{
  "id": 47537,
  "title": "A silver solution(33rd)",
  "url": "/competitions/favorita-grocery-sales-forecasting/writeups/struggle-for-a-silver-medal-a-silver-solution-33rd",
  "author_name": "",
  "post_date": "2018-01-16T02:54:14.335003Z",
  "votes": 8,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I'm a newbie in Kaggle and it's actually a shock for me to reach top 2%...</p>\n\n<p>My solution is based on Lingzhi's LGBM One Step Ahead with some modifications:</p>\n\n<ul>\n<li><p>using 8 weeks of training data starting from May 31st, 2017</p></li>\n<li><p>extract min, max, var, skew and kurt</p></li>\n<li><p>calculate first order difference for the mean values</p></li>\n<li><p>Categorical features(store, item): one hot encoding, count (I learn this trick from <a href=\"https://www.kaggle.com/xiaozhouwang/2nd-place-lightgbm-solution\">https://www.kaggle.com/xiaozhouwang/2nd-place-lightgbm-solution</a>)</p></li>\n<li><p>averaged across different random seeds to stabilize model performance (also learned from Little Boat's solution)</p></li>\n</ul>\n\n<p>I have posted my LGBM model on <a href=\"https://www.kaggle.com/deepdreamer/a-silver-solution-33rd\">https://www.kaggle.com/deepdreamer/a-silver-solution-33rd</a>. It scores 0.517 in private board.</p>\n\n<p>Thanks Little Boat for sharing his solutions in Porto Seguro and Ceshine lee for sharing the framework. I really learn a lot from them.</p>",
  "messages": [
    {
      "id": "268984",
      "postDate": "01/16/2018 02:54:14",
      "content": "<p>I'm a newbie in Kaggle and it's actually a shock for me to reach top 2%...</p>\n\n<p>My solution is based on Lingzhi's LGBM One Step Ahead with some modifications:</p>\n\n<ul>\n<li><p>using 8 weeks of training data starting from May 31st, 2017</p></li>\n<li><p>extract min, max, var, skew and kurt</p></li>\n<li><p>calculate first order difference for the mean values</p></li>\n<li><p>Categorical features(store, item): one hot encoding, count (I learn this trick from <a href=\"https://www.kaggle.com/xiaozhouwang/2nd-place-lightgbm-solution\">https://www.kaggle.com/xiaozhouwang/2nd-place-lightgbm-solution</a>)</p></li>\n<li><p>averaged across different random seeds to stabilize model performance (also learned from Little Boat's solution)</p></li>\n</ul>\n\n<p>I have posted my LGBM model on <a href=\"https://www.kaggle.com/deepdreamer/a-silver-solution-33rd\">https://www.kaggle.com/deepdreamer/a-silver-solution-33rd</a>. It scores 0.517 in private board.</p>\n\n<p>Thanks Little Boat for sharing his solutions in Porto Seguro and Ceshine lee for sharing the framework. I really learn a lot from them.</p>",
      "rawMarkdown": "I'm a newbie in Kaggle and it's actually a shock for me to reach top 2%...\n\nMy solution is based on Lingzhi's LGBM One Step Ahead with some modifications:\n\n- using 8 weeks of training data starting from May 31st, 2017\n\n- extract min, max, var, skew and kurt\n\n- calculate first order difference for the mean values\n\n- Categorical features(store, item): one hot encoding, count (I learn this trick from https://www.kaggle.com/xiaozhouwang/2nd-place-lightgbm-solution)\n\n- averaged across different random seeds to stabilize model performance (also learned from Little Boat's solution)\n\nI have posted my LGBM model on https://www.kaggle.com/deepdreamer/a-silver-solution-33rd. It scores 0.517 in private board.\n\nThanks Little Boat for sharing his solutions in Porto Seguro and Ceshine lee for sharing the framework. I really learn a lot from them.",
      "votes": null
    },
    {
      "id": "269302",
      "postDate": "01/16/2018 15:59:13",
      "content": "<p>Nice! Congratiulations, very good!\nFor me this was also my first competition, learned a lot and had a lot of fun.\nHope to see you in other competitions!</p>",
      "rawMarkdown": "Nice! Congratiulations, very good!\nFor me this was also my first competition, learned a lot and had a lot of fun.\nHope to see you in other competitions!",
      "votes": null
    },
    {
      "id": "573967",
      "postDate": "07/13/2019 03:39:44",
      "content": "<p>Thanks for sharing.</p>",
      "rawMarkdown": "Thanks for sharing.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 269302,
      "author_name": "nickydebeer",
      "author_url": "",
      "post_date": "01/16/2018 15:59:13",
      "content": "<p>Nice! Congratiulations, very good!\nFor me this was also my first competition, learned a lot and had a lot of fun.\nHope to see you in other competitions!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 573967,
      "author_name": "longyin2",
      "author_url": "",
      "post_date": "07/13/2019 03:39:44",
      "content": "<p>Thanks for sharing.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "268984": "I'm a newbie in Kaggle and it's actually a shock for me to reach top 2%...\n\nMy solution is based on Lingzhi's LGBM One Step Ahead with some modifications:\n\n- using 8 weeks of training data starting from May 31st, 2017\n\n- extract min, max, var, skew and kurt\n\n- calculate first order difference for the mean values\n\n- Categorical features(store, item): one hot encoding, count (I learn this trick from https://www.kaggle.com/xiaozhouwang/2nd-place-lightgbm-solution)\n\n- averaged across different random seeds to stabilize model performance (also learned from Little Boat's solution)\n\nI have posted my LGBM model on https://www.kaggle.com/deepdreamer/a-silver-solution-33rd. It scores 0.517 in private board.\n\nThanks Little Boat for sharing his solutions in Porto Seguro and Ceshine lee for sharing the framework. I really learn a lot from them.",
    "269302": "Nice! Congratiulations, very good!\nFor me this was also my first competition, learned a lot and had a lot of fun.\nHope to see you in other competitions!",
    "573967": "Thanks for sharing."
  },
  "source": "meta"
}