{
  "id": 82040,
  "title": "Congratulation to everyone!!!",
  "url": "/competitions/elo-merchant-category-recommendation/writeups/happy-pepper-pig-year-congratulation-to-everyone",
  "author_name": "",
  "post_date": "2019-02-27T06:29:02.573Z",
  "votes": 30,
  "comment_count": 27,
  "views": 0,
  "content": "<p>Hello all, congratulations to the winner, gold medalists,, and especially to those who survived the shake-up :)</p>\n\n<p>I got busy with my work and had to stay away for 4 weeks without doing anything, now looking back that may have prevented me from overfitting my model further!</p>\n\n<p>The shake-up seems to be quite strong, and I am not entirely sure what worked in my solution and what didn't.  I am pretty sure the following were helpful:</p>\n\n<ul>\n<li>feature selection by the mean of feature permutation </li>\n<li>a significant amount of pseudo labelling, even concatenating multiple copies of test data and predict label. one of my best single models was 3 x pseudo labels with public 3.668 private 3.600, but sadly I didn't select it </li>\n<li>a little moderated post-processing for prediction with more than 0.50 probability seems to help - I used a formula of binary_probabilityX1.1XOUTLIER_VALUE, which helped to improve my best final submission</li>\n</ul>\n\n<p>anyway, I have learnt a lot throughout this competition -- even though I didn't have much as fun as I would like to - competing by yourself is hard, turning down team-up invitation from friends is HARD, getting that solo gold is VERY HARD!</p>\n\n<p>See you all around the LB!!!!</p>",
  "messages": [
    {
      "id": "479066",
      "postDate": "02/27/2019 00:35:17",
      "content": "<p>Hello all, congratulations to the winner, gold medalists,, and especially to those who survived the shake-up :)</p>\n\n<p>I got busy with my work and had to stay away for 4 weeks without doing anything, now looking back that may have prevented me from overfitting my model further!</p>\n\n<p>The shake-up seems to be quite strong, and I am not entirely sure what worked in my solution and what didn't.  I am pretty sure the following were helpful:</p>\n\n<ul>\n<li>feature selection by the mean of feature permutation </li>\n<li>a significant amount of pseudo labelling, even concatenating multiple copies of test data and predict label. one of my best single models was 3 x pseudo labels with public 3.668 private 3.600, but sadly I didn't select it </li>\n<li>a little moderated post-processing for prediction with more than 0.50 probability seems to help - I used a formula of binary_probabilityX1.1XOUTLIER_VALUE, which helped to improve my best final submission</li>\n</ul>\n\n<p>anyway, I have learnt a lot throughout this competition -- even though I didn't have much as fun as I would like to - competing by yourself is hard, turning down team-up invitation from friends is HARD, getting that solo gold is VERY HARD!</p>\n\n<p>See you all around the LB!!!!</p>",
      "rawMarkdown": "Hello all, congratulations to the winner, gold medalists,, and especially to those who survived the shake-up :)\n\nI got busy with my work and had to stay away for 4 weeks without doing anything, now looking back that may have prevented me from overfitting my model further!\n\nThe shake-up seems to be quite strong, and I am not entirely sure what worked in my solution and what didn't.  I am pretty sure the following were helpful:\n\n- feature selection by the mean of feature permutation \n- a significant amount of pseudo labelling, even concatenating multiple copies of test data and predict label. one of my best single models was 3 x pseudo labels with public 3.668 private 3.600, but sadly I didn't select it \n- a little moderated post-processing for prediction with more than 0.50 probability seems to help - I used a formula of binary_probabilityX1.1XOUTLIER_VALUE, which helped to improve my best final submission\n\nanyway, I have learnt a lot throughout this competition -- even though I didn't have much as fun as I would like to - competing by yourself is hard, turning down team-up invitation from friends is HARD, getting that solo gold is VERY HARD!\n\nSee you all around the LB!!!!",
      "votes": null
    },
    {
      "id": "479083",
      "postDate": "02/27/2019 00:50:51",
      "content": "<p>Congrats for your solid solo result.  Shakeup did not move you much, did it?</p>",
