I'm participating in this competition as an NLP beginner. I can see that this competition is a classification problem, but how would you label it? Is it a specific case of sentiment analysis? or a completely different kind of classification problem?
\n\nI have to present the implementation of a sentiment analysis algorithm at university, and I thought taking part in a competition and presenting my progress would be a good idea. It has to be sentiment analysis or related though, that's why I'm asking. Thank you!
", "messages": [ { "id": 424582, "postDate": "2018-11-20T11:07:16.920Z", "content": "I'm participating in this competition as an NLP beginner. I can see that this competition is a classification problem, but how would you label it? Is it a specific case of sentiment analysis? or a completely different kind of classification problem?
\n\nI have to present the implementation of a sentiment analysis algorithm at university, and I thought taking part in a competition and presenting my progress would be a good idea. It has to be sentiment analysis or related though, that's why I'm asking. Thank you!
", "rawMarkdown": "I'm participating in this competition as an NLP beginner. I can see that this competition is a classification problem, but how would you label it? Is it a specific case of sentiment analysis? or a completely different kind of classification problem?\n\nI have to present the implementation of a sentiment analysis algorithm at university, and I thought taking part in a competition and presenting my progress would be a good idea. It has to be sentiment analysis or related though, that's why I'm asking. Thank you!", "votes": 1 }, { "id": 424611, "postDate": "2018-11-20T12:23:35.060Z", "content": "For me, sentiment analysis is basically saying how positive/negative a text is. Here, it is a bit different because you are working with questions that are rarely polarized.
\n\nHowever, the two problems are very similar, considering that for both sentiment analysis and this competition you want a score (between 0 and 1) evaluating how positive or insincere the text is. \nTherefore a model for sentiment analysis can be use in this competition, you just have to train it on different data, and it can be a good thing to spend some time on this competition.
", "rawMarkdown": "For me, sentiment analysis is basically saying how positive/negative a text is. Here, it is a bit different because you are working with questions that are rarely polarized. \n\nHowever, the two problems are very similar, considering that for both sentiment analysis and this competition you want a score (between 0 and 1) evaluating how positive or insincere the text is. \nTherefore a model for sentiment analysis can be use in this competition, you just have to train it on different data, and it can be a good thing to spend some time on this competition.", "votes": 2, "replies": [ { "id": 425938, "postDate": "2018-11-22T10:41:11.260Z", "content": "Thank you for your answer Theo! Couldn't be said that insincere reviews could be equivalent to negative reviews, and the rest equivalent to positive reviews? Then again, in this competition the dataset is very imbalanced and that's not usually the case in sentiment analysis right?
", "rawMarkdown": "Thank you for your answer Theo! Couldn't be said that insincere reviews could be equivalent to negative reviews, and the rest equivalent to positive reviews? Then again, in this competition the dataset is very imbalanced and that's not usually the case in sentiment analysis right?", "votes": 1 }, { "id": 425950, "postDate": "2018-11-22T10:55:23.797Z", "content": "Sincere reviews are more neutral than positives. And yes sentiment analysis datasets are pretty much balanced, at least for the ones with 2 classes (0/1)
", "rawMarkdown": "Sincere reviews are more neutral than positives. And yes sentiment analysis datasets are pretty much balanced, at least for the ones with 2 classes (0/1)" } ] }, { "id": 425273, "postDate": "2018-11-21T11:37:41.100Z", "rawMarkdown": "", "isDeleted": true, "replies": [ { "id": 425940, "postDate": "2018-11-22T10:43:17.577Z", "content": "What a coincidence! You mean using a pre-trained sentiment analysis model in this competition?
", "rawMarkdown": "What a coincidence! You mean using a pre-trained sentiment analysis model in this competition?" } ] } ], "comments": [ { "id": 424611, "author_name": "Theo Viel", "author_url": "", "post_date": "2018-11-20T12:23:35.060000", "content": "For me, sentiment analysis is basically saying how positive/negative a text is. Here, it is a bit different because you are working with questions that are rarely polarized.
\n\nHowever, the two problems are very similar, considering that for both sentiment analysis and this competition you want a score (between 0 and 1) evaluating how positive or insincere the text is. \nTherefore a model for sentiment analysis can be use in this competition, you just have to train it on different data, and it can be a good thing to spend some time on this competition.
", "votes": 2, "replies": [ { "id": 425938, "author_name": "Ane", "author_url": "", "post_date": "2018-11-22T10:41:11.260000", "content": "Thank you for your answer Theo! Couldn't be said that insincere reviews could be equivalent to negative reviews, and the rest equivalent to positive reviews? Then again, in this competition the dataset is very imbalanced and that's not usually the case in sentiment analysis right?
", "votes": 1, "replies": [] }, { "id": 425950, "author_name": "Theo Viel", "author_url": "", "post_date": "2018-11-22T10:55:23.797000", "content": "Sincere reviews are more neutral than positives. And yes sentiment analysis datasets are pretty much balanced, at least for the ones with 2 classes (0/1)
", "votes": 0, "replies": [] } ] }, { "id": 425273, "author_name": "", "author_url": "", "post_date": "2018-11-21T11:37:41.100000", "content": "", "votes": 0, "replies": [ { "id": 425940, "author_name": "Ane", "author_url": "", "post_date": "2018-11-22T10:43:17.577000", "content": "What a coincidence! You mean using a pre-trained sentiment analysis model in this competition?
", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "424582": "I'm participating in this competition as an NLP beginner. I can see that this competition is a classification problem, but how would you label it? Is it a specific case of sentiment analysis? or a completely different kind of classification problem?\n\nI have to present the implementation of a sentiment analysis algorithm at university, and I thought taking part in a competition and presenting my progress would be a good idea. It has to be sentiment analysis or related though, that's why I'm asking. Thank you!", "424611": "For me, sentiment analysis is basically saying how positive/negative a text is. Here, it is a bit different because you are working with questions that are rarely polarized. \n\nHowever, the two problems are very similar, considering that for both sentiment analysis and this competition you want a score (between 0 and 1) evaluating how positive or insincere the text is. \nTherefore a model for sentiment analysis can be use in this competition, you just have to train it on different data, and it can be a good thing to spend some time on this competition.", "425273": "" } }