{ "id": 188901, "title": "Paper and more datasets by the organizer Riiid", "url": "/competitions/riiid-test-answer-prediction/discussion/188901", "author_name": "Sirish Somanchi", "post_date": "2020-10-05T19:39:24.837000", "votes": 33, "comment_count": 8, "views": 0, "content": "

[Paper]: EdNet: A Large-Scale Hierarchical Dataset in Education

\n

See Github for additional datasets by the organizer:

\n\n

See EDA jupyter notebooks

", "messages": [ { "id": 1038443, "postDate": "2020-10-05T19:39:24.837Z", "content": "

[Paper]: EdNet: A Large-Scale Hierarchical Dataset in Education

\n

See Github for additional datasets by the organizer:

\n\n

See EDA jupyter notebooks

", "rawMarkdown": "[[Paper](https://arxiv.org/pdf/1912.03072.pdf)]: [EdNet](https://arxiv.org/abs/1912.03072): A Large-Scale Hierarchical Dataset in Education\n\nSee [Github](https://github.com/riiid/ednet) for additional datasets by the organizer:\n- EdNet-KT1 : bit.ly/ednet_kt1\n- EdNet-KT2 : bit.ly/ednet-kt2\n- EdNet-KT3 : bit.ly/ednet-kt3\n- EdNet-KT4 : bit.ly/ednet-kt4\n- Contents : bit.ly/ednet-content\n\nSee EDA [jupyter notebooks](https://github.com/premonish/EdNet/blob/master/notebooks/)\n", "votes": 33 }, { "id": 1045560, "postDate": "2020-10-10T18:42:27.020Z", "content": "

Among the above for datasets, which is the competition dataset provided?

", "rawMarkdown": "Among the above for datasets, which is the competition dataset provided?", "votes": 1, "replies": [ { "id": 1045574, "postDate": "2020-10-10T19:09:41.633Z", "content": "

Yes, are we sure these are not contained within the competition dataset?

", "rawMarkdown": "Yes, are we sure these are not contained within the competition dataset?" }, { "id": 1045888, "postDate": "2020-10-11T06:10:33.860Z", "content": "

After my inspection, I think KT4 is provided to us. Since EDNet is a Hierarchical dataset, as we go from KT1 to KT4 the granularity of the data increases. So Since we are provided with the one with the highest granularity it is possible to extract any other lower Dataset by aggregation. The attached screenshot is taken from EDNet paper which explains my findings

", "rawMarkdown": "After my inspection, I think KT4 is provided to us. Since EDNet is a Hierarchical dataset, as we go from KT1 to KT4 the granularity of the data increases. So Since we are provided with the one with the highest granularity it is possible to extract any other lower Dataset by aggregation. The attached screenshot is taken from EDNet paper which explains my findings\n\n", "votes": 1 }, { "id": 1047503, "postDate": "2020-10-12T17:06:26.167Z", "content": "

The dataset provided to us doesn't strictly fall into any category in KT1-4 hierarchy.

\n

We are provided (q,a) pairs like in KT1.
\nWe are not provided the info on a student alternating between choices like in KT2
\nWe are provided the provided info on lecture and explanation viewing like in KT3
\nWe are not provided the fine details like payment, audio plays etc. like in KT4

", "rawMarkdown": "The dataset provided to us doesn't strictly fall into any category in KT1-4 hierarchy. \n\nWe are provided (q,a) pairs like in KT1.\nWe are **not** provided the info on a student alternating between choices like in KT2\nWe are provided the provided info on lecture and explanation viewing like in KT3\nWe are **not** provided the fine details like payment, audio plays etc. like in KT4", "votes": 3 }, { "id": 1105752, "postDate": "2020-12-08T06:59:35.927Z", "content": "

@gautham11 is there any more data also there apart from Train.csv i see people in SAKT public kernel talking about training with more no of Users

", "rawMarkdown": "@gautham11 is there any more data also there apart from Train.csv i see people in SAKT public kernel talking about training with more no of Users " } ] }, { "id": 1042770, "postDate": "2020-10-08T13:04:37.783Z", "rawMarkdown": "", "votes": 1, "isDeleted": true }, { "id": 1045566, "postDate": "2020-10-10T18:57:58.820Z", "content": "

Amazing, thanks for sharing :)

", "rawMarkdown": "Amazing, thanks for sharing :)", "votes": 2 }, { "id": 1070935, "postDate": "2020-11-06T10:27:34.843Z", "content": "

Thanks, very useful indeed!

