I'm looking to taking part in my first featured competition and found this interesting. My goal is mainly learning how to model text classification tasks. I've looked at a few kernels and they seem quite simple although they seem to hide what's going on in the background. Any advice on what learning path I should take that is learn while I \"compete\"?!\n I'm currently (more) comfortable with R although I've done a bit of Python.\nThanks in advance.
", "messages": [ { "id": "427954", "postDate": "11/26/2018 13:10:32", "content": "I'm looking to taking part in my first featured competition and found this interesting. My goal is mainly learning how to model text classification tasks. I've looked at a few kernels and they seem quite simple although they seem to hide what's going on in the background. Any advice on what learning path I should take that is learn while I \"compete\"?!\n I'm currently (more) comfortable with R although I've done a bit of Python.\nThanks in advance.
", "rawMarkdown": "I'm looking to taking part in my first featured competition and found this interesting. My goal is mainly learning how to model text classification tasks. I've looked at a few kernels and they seem quite simple although they seem to hide what's going on in the background. Any advice on what learning path I should take that is learn while I \"compete\"?!\n I'm currently (more) comfortable with R although I've done a bit of Python.\nThanks in advance.", "votes": null }, { "id": "427971", "postDate": "11/26/2018 13:50:59", "content": "You probably won't get very far with R... Virtually all the winning entries are Deep Learning approaches, so I would heavily recommend you learn some Python and do a few Keras tutorials (they also have some great examples on their GitHub repo). Keras is extremely easy to use, has a great API, and is pretty much the standard for Kernels here on Kaggle.
\n\nTheoretical parts you can read up on include:\n- Word Embeddings\n- RNNs / LSTMs / GRUs
\n\nThe problem is closely related to sentiment analysis as well, so reading tutorials / papers about that subject will also help you greatly.
\n\nAnd a final tip: When reading papers / tutorials, do keep in mind that the state of the art gets set every few months at the moment, so any material you find is probably outdated unless written in the last few months, maybe year.
", "rawMarkdown": "You probably won't get very far with R... Virtually all the winning entries are Deep Learning approaches, so I would heavily recommend you learn some Python and do a few Keras tutorials (they also have some great examples on their GitHub repo). Keras is extremely easy to use, has a great API, and is pretty much the standard for Kernels here on Kaggle.\n\nTheoretical parts you can read up on include:\n- Word Embeddings\n- RNNs / LSTMs / GRUs\n\nThe problem is closely related to sentiment analysis as well, so reading tutorials / papers about that subject will also help you greatly.\n\nAnd a final tip: When reading papers / tutorials, do keep in mind that the state of the art gets set every few months at the moment, so any material you find is probably outdated unless written in the last few months, maybe year.", "votes": null }, { "id": "429995", "postDate": "11/29/2018 16:26:11", "content": "Thank you very much. I will resume my python journey. Thanks again!
", "rawMarkdown": "Thank you very much. I will resume my python journey. Thanks again!", "votes": null } ], "comments": [ { "id": 427971, "author_name": "mschumacher", "author_url": "", "post_date": "11/26/2018 13:50:59", "content": "You probably won't get very far with R... Virtually all the winning entries are Deep Learning approaches, so I would heavily recommend you learn some Python and do a few Keras tutorials (they also have some great examples on their GitHub repo). Keras is extremely easy to use, has a great API, and is pretty much the standard for Kernels here on Kaggle.
\n\nTheoretical parts you can read up on include:\n- Word Embeddings\n- RNNs / LSTMs / GRUs
\n\nThe problem is closely related to sentiment analysis as well, so reading tutorials / papers about that subject will also help you greatly.
\n\nAnd a final tip: When reading papers / tutorials, do keep in mind that the state of the art gets set every few months at the moment, so any material you find is probably outdated unless written in the last few months, maybe year.
", "votes": null, "replies": [ { "id": 429995, "author_name": "gonnel", "author_url": "", "post_date": "11/29/2018 16:26:11", "content": "Thank you very much. I will resume my python journey. Thanks again!
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "427954": "I'm looking to taking part in my first featured competition and found this interesting. My goal is mainly learning how to model text classification tasks. I've looked at a few kernels and they seem quite simple although they seem to hide what's going on in the background. Any advice on what learning path I should take that is learn while I \"compete\"?!\n I'm currently (more) comfortable with R although I've done a bit of Python.\nThanks in advance.", "427971": "You probably won't get very far with R... Virtually all the winning entries are Deep Learning approaches, so I would heavily recommend you learn some Python and do a few Keras tutorials (they also have some great examples on their GitHub repo). Keras is extremely easy to use, has a great API, and is pretty much the standard for Kernels here on Kaggle.\n\nTheoretical parts you can read up on include:\n- Word Embeddings\n- RNNs / LSTMs / GRUs\n\nThe problem is closely related to sentiment analysis as well, so reading tutorials / papers about that subject will also help you greatly.\n\nAnd a final tip: When reading papers / tutorials, do keep in mind that the state of the art gets set every few months at the moment, so any material you find is probably outdated unless written in the last few months, maybe year.", "429995": "Thank you very much. I will resume my python journey. Thanks again!" }, "source": "meta" }