      "rawMarkdown": "Congrats for your solid solo result.  Shakeup did not move you much, did it?",
      "votes": null
    },
    {
      "id": "479085",
      "postDate": "02/27/2019 00:51:01",
      "content": "<p>Congrats..!!</p>",
      "rawMarkdown": "Congrats..!!",
      "votes": null
    },
    {
      "id": "479086",
      "postDate": "02/27/2019 00:51:45",
      "content": "<p>Congrats!</p>\n\n<p>How exactly does this <code>binaryprobability1.1OUTLIERVALUE</code> formula work? I don't quite get it. </p>",
      "rawMarkdown": "Congrats!\n\nHow exactly does this `binaryprobability1.1OUTLIERVALUE` formula work? I don't quite get it.",
      "votes": null
    },
    {
      "id": "479089",
      "postDate": "02/27/2019 00:53:25",
      "content": "<p>Congrats! Your post help me a lot.\nThanks!</p>",
      "rawMarkdown": "Congrats! Your post help me a lot.\nThanks!",
      "votes": null
    },
    {
      "id": "479092",
      "postDate": "02/27/2019 00:55:58",
      "content": "<p>sorry, it was meant to be a multiplication formula, I put X in now.</p>",
      "rawMarkdown": "sorry, it was meant to be a multiplication formula, I put X in now.",
      "votes": null
    },
    {
      "id": "479094",
      "postDate": "02/27/2019 00:56:35",
      "content": "<p>not being able to compete for the last month definitely helped, although I honestly ran out of idea by that time :)</p>",
      "rawMarkdown": "not being able to compete for the last month definitely helped, although I honestly ran out of idea by that time :)",
      "votes": null
    },
    {
      "id": "479097",
      "postDate": "02/27/2019 00:57:35",
      "content": "<blockquote>\n  <p>I honestly ran out of idea </p>\n</blockquote>\n\n<p>That's when teaming makes sense ;)</p>",
      "rawMarkdown": "&gt;  I honestly ran out of idea \n\nThat's when teaming makes sense ;)",
      "votes": null
    },
    {
      "id": "479098",
      "postDate": "02/27/2019 01:00:06",
      "content": "<p><code>OUTLIERVALUE</code> is the regression value from a model trained without excluding outliers? </p>\n\n<p>Geez, I really suck at interpreting variable names.</p>",
      "rawMarkdown": "`OUTLIERVALUE` is the regression value from a model trained without excluding outliers? \n\nGeez, I really suck at interpreting variable names.",
      "votes": null
    },
    {
      "id": "479103",
      "postDate": "02/27/2019 01:07:24",
      "content": "<p>yup</p>",
      "rawMarkdown": "yup",
      "votes": null
    },
    {
      "id": "479106",
      "postDate": "02/27/2019 01:08:31",
      "content": "<p>all right, all right, I get the message :)</p>",
      "rawMarkdown": "all right, all right, I get the message :)",
      "votes": null
    },
    {
      "id": "479112",
      "postDate": "02/27/2019 01:14:03",
      "content": "<p>glad it helped, and CONGRATULATION for the prize-winning performance! </p>",
      "rawMarkdown": "glad it helped, and CONGRATULATION for the prize-winning performance!",
      "votes": null
    },
    {
      "id": "479116",
      "postDate": "02/27/2019 01:17:25",
      "content": "<p>Thanks for your sharing in this game, you should get gold medal. Thanks again.</p>",
      "rawMarkdown": "Thanks for your sharing in this game, you should get gold medal. Thanks again.",
      "votes": null
    },
    {
      "id": "479126",
      "postDate": "02/27/2019 01:24:05",
      "content": "<p>Congratulations Yifan. Can you explain a bit more about your feature engineering process ? Just used simple aggregation or other strategies ?</p>",
      "rawMarkdown": "Congratulations Yifan. Can you explain a bit more about your feature engineering process ? Just used simple aggregation or other strategies ?",
      "votes": null
    },
    {
      "id": "479150",
      "postDate": "02/27/2019 01:45:16",
      "content": "<p>Congratulations to you @Yifan Xie\nFor your advice:</p>\n\n<ul>\n<li>feature selection by the mean of feature permutation</li>\n</ul>\n\n<p>I want to know is it 'feature permutation' or 'target permutation'? there seem's have two methods, I want to know which is useful for this competition, thanks!</p>",
      "rawMarkdown": "Congratulations to you @Yifan Xie\nFor your advice:\n\n- feature selection by the mean of feature permutation\n\nI want to know is it 'feature permutation' or 'target permutation'? there seem's have two methods, I want to know which is useful for this competition, thanks!",