", "rawMarkdown": "Thanks, very useful indeed!" } ], "comments": [ { "id": 1045560, "author_name": "Aravind P", "author_url": "", "post_date": "2020-10-10T18:42:27.020000", "content": "

Among the above for datasets, which is the competition dataset provided?

", "votes": 1, "replies": [ { "id": 1045574, "author_name": "Alex", "author_url": "", "post_date": "2020-10-10T19:09:41.633000", "content": "

Yes, are we sure these are not contained within the competition dataset?

", "votes": 0, "replies": [] }, { "id": 1045888, "author_name": "Aravind P", "author_url": "", "post_date": "2020-10-11T06:10:33.860000", "content": "

After my inspection, I think KT4 is provided to us. Since EDNet is a Hierarchical dataset, as we go from KT1 to KT4 the granularity of the data increases. So Since we are provided with the one with the highest granularity it is possible to extract any other lower Dataset by aggregation. The attached screenshot is taken from EDNet paper which explains my findings

", "votes": 1, "replies": [] }, { "id": 1047503, "author_name": "GauthamKumaran", "author_url": "", "post_date": "2020-10-12T17:06:26.167000", "content": "

The dataset provided to us doesn't strictly fall into any category in KT1-4 hierarchy.

\n

We are provided (q,a) pairs like in KT1.
\nWe are not provided the info on a student alternating between choices like in KT2
\nWe are provided the provided info on lecture and explanation viewing like in KT3
\nWe are not provided the fine details like payment, audio plays etc. like in KT4

", "votes": 3, "replies": [] }, { "id": 1105752, "author_name": "Jaideep", "author_url": "", "post_date": "2020-12-08T06:59:35.927000", "content": "

@gautham11 is there any more data also there apart from Train.csv i see people in SAKT public kernel talking about training with more no of Users

", "votes": 0, "replies": [] } ] }, { "id": 1042770, "author_name": "", "author_url": "", "post_date": "2020-10-08T13:04:37.783000", "content": "", "votes": 1, "replies": [] }, { "id": 1045566, "author_name": "domiziano.stingi95", "author_url": "", "post_date": "2020-10-10T18:57:58.820000", "content": "

Amazing, thanks for sharing :)

", "votes": 2, "replies": [] }, { "id": 1070935, "author_name": "Jorge Marcher", "author_url": "", "post_date": "2020-11-06T10:27:34.843000", "content": "

Thanks, very useful indeed!

", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "1038443": "[[Paper](https://arxiv.org/pdf/1912.03072.pdf)]: [EdNet](https://arxiv.org/abs/1912.03072): A Large-Scale Hierarchical Dataset in Education\n\nSee [Github](https://github.com/riiid/ednet) for additional datasets by the organizer:\n- EdNet-KT1 : bit.ly/ednet_kt1\n- EdNet-KT2 : bit.ly/ednet-kt2\n- EdNet-KT3 : bit.ly/ednet-kt3\n- EdNet-KT4 : bit.ly/ednet-kt4\n- Contents : bit.ly/ednet-content\n\nSee EDA [jupyter notebooks](https://github.com/premonish/EdNet/blob/master/notebooks/)\n", "1045560": "Among the above for datasets, which is the competition dataset provided?", "1042770": "", "1045566": "Amazing, thanks for sharing :)", "1070935": "Thanks, very useful indeed!" } }