most of us can enhance 0.003 while he used Count vectorization based on category, but I sharply decrease my score from 3.687->3.697 while I do it......
\n\nI want to know why? I haven't make FM/remove outlier/category_encode/feature selection
", "messages": [ { "id": 452644, "postDate": "2019-01-09T01:34:06.947Z", "content": "most of us can enhance 0.003 while he used Count vectorization based on category, but I sharply decrease my score from 3.687->3.697 while I do it......
\n\nI want to know why? I haven't make FM/remove outlier/category_encode/feature selection
", "rawMarkdown": "most of us can enhance 0.003 while he used Count vectorization based on category, but I sharply decrease my score from 3.687->3.697 while I do it......\n\nI want to know why? I haven't make FM/remove outlier/category_encode/feature selection", "votes": 3 }, { "id": 452829, "postDate": "2019-01-09T08:03:07.240Z", "content": "don't help me too!
", "rawMarkdown": "don't help me too!", "replies": [ { "id": 452887, "postDate": "2019-01-09T10:09:38.363Z", "content": "Yeah,I think it's only based on your data pattern,some people maybe can't improve score beacause of the infomation provided by features is similar
", "rawMarkdown": "Yeah,I think it's only based on your data pattern,some people maybe can't improve score beacause of the infomation provided by features is similar" }, { "id": 452890, "postDate": "2019-01-09T10:11:23.263Z", "content": "But I wanna know the count-method u did it , similar to me or the other?(I make dummies, followed by sum_agg)
", "rawMarkdown": "But I wanna know the count-method u did it , similar to me or the other?(I make dummies, followed by sum_agg)" }, { "id": 452933, "postDate": "2019-01-09T11:27:09.390Z", "content": "I just make a aggregate by card_id and then join them together as a string, then use count vectorzer, search some example about count Vectorizer. let me know if this work for you!
", "rawMarkdown": "I just make a aggregate by card_id and then join them together as a string, then use count vectorzer, search some example about count Vectorizer. let me know if this work for you!", "votes": 1 }, { "id": 452968, "postDate": "2019-01-09T12:24:52.990Z", "content": "u means that join category2/3 as a string?such as A1/A3/B3 and so on ?
", "rawMarkdown": "u means that join category2/3 as a string?such as A1/A3/B3 and so on ?" }, { "id": 453370, "postDate": "2019-01-10T04:39:31.933Z", "content": "Yes, like 'A1 A2 A3' , then you can use count Vectorizer.
", "rawMarkdown": "Yes, like 'A1 A2 A3' , then you can use count Vectorizer. " }, { "id": 453376, "postDate": "2019-01-10T04:54:34.547Z", "content": "yeah,I will tell u if this works for me
", "rawMarkdown": "yeah,I will tell u if this works for me" }, { "id": 453491, "postDate": "2019-01-10T09:18:43.470Z", "content": "this cannot work for me
", "rawMarkdown": "this cannot work for me", "votes": 1 }, { "id": 454109, "postDate": "2019-01-11T06:23:08.633Z", "content": "why?
", "rawMarkdown": "why?" }, { "id": 454504, "postDate": "2019-01-11T18:56:23.770Z", "rawMarkdown": "", "isDeleted": true }, { "id": 454505, "postDate": "2019-01-11T18:57:15.307Z", "content": "Count vectorizer, is for text in a set of documents. It just converts a collection of text documents to a matrix of token counts. How can it help us? Or do you mean to use it at character level?
", "rawMarkdown": "Count vectorizer, is for text in a set of documents. It just converts a collection of text documents to a matrix of token counts. How can it help us? Or do you mean to use it at character level?" }, { "id": 454696, "postDate": "2019-01-12T03:12:27.197Z", "content": "I also can't figure it out,count A1/A2/B1 maybe help me but lee than 0.001
", "rawMarkdown": "I also can't figure it out,count A1/A2/B1 maybe help me but lee than 0.001" } ] }, { "id": 452667, "postDate": "2019-01-09T02:40:37.707Z", "content": "how r u doing that?
", "rawMarkdown": "how r u doing that?", "replies": [ { "id": 452676, "postDate": "2019-01-09T03:01:21.417Z", "rawMarkdown": "", "isDeleted": true }, { "id": 452679, "postDate": "2019-01-09T03:03:20.803Z", "content": "Hi,SubikashPal . I make dummies for category ,followed by agg(groupby card_id) sum,May be I haven't figured out the mean of COUNT?