      "votes": null
    },
    {
      "id": "479161",
      "postDate": "02/27/2019 01:57:36",
      "content": "<p>Congrats Yifan (laoshi)! Your \"Sharing of my experience so far\" post inspired me and helped me to set a target line (CV 3.62, LB 3.67). I keep moving forward with feature engineering + feature selection and finally it works. I wonder if you will share more about your feature engineering stuff when you have time ;)? Thank you!</p>",
      "rawMarkdown": "Congrats Yifan (laoshi)! Your \"Sharing of my experience so far\" post inspired me and helped me to set a target line (CV 3.62, LB 3.67). I keep moving forward with feature engineering + feature selection and finally it works. I wonder if you will share more about your feature engineering stuff when you have time ;)? Thank you!",
      "votes": null
    },
    {
      "id": "479184",
      "postDate": "02/27/2019 02:20:07",
      "content": "<p>Thanks for your post! </p>",
      "rawMarkdown": "Thanks for your post!",
      "votes": null
    },
    {
      "id": "479200",
      "postDate": "02/27/2019 02:34:56",
      "content": "<p>Congratulations <a href=\"/yifanxie\">@yifanxie</a>, well deserved. </p>\n\n<blockquote>\n  <p>feature selection by the mean of feature permutation</p>\n</blockquote>\n\n<p>I agree, here is my post on feature selection techniques <a href=\"https://www.kaggle.com/c/elo-merchant-category-recommendation/discussion/82052\">https://www.kaggle.com/c/elo-merchant-category-recommendation/discussion/82052</a></p>",
      "rawMarkdown": "Congratulations @yifanxie, well deserved. \n\n&gt; feature selection by the mean of feature permutation\n\nI agree, here is my post on feature selection techniques https://www.kaggle.com/c/elo-merchant-category-recommendation/discussion/82052",
      "votes": null
    },
    {
      "id": "479297",
      "postDate": "02/27/2019 03:50:58",
      "content": "<p>so close to gold,I believe next time you can win.</p>",
      "rawMarkdown": "so close to gold,I believe next time you can win.",
      "votes": null
    },
    {
      "id": "479364",
      "postDate": "02/27/2019 05:05:53",
      "content": "<p>Congratulations <a href=\"/yifanxie\">@yifanxie</a>. Could you please ellaborate more on the process of <code>pseudo labelling</code>?</p>",
      "rawMarkdown": "Congratulations @yifanxie. Could you please ellaborate more on the process of `pseudo labelling`?",
      "votes": null
    },
    {
      "id": "479368",
      "postDate": "02/27/2019 05:07:51",
      "content": "<p>Congrats to you !!\nGuess the post-processing for prediction with more than 0.50 may have done the trick for you. I was unable to understand and if you are ok to share more details if possible a code sample would be helpful, thanks in advance..</p>",
      "rawMarkdown": "Congrats to you !!\nGuess the post-processing for prediction with more than 0.50 may have done the trick for you. I was unable to understand and if you are ok to share more details if possible a code sample would be helpful, thanks in advance..",
      "votes": null
    },
    {
      "id": "479483",
      "postDate": "02/27/2019 07:25:15",
      "content": "<p>Congratulations  @YifanXie . </p>",
      "rawMarkdown": "Congratulations  @YifanXie .",
      "votes": null
    },
    {
      "id": "479653",
      "postDate": "02/27/2019 10:12:19",
      "content": "<p>Congratulations!</p>",
      "rawMarkdown": "Congratulations!",
      "votes": null
    },
    {
      "id": "479722",
      "postDate": "02/27/2019 11:13:05",
      "content": "<p>congrats Yifan! </p>",
      "rawMarkdown": "congrats Yifan!",
      "votes": null
    },
    {
      "id": "479815",
      "postDate": "02/27/2019 12:54:59",
      "content": "<p>Congratulations,I,m your fans</p>",
      "rawMarkdown": "Congratulations,I,m your fans",
      "votes": null
    },
    {
      "id": "479957",
      "postDate": "02/27/2019 16:04:32",
      "content": "<p>Well done, and thanks for all the valuable discussion during the comp!</p>",
      "rawMarkdown": "Well done, and thanks for all the valuable discussion during the comp!",
      "votes": null
    },
    {
      "id": "480513",
      "postDate": "02/28/2019 09:18:37",
      "content": "<p>congrats to you, your post give me lot of help and keep me fighting, thx.  I hope and believe you will get a solo gold next time</p>",
