{
  "id": 78478,
  "title": "For Andrew Lukyanenko",
  "url": "/competitions/LANL-Earthquake-Prediction/discussion/78478",
  "author_name": "",
  "post_date": "2019-01-24T10:01:37.865301100Z",
  "votes": 22,
  "comment_count": 12,
  "views": 0,
  "content": "<p>Having looked at your excellent kernel I used X_train_scaled and X_test_scaled to evolve a function using Genetic Programming.  Hope you like it!</p>\n\n<p>You did all the work on data munging so I thought it best to leave the code here rather than forking your script as you deserve the votes!</p>\n\n<p><img src=\"https://storage.googleapis.com/kaggle-forum-message-attachments/460720/11100/gpfeatures.png\" alt=\"Freatures\"></p>",
  "messages": [
    {
      "id": "460720",
      "postDate": "01/24/2019 10:01:37",
      "content": "<p>Having looked at your excellent kernel I used X_train_scaled and X_test_scaled to evolve a function using Genetic Programming.  Hope you like it!</p>\n\n<p>You did all the work on data munging so I thought it best to leave the code here rather than forking your script as you deserve the votes!</p>\n\n<p><img src=\"https://storage.googleapis.com/kaggle-forum-message-attachments/460720/11100/gpfeatures.png\" alt=\"Freatures\"></p>",
      "rawMarkdown": "Having looked at your excellent kernel I used X_train_scaled and X_test_scaled to evolve a function using Genetic Programming.  Hope you like it!\n\nYou did all the work on data munging so I thought it best to leave the code here rather than forking your script as you deserve the votes!\n\n![Freatures][1]\n\n\n  [1]: https://storage.googleapis.com/kaggle-forum-message-attachments/460720/11100/gpfeatures.png",
      "votes": null
    },
    {
      "id": "460787",
      "postDate": "01/24/2019 12:19:35",
      "content": "<p>Hi Scirpus, how did you obtain that function? </p>",
      "rawMarkdown": "Hi Scirpus, how did you obtain that function?",
      "votes": null
    },
    {
      "id": "460788",
      "postDate": "01/24/2019 12:21:45",
      "content": "<p>Imported the data pressed play on my own Genetic Programming Tool - took 20 minutes. (Edit the GP stuff took 5 minutes the data munging took 15 minutes)</p>",
      "rawMarkdown": "Imported the data pressed play on my own Genetic Programming Tool - took 20 minutes. (Edit the GP stuff took 5 minutes the data munging took 15 minutes)",
      "votes": null
    },
    {
      "id": "461170",
      "postDate": "01/25/2019 12:49:55",
      "content": "<p>Hi Sirpus, so intresting! Do you have any example how build GP algorithm for time series? May be kernel or link? </p>",
      "rawMarkdown": "Hi Sirpus, so intresting! Do you have any example how build GP algorithm for time series? May be kernel or link?",
      "votes": null
    },
    {
      "id": "461173",
      "postDate": "01/25/2019 12:55:40",
      "content": "<p>Sorry, I saw this topic only now!</p>\n\n<p>This is cool and I really appreciate you doing this :)</p>",
      "rawMarkdown": "Sorry, I saw this topic only now!\n\nThis is cool and I really appreciate you doing this :)",
      "votes": null
    },
    {
      "id": "461174",
      "postDate": "01/25/2019 12:59:45",
      "content": "<p>Hi,</p>\n\n<p>Do you mean like an RNN type of things using a windowing technique on the raw data?</p>",
      "rawMarkdown": "Hi,\n\nDo you mean like an RNN type of things using a windowing technique on the raw data?",
      "votes": null
    },
    {
      "id": "461359",
      "postDate": "01/25/2019 20:39:59",
      "content": "<p>I'm playing with the same baseline model, but your achievement is amazing. Could you explain what parameter named 'roll' is?</p>",
      "rawMarkdown": "I'm playing with the same baseline model, but your achievement is amazing. Could you explain what parameter named 'roll' is?",
      "votes": null
    },
    {
      "id": "461523",
      "postDate": "01/26/2019 08:50:03",
      "content": "<p><a href=\"https://www.kaggle.com/artgor/earthquakes-fe-more-features-and-samples\">https://www.kaggle.com/artgor/earthquakes-fe-more-features-and-samples</a></p>",
      "rawMarkdown": "https://www.kaggle.com/artgor/earthquakes-fe-more-features-and-samples",
      "votes": null
    },
    {
      "id": "461592",
      "postDate": "01/26/2019 13:07:55",
      "content": "<p>Not entirely. You wrote that you used your Genetic Programming Tool. What do you mean, \"Genetic Programming Tool\"? It has something to do with <a href=\"https://en.wikipedia.org/wiki/Genetic_programming\">https://en.wikipedia.org/wiki/Genetic_programming</a> ?</p>",
      "rawMarkdown": "Not entirely. You wrote that you used your Genetic Programming Tool. What do you mean, \"Genetic Programming Tool\"? It has something to do with https://en.wikipedia.org/wiki/Genetic_programming ?",
      "votes": null
    },
    {
      "id": "461606",
      "postDate": "01/26/2019 13:53:50",
