{
  "id": 347324,
  "title": "Does anyone try model based on boosting or bagging?",
  "url": "/competitions/open-problems-multimodal/discussion/347324",
  "author_name": "",
  "post_date": "2022-08-23T20:50:25.266215800Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hi, I find that for multiome data, the number of cells in the training dataset is too large to train efficiently, and I also cannot filter too many peaks. Does anyone have better solutions for this thing? Thanks.</p>",
  "messages": [
    {
      "id": "1910986",
      "postDate": "08/23/2022 20:50:25",
      "content": "<p>Hi, I find that for multiome data, the number of cells in the training dataset is too large to train efficiently, and I also cannot filter too many peaks. Does anyone have better solutions for this thing? Thanks.</p>",
      "rawMarkdown": "Hi, I find that for multiome data, the number of cells in the training dataset is too large to train efficiently, and I also cannot filter too many peaks. Does anyone have better solutions for this thing? Thanks.",
      "votes": null
    },
    {
      "id": "1912386",
      "postDate": "08/24/2022 17:33:59",
      "content": "<p>Not sure about the memory requirements, but you can use lightgbm for a multiple output regression problem as follows:</p>\n<p>from sklearn.multioutput import MultiOutputRegressor<br>\nimport lightgbm as lgb</p>\n<p>model = MultiOutputRegressor(lgb.LGBMRegressor(random_state=42), n_jobs=-1)</p>\n<p>model.fit(X,y)</p>",
      "rawMarkdown": "Not sure about the memory requirements, but you can use lightgbm for a multiple output regression problem as follows:\n\nfrom sklearn.multioutput import MultiOutputRegressor\nimport lightgbm as lgb\n\nmodel = MultiOutputRegressor(lgb.LGBMRegressor(random_state=42), n_jobs=-1)\n\nmodel.fit(X,y)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1912386,
      "author_name": "nymfree",
      "author_url": "",
      "post_date": "08/24/2022 17:33:59",
      "content": "<p>Not sure about the memory requirements, but you can use lightgbm for a multiple output regression problem as follows:</p>\n<p>from sklearn.multioutput import MultiOutputRegressor<br>\nimport lightgbm as lgb</p>\n<p>model = MultiOutputRegressor(lgb.LGBMRegressor(random_state=42), n_jobs=-1)</p>\n<p>model.fit(X,y)</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1910986": "Hi, I find that for multiome data, the number of cells in the training dataset is too large to train efficiently, and I also cannot filter too many peaks. Does anyone have better solutions for this thing? Thanks.",
    "1912386": "Not sure about the memory requirements, but you can use lightgbm for a multiple output regression problem as follows:\n\nfrom sklearn.multioutput import MultiOutputRegressor\nimport lightgbm as lgb\n\nmodel = MultiOutputRegressor(lgb.LGBMRegressor(random_state=42), n_jobs=-1)\n\nmodel.fit(X,y)"
  },
  "source": "meta"
}