      "rawMarkdown": "congrats to you, your post give me lot of help and keep me fighting, thx.  I hope and believe you will get a solo gold next time",
      "votes": null
    },
    {
      "id": "480604",
      "postDate": "02/28/2019 12:09:42",
      "content": "<p>congrats to all the top rankers on LB.This was an interesting competition.Gave me opportunity to learn a lot from the fellow kagglers . And finally thanks to Elo for organizing this data science competition</p>",
      "rawMarkdown": "congrats to all the top rankers on LB.This was an interesting competition.Gave me opportunity to learn a lot from the fellow kagglers . And finally thanks to Elo for organizing this data science competition",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 479083,
      "author_name": "cpmpml",
      "author_url": "",
      "post_date": "02/27/2019 00:50:51",
      "content": "<p>Congrats for your solid solo result.  Shakeup did not move you much, did it?</p>",
      "votes": null,
      "replies": [
        {
          "id": 479094,
          "author_name": "yifanxie",
          "author_url": "",
          "post_date": "02/27/2019 00:56:35",
          "content": "<p>not being able to compete for the last month definitely helped, although I honestly ran out of idea by that time :)</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 479097,
          "author_name": "cpmpml",
          "author_url": "",
          "post_date": "02/27/2019 00:57:35",
          "content": "<blockquote>\n  <p>I honestly ran out of idea </p>\n</blockquote>\n\n<p>That's when teaming makes sense ;)</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 479106,
          "author_name": "yifanxie",
          "author_url": "",
          "post_date": "02/27/2019 01:08:31",
          "content": "<p>all right, all right, I get the message :)</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 479085,
      "author_name": "sreeedevi",
      "author_url": "",
      "post_date": "02/27/2019 00:51:01",
      "content": "<p>Congrats..!!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 479086,
      "author_name": "ryanzhang",
      "author_url": "",
      "post_date": "02/27/2019 00:51:45",
      "content": "<p>Congrats!</p>\n\n<p>How exactly does this <code>binaryprobability1.1OUTLIERVALUE</code> formula work? I don't quite get it. </p>",
      "votes": null,
      "replies": [
        {
          "id": 479092,
          "author_name": "yifanxie",
          "author_url": "",
          "post_date": "02/27/2019 00:55:58",
          "content": "<p>sorry, it was meant to be a multiplication formula, I put X in now.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 479098,
          "author_name": "ryanzhang",
          "author_url": "",
          "post_date": "02/27/2019 01:00:06",
          "content": "<p><code>OUTLIERVALUE</code> is the regression value from a model trained without excluding outliers? </p>\n\n<p>Geez, I really suck at interpreting variable names.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 479103,
          "author_name": "yifanxie",
          "author_url": "",
          "post_date": "02/27/2019 01:07:24",
          "content": "<p>yup</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 479089,
      "author_name": "daishu",
      "author_url": "",
      "post_date": "02/27/2019 00:53:25",
      "content": "<p>Congrats! Your post help me a lot.\nThanks!</p>",
      "votes": null,
      "replies": [
        {
          "id": 479112,
          "author_name": "yifanxie",
          "author_url": "",
          "post_date": "02/27/2019 01:14:03",
          "content": "<p>glad it helped, and CONGRATULATION for the prize-winning performance! </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 479116,
      "author_name": "longyin2",
      "author_url": "",
      "post_date": "02/27/2019 01:17:25",
      "content": "<p>Thanks for your sharing in this game, you should get gold medal. Thanks again.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 479126,
      "author_name": "recruta42",
      "author_url": "",
      "post_date": "02/27/2019 01:24:05",
      "content": "<p>Congratulations Yifan. Can you explain a bit more about your feature engineering process ? Just used simple aggregation or other strategies ?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 479150,
      "author_name": "jiazhuang",
      "author_url": "",
      "post_date": "02/27/2019 01:45:16",
