{
  "id": 82919,
  "title": "Sentiment analysis / Polarity ",
  "url": "/competitions/quora-insincere-questions-classification/discussion/82919",
  "author_name": "",
  "post_date": "2019-03-05T10:50:39.411540600Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Any NLP wizards here!!! Wana know better ways of having the polarity of a sentence and the words/ phrases contributing to it!!</p>\n\n<p>Tried: Txtblob, NLTK, Flair, used word2vec/glove for having polarized words/phrases.</p>\n\n<p>Is any other way like … stacking models or some other!</p>\n\n<p>Thanks in advance</p>",
  "messages": [
    {
      "id": "483926",
      "postDate": "03/05/2019 10:50:39",
      "content": "<p>Any NLP wizards here!!! Wana know better ways of having the polarity of a sentence and the words/ phrases contributing to it!!</p>\n\n<p>Tried: Txtblob, NLTK, Flair, used word2vec/glove for having polarized words/phrases.</p>\n\n<p>Is any other way like … stacking models or some other!</p>\n\n<p>Thanks in advance</p>",
      "rawMarkdown": "Any NLP wizards here!!! Wana know better ways of having the polarity of a sentence and the words/ phrases contributing to it!!\n\nTried: Txtblob, NLTK, Flair, used word2vec/glove for having polarized words/phrases.\n\nIs any other way like … stacking models or some other!\n\nThanks in advance",
      "votes": null
    },
    {
      "id": "483948",
      "postDate": "03/05/2019 11:43:15",
      "content": "<p><a href=\"https://eli5.readthedocs.io/en/latest/tutorials/black-box-text-classifiers.html\">https://eli5.readthedocs.io/en/latest/tutorials/black-box-text-classifiers.html</a></p>",
      "rawMarkdown": "https://eli5.readthedocs.io/en/latest/tutorials/black-box-text-classifiers.html",
      "votes": null
    },
    {
      "id": "485487",
      "postDate": "03/07/2019 12:49:00",
      "content": "<p>Thanks for the info mate. Good to know that! \nBut It took a lot of time for larger articles(used flair package for getting polarity), as it tires different combinations of the text and having the polairty.</p>\n\n<p>Can you suggest a way to find polarity without using any pretrained models/packages. </p>\n\n<p>Thanks and sorry for late reply. </p>",
      "rawMarkdown": "Thanks for the info mate. Good to know that! \nBut It took a lot of time for larger articles(used flair package for getting polarity), as it tires different combinations of the text and having the polairty.\n\nCan you suggest a way to find polarity without using any pretrained models/packages. \n\nThanks and sorry for late reply.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 483948,
      "author_name": "artgor",
      "author_url": "",
      "post_date": "03/05/2019 11:43:15",
      "content": "<p><a href=\"https://eli5.readthedocs.io/en/latest/tutorials/black-box-text-classifiers.html\">https://eli5.readthedocs.io/en/latest/tutorials/black-box-text-classifiers.html</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 485487,
          "author_name": "vijay09",
          "author_url": "",
          "post_date": "03/07/2019 12:49:00",
          "content": "<p>Thanks for the info mate. Good to know that! \nBut It took a lot of time for larger articles(used flair package for getting polarity), as it tires different combinations of the text and having the polairty.</p>\n\n<p>Can you suggest a way to find polarity without using any pretrained models/packages. </p>\n\n<p>Thanks and sorry for late reply. </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "483926": "Any NLP wizards here!!! Wana know better ways of having the polarity of a sentence and the words/ phrases contributing to it!!\n\nTried: Txtblob, NLTK, Flair, used word2vec/glove for having polarized words/phrases.\n\nIs any other way like … stacking models or some other!\n\nThanks in advance",
    "483948": "https://eli5.readthedocs.io/en/latest/tutorials/black-box-text-classifiers.html",
    "485487": "Thanks for the info mate. Good to know that! \nBut It took a lot of time for larger articles(used flair package for getting polarity), as it tires different combinations of the text and having the polairty.\n\nCan you suggest a way to find polarity without using any pretrained models/packages. \n\nThanks and sorry for late reply."
  },
  "source": "meta"
}