/2018 23:51:11",
      "content": "<p>Hi Jack, thank you for hosting this competition, I have a question, the end goal is to detect fraudulent clicks, but the goal of the competition seems to predicting whether an app will be downloaded or not. How does it differentiate between fraudulent clickers v.s people who are just browsing apps but don't end up downloading ?</p>",
      "votes": null,
      "replies": [
        {
          "id": 296024,
          "author_name": "enricospada",
          "author_url": "",
          "post_date": "03/14/2018 15:55:25",
          "content": "<p>I think the only goal iis to predict who will download the app.</p>\n\n<p>The explanation about fraudulent clickers was just to give you some business context.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 302498,
      "author_name": "jitengma",
      "author_url": "",
      "post_date": "03/24/2018 07:11:19",
      "content": "<p>Hi! I am new to Kaggle and data science. I have read the description carefully but didn't get it. The submission should contain 0s and 1s or the probabilities of a download after clicking.</p>\n\n<p>谢谢！</p>",
      "votes": null,
      "replies": [
        {
          "id": 302718,
          "author_name": "aharless",
          "author_url": "",
          "post_date": "03/24/2018 16:23:56",
          "content": "<p>It should contain probabilities. (But note that the evaluation criterion depends only on order, so any order-preserving transformation of the probabilities will not affect the score.)</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 310321,
      "author_name": "saurabhkamble111",
      "author_url": "",
      "post_date": "04/07/2018 05:17:16",
      "content": "<p>i am working on this dataset but i m not able to read my file in jupyter notebook. Is there any platform  or environment  where i can read and do EDA on this dataset. Also help me working on encoding the variables .</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "291473": "Hi Kagglers,\n\nWelcome to the TalkingData AdTracking Fraud Detection Challenge! We are excited about our second competition with Kaggle, and we hope that you will enjoy working on this problem. Our data science team will be on this forum throughout the competition to answer your questions.\n\nWe hope you will gain insight into the Chinese mobile market through looking at our dataset. Since click fraud happens at such a large volume, the problem you are solving is very important to us. We are looking forward to your innovative solutions. Thank you for participating and good luck! 祝你好运！\n\nBest Wishes,\n\nThe TalkingData Team",
    "295626": "Hi Jack, thank you for hosting this competition, I have a question, the end goal is to detect fraudulent clicks, but the goal of the competition seems to predicting whether an app will be downloaded or not. How does it differentiate between fraudulent clickers v.s people who are just browsing apps but don't end up downloading ?",
    "296024": "I think the only goal iis to predict who will download the app.\n\nThe explanation about fraudulent clickers was just to give you some business context.",
    "302498": "Hi! I am new to Kaggle and data science. I have read the description carefully but didn't get it. The submission should contain 0s and 1s or the probabilities of a download after clicking.\n\n谢谢！",
    "302718": "It should contain probabilities. (But note that the evaluation criterion depends only on order, so any order-preserving transformation of the probabilities will not affect the score.)",
    "310321": "i am working on this dataset but i m not able to read my file in jupyter notebook. Is there any platform  or environment  where i can read and do EDA on this dataset. Also help me working on encoding the variables ."
  },
  "source": "meta"
}