      "content": "<p>Congratulations to you @Yifan Xie\nFor your advice:</p>\n\n<ul>\n<li>feature selection by the mean of feature permutation</li>\n</ul>\n\n<p>I want to know is it 'feature permutation' or 'target permutation'? there seem's have two methods, I want to know which is useful for this competition, thanks!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 479161,
      "author_name": "bangdasun",
      "author_url": "",
      "post_date": "02/27/2019 01:57:36",
      "content": "<p>Congrats Yifan (laoshi)! Your \"Sharing of my experience so far\" post inspired me and helped me to set a target line (CV 3.62, LB 3.67). I keep moving forward with feature engineering + feature selection and finally it works. I wonder if you will share more about your feature engineering stuff when you have time ;)? Thank you!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 479184,
      "author_name": "xiaojiu1414",
      "author_url": "",
      "post_date": "02/27/2019 02:20:07",
      "content": "<p>Thanks for your post! </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 479200,
      "author_name": "sheriytm",
      "author_url": "",
      "post_date": "02/27/2019 02:34:56",
      "content": "<p>Congratulations <a href=\"/yifanxie\">@yifanxie</a>, well deserved. </p>\n\n<blockquote>\n  <p>feature selection by the mean of feature permutation</p>\n</blockquote>\n\n<p>I agree, here is my post on feature selection techniques <a href=\"https://www.kaggle.com/c/elo-merchant-category-recommendation/discussion/82052\">https://www.kaggle.com/c/elo-merchant-category-recommendation/discussion/82052</a></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 479297,
      "author_name": "senkin13",
      "author_url": "",
      "post_date": "02/27/2019 03:50:58",
      "content": "<p>so close to gold,I believe next time you can win.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 479364,
      "author_name": "liewmanchoi",
      "author_url": "",
      "post_date": "02/27/2019 05:05:53",
      "content": "<p>Congratulations <a href=\"/yifanxie\">@yifanxie</a>. Could you please ellaborate more on the process of <code>pseudo labelling</code>?</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 479368,
      "author_name": "viswanathravindran",
      "author_url": "",
      "post_date": "02/27/2019 05:07:51",
      "content": "<p>Congrats to you !!\nGuess the post-processing for prediction with more than 0.50 may have done the trick for you. I was unable to understand and if you are ok to share more details if possible a code sample would be helpful, thanks in advance..</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 479483,
      "author_name": "karthik7395",
      "author_url": "",
      "post_date": "02/27/2019 07:25:15",
      "content": "<p>Congratulations  @YifanXie . </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 479653,
      "author_name": "baltuky",
      "author_url": "",
      "post_date": "02/27/2019 10:12:19",
      "content": "<p>Congratulations!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 479722,
      "author_name": "thisisalpha",
      "author_url": "",
      "post_date": "02/27/2019 11:13:05",
      "content": "<p>congrats Yifan! </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 479815,
      "author_name": "fuengwul5",
      "author_url": "",
      "post_date": "02/27/2019 12:54:59",
      "content": "<p>Congratulations,I,m your fans</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 479957,
      "author_name": "brunogc",
      "author_url": "",
      "post_date": "02/27/2019 16:04:32",
      "content": "<p>Well done, and thanks for all the valuable discussion during the comp!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 480513,
      "author_name": "yangddd",
      "author_url": "",
      "post_date": "02/28/2019 09:18:37",
      "content": "<p>congrats to you, your post give me lot of help and keep me fighting, thx.  I hope and believe you will get a solo gold next time</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 480604,
      "author_name": "econdata",
      "author_url": "",
      "post_date": "02/28/2019 12:09:42",
