{
  "id": 326339,
  "title": "Intuition behind creating the air_temperature_mean_lag73 and other lag fields?",
  "url": "/competitions/energy-anomaly-detection/discussion/326339",
  "author_name": "",
  "post_date": "2022-05-21T17:09:50.602607300Z",
  "votes": 4,
  "comment_count": 2,
  "views": 0,
  "content": "<p>There are multiple fields derived from air_temperature that are basically different lag operations on top of the field, air_temperature_mean_lag73 being one of them. Intuitively it makes sense that recent temperature values can have an impact on generating an anomalous reading. What I fail to understand is how did we arrive at a conclusion that air_temprature lag should be taken and defined as a min, max,std of varying range? What was the EDA that went into the same?</p>",
  "messages": [
    {
      "id": "1797193",
      "postDate": "05/21/2022 17:09:50",
      "content": "<p>There are multiple fields derived from air_temperature that are basically different lag operations on top of the field, air_temperature_mean_lag73 being one of them. Intuitively it makes sense that recent temperature values can have an impact on generating an anomalous reading. What I fail to understand is how did we arrive at a conclusion that air_temprature lag should be taken and defined as a min, max,std of varying range? What was the EDA that went into the same?</p>",
      "rawMarkdown": "There are multiple fields derived from air_temperature that are basically different lag operations on top of the field, air_temperature_mean_lag73 being one of them. Intuitively it makes sense that recent temperature values can have an impact on generating an anomalous reading. What I fail to understand is how did we arrive at a conclusion that air_temprature lag should be taken and defined as a min, max,std of varying range? What was the EDA that went into the same?",
      "votes": null
    },
    {
      "id": "1871335",
      "postDate": "07/26/2022 08:08:52",
      "content": "<p>you should probably check EDA nbs from previous competition:  ASHRAE - Great Energy Predictor III competition </p>\n<p>I think all features we are given here are generated there </p>",
      "rawMarkdown": "you should probably check EDA nbs from previous competition:  ASHRAE - Great Energy Predictor III competition \n\nI think all features we are given here are generated there",
      "votes": null
    },
    {
      "id": "1959312",
      "postDate": "09/28/2022 04:49:52",
      "content": "<p>Sorry for the late reply, I did check this but couldn't find the description of this specific feature, neither could I infer how exactly this feature was being calculated based on the code they provided on github.</p>",
      "rawMarkdown": "Sorry for the late reply, I did check this but couldn't find the description of this specific feature, neither could I infer how exactly this feature was being calculated based on the code they provided on github.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1871335,
      "author_name": "imeintanis",
      "author_url": "",
      "post_date": "07/26/2022 08:08:52",
      "content": "<p>you should probably check EDA nbs from previous competition:  ASHRAE - Great Energy Predictor III competition </p>\n<p>I think all features we are given here are generated there </p>",
      "votes": null,
      "replies": [
        {
          "id": 1959312,
          "author_name": "rohannanaware",
          "author_url": "",
          "post_date": "09/28/2022 04:49:52",
          "content": "<p>Sorry for the late reply, I did check this but couldn't find the description of this specific feature, neither could I infer how exactly this feature was being calculated based on the code they provided on github.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1797193": "There are multiple fields derived from air_temperature that are basically different lag operations on top of the field, air_temperature_mean_lag73 being one of them. Intuitively it makes sense that recent temperature values can have an impact on generating an anomalous reading. What I fail to understand is how did we arrive at a conclusion that air_temprature lag should be taken and defined as a min, max,std of varying range? What was the EDA that went into the same?",
    "1871335": "you should probably check EDA nbs from previous competition:  ASHRAE - Great Energy Predictor III competition \n\nI think all features we are given here are generated there",
    "1959312": "Sorry for the late reply, I did check this but couldn't find the description of this specific feature, neither could I infer how exactly this feature was being calculated based on the code they provided on github."
  },
  "source": "meta"
}