", "rawMarkdown": "Hi,SubikashPal . I make dummies for category ,followed by agg(groupby card_id) sum,May be I haven't figured out the mean of COUNT?" } ] } ], "comments": [ { "id": 452829, "author_name": "Jianfei Wang", "author_url": "", "post_date": "2019-01-09T08:03:07.240000", "content": "don't help me too!
", "votes": 0, "replies": [ { "id": 452887, "author_name": "Neo Zhao", "author_url": "", "post_date": "2019-01-09T10:09:38.363000", "content": "Yeah,I think it's only based on your data pattern,some people maybe can't improve score beacause of the infomation provided by features is similar
", "votes": 0, "replies": [] }, { "id": 452890, "author_name": "Neo Zhao", "author_url": "", "post_date": "2019-01-09T10:11:23.263000", "content": "But I wanna know the count-method u did it , similar to me or the other?(I make dummies, followed by sum_agg)
", "votes": 0, "replies": [] }, { "id": 452933, "author_name": "Jianfei Wang", "author_url": "", "post_date": "2019-01-09T11:27:09.390000", "content": "I just make a aggregate by card_id and then join them together as a string, then use count vectorzer, search some example about count Vectorizer. let me know if this work for you!
", "votes": 1, "replies": [] }, { "id": 452968, "author_name": "Neo Zhao", "author_url": "", "post_date": "2019-01-09T12:24:52.990000", "content": "u means that join category2/3 as a string?such as A1/A3/B3 and so on ?
", "votes": 0, "replies": [] }, { "id": 453370, "author_name": "Jianfei Wang", "author_url": "", "post_date": "2019-01-10T04:39:31.933000", "content": "Yes, like 'A1 A2 A3' , then you can use count Vectorizer.
", "votes": 0, "replies": [] }, { "id": 453376, "author_name": "Neo Zhao", "author_url": "", "post_date": "2019-01-10T04:54:34.547000", "content": "yeah,I will tell u if this works for me
", "votes": 0, "replies": [] }, { "id": 453491, "author_name": "Neo Zhao", "author_url": "", "post_date": "2019-01-10T09:18:43.470000", "content": "this cannot work for me
", "votes": 1, "replies": [] }, { "id": 454109, "author_name": "Jianfei Wang", "author_url": "", "post_date": "2019-01-11T06:23:08.633000", "content": "why?
", "votes": 0, "replies": [] }, { "id": 454504, "author_name": "", "author_url": "", "post_date": "2019-01-11T18:56:23.770000", "content": "", "votes": 0, "replies": [] }, { "id": 454505, "author_name": "stout", "author_url": "", "post_date": "2019-01-11T18:57:15.307000", "content": "Count vectorizer, is for text in a set of documents. It just converts a collection of text documents to a matrix of token counts. How can it help us? Or do you mean to use it at character level?
", "votes": 0, "replies": [] }, { "id": 454696, "author_name": "Neo Zhao", "author_url": "", "post_date": "2019-01-12T03:12:27.197000", "content": "I also can't figure it out,count A1/A2/B1 maybe help me but lee than 0.001
", "votes": 0, "replies": [] } ] }, { "id": 452667, "author_name": "SubikashPal", "author_url": "", "post_date": "2019-01-09T02:40:37.707000", "content": "how r u doing that?
", "votes": 0, "replies": [ { "id": 452676, "author_name": "", "author_url": "", "post_date": "2019-01-09T03:01:21.417000", "content": "", "votes": 0, "replies": [] }, { "id": 452679, "author_name": "Neo Zhao", "author_url": "", "post_date": "2019-01-09T03:03:20.803000", "content": "Hi,SubikashPal . I make dummies for category ,followed by agg(groupby card_id) sum,May be I haven't figured out the mean of COUNT?
", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "452644": "most of us can enhance 0.003 while he used Count vectorization based on category, but I sharply decrease my score from 3.687->3.697 while I do it......\n\nI want to know why? I haven't make FM/remove outlier/category_encode/feature selection", "452829": "don't help me too!", "452667": "how r u doing that?" } }