      "content": "<p>congrats to all the top rankers on LB.This was an interesting competition.Gave me opportunity to learn a lot from the fellow kagglers . And finally thanks to Elo for organizing this data science competition</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "479066": "Hello all, congratulations to the winner, gold medalists,, and especially to those who survived the shake-up :)\n\nI got busy with my work and had to stay away for 4 weeks without doing anything, now looking back that may have prevented me from overfitting my model further!\n\nThe shake-up seems to be quite strong, and I am not entirely sure what worked in my solution and what didn't.  I am pretty sure the following were helpful:\n\n- feature selection by the mean of feature permutation \n- a significant amount of pseudo labelling, even concatenating multiple copies of test data and predict label. one of my best single models was 3 x pseudo labels with public 3.668 private 3.600, but sadly I didn't select it \n- a little moderated post-processing for prediction with more than 0.50 probability seems to help - I used a formula of binary_probabilityX1.1XOUTLIER_VALUE, which helped to improve my best final submission\n\nanyway, I have learnt a lot throughout this competition -- even though I didn't have much as fun as I would like to - competing by yourself is hard, turning down team-up invitation from friends is HARD, getting that solo gold is VERY HARD!\n\nSee you all around the LB!!!!",
    "479083": "Congrats for your solid solo result.  Shakeup did not move you much, did it?",
    "479085": "Congrats..!!",
    "479086": "Congrats!\n\nHow exactly does this `binaryprobability1.1OUTLIERVALUE` formula work? I don't quite get it.",
    "479089": "Congrats! Your post help me a lot.\nThanks!",
    "479092": "sorry, it was meant to be a multiplication formula, I put X in now.",
    "479094": "not being able to compete for the last month definitely helped, although I honestly ran out of idea by that time :)",
    "479097": "&gt;  I honestly ran out of idea \n\nThat's when teaming makes sense ;)",
    "479098": "`OUTLIERVALUE` is the regression value from a model trained without excluding outliers? \n\nGeez, I really suck at interpreting variable names.",
    "479103": "yup",
    "479106": "all right, all right, I get the message :)",
    "479112": "glad it helped, and CONGRATULATION for the prize-winning performance!",
    "479116": "Thanks for your sharing in this game, you should get gold medal. Thanks again.",
    "479126": "Congratulations Yifan. Can you explain a bit more about your feature engineering process ? Just used simple aggregation or other strategies ?",
    "479150": "Congratulations to you @Yifan Xie\nFor your advice:\n\n- feature selection by the mean of feature permutation\n\nI want to know is it 'feature permutation' or 'target permutation'? there seem's have two methods, I want to know which is useful for this competition, thanks!",
    "479161": "Congrats Yifan (laoshi)! Your \"Sharing of my experience so far\" post inspired me and helped me to set a target line (CV 3.62, LB 3.67). I keep moving forward with feature engineering + feature selection and finally it works. I wonder if you will share more about your feature engineering stuff when you have time ;)? Thank you!",
    "479184": "Thanks for your post!",
    "479200": "Congratulations @yifanxie, well deserved. \n\n&gt; feature selection by the mean of feature permutation\n\nI agree, here is my post on feature selection techniques https://www.kaggle.com/c/elo-merchant-category-recommendation/discussion/82052",
    "479297": "so close to gold,I believe next time you can win.",
    "479364": "Congratulations @yifanxie. Could you please ellaborate more on the process of `pseudo labelling`?",
    "479368": "Congrats to you !!\nGuess the post-processing for prediction with more than 0.50 may have done the trick for you. I was unable to understand and if you are ok to share more details if possible a code sample would be helpful, thanks in advance..",
    "479483": "Congratulations  @YifanXie .",
    "479653": "Congratulations!",
    "479722": "congrats Yifan!",
    "479815": "Congratulations,I,m your fans",
    "479957": "Well done, and thanks for all the valuable discussion during the comp!",
    "480513": "congrats to you, your post give me lot of help and keep me fighting, thx.  I hope and believe you will get a solo gold next time",
    "480604": "congrats to all the top rankers on LB.This was an interesting competition.Gave me opportunity to learn a lot from the fellow kagglers . And finally thanks to Elo for organizing this data science competition"
  },
  "source": "meta"
}