{
  "id": 181522,
  "title": "More explanation needed regarding confidence",
  "url": "/competitions/osic-pulmonary-fibrosis-progression/discussion/181522",
  "author_name": "",
  "post_date": "2020-09-09T07:36:03.052533500Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hi there,</p>\n<p>For a couple of days, I've been so confused with term <strong>confidence</strong>.<br>\nI have checked several other discussion topics and understood it's nothing but <strong>std. dev</strong>.<br>\n<strong>BUT</strong><br>\n<em>What I thought was,</em> <br>\nFirst,<br>\nI predict all the FVC values for each patient in all weeks(-12 to 133).<br>\nThen,<br>\nI calculate <strong>std dev.</strong> on all FVC values for each patient and then submit.</p>\n<p>So is that it?!<br>\nPlease provide your answer in clean and detailed manner.</p>\n<p>Thank you!</p>",
  "messages": [
    {
      "id": "1003711",
      "postDate": "09/09/2020 07:36:03",
      "content": "<p>Hi there,</p>\n<p>For a couple of days, I've been so confused with term <strong>confidence</strong>.<br>\nI have checked several other discussion topics and understood it's nothing but <strong>std. dev</strong>.<br>\n<strong>BUT</strong><br>\n<em>What I thought was,</em> <br>\nFirst,<br>\nI predict all the FVC values for each patient in all weeks(-12 to 133).<br>\nThen,<br>\nI calculate <strong>std dev.</strong> on all FVC values for each patient and then submit.</p>\n<p>So is that it?!<br>\nPlease provide your answer in clean and detailed manner.</p>\n<p>Thank you!</p>",
      "rawMarkdown": "Hi there,\n\nFor a couple of days, I've been so confused with term **confidence**.\nI have checked several other discussion topics and understood it's nothing but **std. dev**.\n**BUT**\n*What I thought was,* \nFirst,\nI predict all the FVC values for each patient in all weeks(-12 to 133).\nThen,\nI calculate **std dev.** on all FVC values for each patient and then submit.\n\nSo is that it?!\nPlease provide your answer in clean and detailed manner.\n\nThank you!",
      "votes": null
    },
    {
      "id": "1004252",
      "postDate": "09/09/2020 15:01:44",
      "content": "<p>No, it's not. The confidence is a measure of how precise is each prediction. The smallest it is, the more precise you are. <br>\n Each prediction should have a different confidence. There is no official way to predict it but the more classic way is to predict quantiles (one smaller and one bigger than your prediction). Then you can make the difference between the bigger and the smaller. The less difference there is between those two values, the more confident you are on your prediction. Take a look at the this <a href=\"https://www.kaggle.com/ulrich07/osic-multiple-quantile-regression-starter\" target=\"_blank\">notebook</a>. It's an implementation of multiple quantiles prediction.</p>",
      "rawMarkdown": "No, it's not. The confidence is a measure of how precise is each prediction. The smallest it is, the more precise you are. \n Each prediction should have a different confidence. There is no official way to predict it but the more classic way is to predict quantiles (one smaller and one bigger than your prediction). Then you can make the difference between the bigger and the smaller. The less difference there is between those two values, the more confident you are on your prediction. Take a look at the this [notebook](https://www.kaggle.com/ulrich07/osic-multiple-quantile-regression-starter). It's an implementation of multiple quantiles prediction.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1004252,
      "author_name": "yohannwattiez",
      "author_url": "",
      "post_date": "09/09/2020 15:01:44",
      "content": "<p>No, it's not. The confidence is a measure of how precise is each prediction. The smallest it is, the more precise you are. <br>\n Each prediction should have a different confidence. There is no official way to predict it but the more classic way is to predict quantiles (one smaller and one bigger than your prediction). Then you can make the difference between the bigger and the smaller. The less difference there is between those two values, the more confident you are on your prediction. Take a look at the this <a href=\"https://www.kaggle.com/ulrich07/osic-multiple-quantile-regression-starter\" target=\"_blank\">notebook</a>. It's an implementation of multiple quantiles prediction.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1003711": "Hi there,\n\nFor a couple of days, I've been so confused with term **confidence**.\nI have checked several other discussion topics and understood it's nothing but **std. dev**.\n**BUT**\n*What I thought was,* \nFirst,\nI predict all the FVC values for each patient in all weeks(-12 to 133).\nThen,\nI calculate **std dev.** on all FVC values for each patient and then submit.\n\nSo is that it?!\nPlease provide your answer in clean and detailed manner.\n\nThank you!",
    "1004252": "No, it's not. The confidence is a measure of how precise is each prediction. The smallest it is, the more precise you are. \n Each prediction should have a different confidence. There is no official way to predict it but the more classic way is to predict quantiles (one smaller and one bigger than your prediction). Then you can make the difference between the bigger and the smaller. The less difference there is between those two values, the more confident you are on your prediction. Take a look at the this [notebook](https://www.kaggle.com/ulrich07/osic-multiple-quantile-regression-starter). It's an implementation of multiple quantiles prediction."
  },
  "source": "meta"
}