      "content": "<p>Yes it uses darwinian evolution - a population of random functions are evaluated and the best get to breed.  The children produced are a mix of their parents DNA . The steps are repeated and after a 1000 generations you get the output you see in the file for the best individual. ;)  The tool I wrote myself</p>",
      "rawMarkdown": "Yes it uses darwinian evolution - a population of random functions are evaluated and the best get to breed.  The children produced are a mix of their parents DNA . The steps are repeated and after a 1000 generations you get the output you see in the file for the best individual. ;)  The tool I wrote myself",
      "votes": null
    },
    {
      "id": "463058",
      "postDate": "01/29/2019 10:18:06",
      "content": "<p>Hi Scirpus, super interesting!</p>\n\n<p>I’ve seen you post some of your gp stuff before and I find it very fascinating. I’ve found an enormous amount of literature about genetic programming that will take a while to make sense of, can you recommend any papers or sources you find to be particularly good? Thanks, and keep up the good work!</p>",
      "rawMarkdown": "Hi Scirpus, super interesting!\n\nI’ve seen you post some of your gp stuff before and I find it very fascinating. I’ve found an enormous amount of literature about genetic programming that will take a while to make sense of, can you recommend any papers or sources you find to be particularly good? Thanks, and keep up the good work!",
      "votes": null
    },
    {
      "id": "463172",
      "postDate": "01/29/2019 15:13:24",
      "content": "<p>Start by looking at the kernel I produced which at least gives you an output and has context on what the data columns actually represent.</p>\n\n<p>Have a look for John Koza on google as his literature got me into it.<a href=\"http://geneticprogramming.com/\">http://geneticprogramming.com/</a> is a good start too!</p>",
      "rawMarkdown": "Start by looking at the kernel I produced which at least gives you an output and has context on what the data columns actually represent.\n\nHave a look for John Koza on google as his literature got me into it.http://geneticprogramming.com/ is a good start too!",
      "votes": null
    },
    {
      "id": "463291",
      "postDate": "01/29/2019 18:47:06",
      "content": "<p>Awesome, thanks a lot!</p>\n\n<p>I'm very much enjoying learning about John Koza and his work. From his wikipedia page it says he invented scratch-off lottery tickets and was the second person ever to get a bachelors degree in computer science. Badass!</p>",
      "rawMarkdown": "Awesome, thanks a lot!\n\nI'm very much enjoying learning about John Koza and his work. From his wikipedia page it says he invented scratch-off lottery tickets and was the second person ever to get a bachelors degree in computer science. Badass!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 460787,
      "author_name": "khahuras",
      "author_url": "",
      "post_date": "01/24/2019 12:19:35",
      "content": "<p>Hi Scirpus, how did you obtain that function? </p>",
      "votes": null,
      "replies": [
        {
          "id": 460788,
          "author_name": "scirpus",
          "author_url": "",
          "post_date": "01/24/2019 12:21:45",
          "content": "<p>Imported the data pressed play on my own Genetic Programming Tool - took 20 minutes. (Edit the GP stuff took 5 minutes the data munging took 15 minutes)</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 461170,
      "author_name": "nikitagribov",
      "author_url": "",
      "post_date": "01/25/2019 12:49:55",
      "content": "<p>Hi Sirpus, so intresting! Do you have any example how build GP algorithm for time series? May be kernel or link? </p>",
      "votes": null,
      "replies": [
        {
          "id": 461174,
          "author_name": "scirpus",
          "author_url": "",
          "post_date": "01/25/2019 12:59:45",
          "content": "<p>Hi,</p>\n\n<p>Do you mean like an RNN type of things using a windowing technique on the raw data?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 461592,
          "author_name": "nikitagribov",
          "author_url": "",
          "post_date": "01/26/2019 13:07:55",
          "content": "<p>Not entirely. You wrote that you used your Genetic Programming Tool. What do you mean, \"Genetic Programming Tool\"? It has something to do with <a href=\"https://en.wikipedia.org/wiki/Genetic_programming\">https://en.wikipedia.org/wiki/Genetic_programming</a> ?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 461606,
          "author_name": "scirpus",
          "author_url": "",
          "post_date": "01/26/2019 13:53:50",
          "content": "<p>Yes it uses darwinian evolution - a population of random functions are evaluated and the best get to breed.  The children produced are a mix of their parents DNA . The steps are repeated and after a 1000 generations you get the output you see in the file for the best individual. ;)  The tool I wrote myself</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 461173,
      "author_name": "artgor",
      "author_url": "",
      "post_date": "01/25/2019 12:55:40",
