{
  "id": 79743,
  "title": "Approaching it as a classification task",
  "url": "/competitions/humpback-whale-identification/discussion/79743",
  "author_name": "",
  "post_date": "2019-02-07T05:53:28.118361100Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I'm trying to solve this problem as a classification task. My solution is <a href=\"https://www.kaggle.com/game1level2/cnn-whalev2?scriptVersionId=10324242\">here</a>. </p>\n\n<p>Removed <code>new_whale</code> from training data, implemented some basic <code>data augmentation</code>, normalized the images. Getting high accuracy on the <code>training data</code>, but the score on <code>test data</code> is low, <code>0.290</code>. </p>\n\n<p>My question is, what is the best way to approach it as a classification task? Is there something obvious that I'm doing wrong?</p>",
  "messages": [
    {
      "id": "467435",
      "postDate": "02/07/2019 05:53:28",
      "content": "<p>I'm trying to solve this problem as a classification task. My solution is <a href=\"https://www.kaggle.com/game1level2/cnn-whalev2?scriptVersionId=10324242\">here</a>. </p>\n\n<p>Removed <code>new_whale</code> from training data, implemented some basic <code>data augmentation</code>, normalized the images. Getting high accuracy on the <code>training data</code>, but the score on <code>test data</code> is low, <code>0.290</code>. </p>\n\n<p>My question is, what is the best way to approach it as a classification task? Is there something obvious that I'm doing wrong?</p>",
      "rawMarkdown": "I'm trying to solve this problem as a classification task. My solution is [here][1]. \n\nRemoved `new_whale` from training data, implemented some basic `data augmentation`, normalized the images. Getting high accuracy on the `training data`, but the score on `test data` is low, `0.290`. \n\nMy question is, what is the best way to approach it as a classification task? Is there something obvious that I'm doing wrong?\n\n\n  [1]: https://www.kaggle.com/game1level2/cnn-whalev2?scriptVersionId=10324242",
      "votes": null
    },
    {
      "id": "467524",
      "postDate": "02/07/2019 09:18:26",
      "content": "<p>This <a href=\"https://www.kaggle.com/c/humpback-whale-identification/discussion/74647\">https://www.kaggle.com/c/humpback-whale-identification/discussion/74647</a></p>",
      "rawMarkdown": "This https://www.kaggle.com/c/humpback-whale-identification/discussion/74647",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 467524,
      "author_name": "oldufo",
      "author_url": "",
      "post_date": "02/07/2019 09:18:26",
      "content": "<p>This <a href=\"https://www.kaggle.com/c/humpback-whale-identification/discussion/74647\">https://www.kaggle.com/c/humpback-whale-identification/discussion/74647</a></p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "467435": "I'm trying to solve this problem as a classification task. My solution is [here][1]. \n\nRemoved `new_whale` from training data, implemented some basic `data augmentation`, normalized the images. Getting high accuracy on the `training data`, but the score on `test data` is low, `0.290`. \n\nMy question is, what is the best way to approach it as a classification task? Is there something obvious that I'm doing wrong?\n\n\n  [1]: https://www.kaggle.com/game1level2/cnn-whalev2?scriptVersionId=10324242",
    "467524": "This https://www.kaggle.com/c/humpback-whale-identification/discussion/74647"
  },
  "source": "meta"
}