      "content": "<p>Sorry, I saw this topic only now!</p>\n\n<p>This is cool and I really appreciate you doing this :)</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 461359,
      "author_name": "jhgeeyang",
      "author_url": "",
      "post_date": "01/25/2019 20:39:59",
      "content": "<p>I'm playing with the same baseline model, but your achievement is amazing. Could you explain what parameter named 'roll' is?</p>",
      "votes": null,
      "replies": [
        {
          "id": 461523,
          "author_name": "scirpus",
          "author_url": "",
          "post_date": "01/26/2019 08:50:03",
          "content": "<p><a href=\"https://www.kaggle.com/artgor/earthquakes-fe-more-features-and-samples\">https://www.kaggle.com/artgor/earthquakes-fe-more-features-and-samples</a></p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 463058,
      "author_name": "interneuron",
      "author_url": "",
      "post_date": "01/29/2019 10:18:06",
      "content": "<p>Hi Scirpus, super interesting!</p>\n\n<p>I’ve seen you post some of your gp stuff before and I find it very fascinating. I’ve found an enormous amount of literature about genetic programming that will take a while to make sense of, can you recommend any papers or sources you find to be particularly good? Thanks, and keep up the good work!</p>",
      "votes": null,
      "replies": [
        {
          "id": 463172,
          "author_name": "scirpus",
          "author_url": "",
          "post_date": "01/29/2019 15:13:24",
          "content": "<p>Start by looking at the kernel I produced which at least gives you an output and has context on what the data columns actually represent.</p>\n\n<p>Have a look for John Koza on google as his literature got me into it.<a href=\"http://geneticprogramming.com/\">http://geneticprogramming.com/</a> is a good start too!</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 463291,
          "author_name": "interneuron",
          "author_url": "",
          "post_date": "01/29/2019 18:47:06",
          "content": "<p>Awesome, thanks a lot!</p>\n\n<p>I'm very much enjoying learning about John Koza and his work. From his wikipedia page it says he invented scratch-off lottery tickets and was the second person ever to get a bachelors degree in computer science. Badass!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "460720": "Having looked at your excellent kernel I used X_train_scaled and X_test_scaled to evolve a function using Genetic Programming.  Hope you like it!\n\nYou did all the work on data munging so I thought it best to leave the code here rather than forking your script as you deserve the votes!\n\n![Freatures][1]\n\n\n  [1]: https://storage.googleapis.com/kaggle-forum-message-attachments/460720/11100/gpfeatures.png",
    "460787": "Hi Scirpus, how did you obtain that function?",
    "460788": "Imported the data pressed play on my own Genetic Programming Tool - took 20 minutes. (Edit the GP stuff took 5 minutes the data munging took 15 minutes)",
    "461170": "Hi Sirpus, so intresting! Do you have any example how build GP algorithm for time series? May be kernel or link?",
    "461173": "Sorry, I saw this topic only now!\n\nThis is cool and I really appreciate you doing this :)",
    "461174": "Hi,\n\nDo you mean like an RNN type of things using a windowing technique on the raw data?",
    "461359": "I'm playing with the same baseline model, but your achievement is amazing. Could you explain what parameter named 'roll' is?",
    "461523": "https://www.kaggle.com/artgor/earthquakes-fe-more-features-and-samples",
    "461592": "Not entirely. You wrote that you used your Genetic Programming Tool. What do you mean, \"Genetic Programming Tool\"? It has something to do with https://en.wikipedia.org/wiki/Genetic_programming ?",
    "461606": "Yes it uses darwinian evolution - a population of random functions are evaluated and the best get to breed.  The children produced are a mix of their parents DNA . The steps are repeated and after a 1000 generations you get the output you see in the file for the best individual. ;)  The tool I wrote myself",
    "463058": "Hi Scirpus, super interesting!\n\nI’ve seen you post some of your gp stuff before and I find it very fascinating. I’ve found an enormous amount of literature about genetic programming that will take a while to make sense of, can you recommend any papers or sources you find to be particularly good? Thanks, and keep up the good work!",
    "463172": "Start by looking at the kernel I produced which at least gives you an output and has context on what the data columns actually represent.\n\nHave a look for John Koza on google as his literature got me into it.http://geneticprogramming.com/ is a good start too!",
    "463291": "Awesome, thanks a lot!\n\nI'm very much enjoying learning about John Koza and his work. From his wikipedia page it says he invented scratch-off lottery tickets and was the second person ever to get a bachelors degree in computer science. Badass!"
  },
  "source": "meta"
}