{ "id": 437262, "title": "Possible technique? Temporal Convolutional Networks", "url": "/competitions/child-mind-institute-detect-sleep-states/discussion/437262", "author_name": "", "post_date": "2023-09-06T04:45:16.939411900Z", "votes": 12, "comment_count": 1, "views": 0, "content": "

Having taken a cursory look at the data, one possible technique that could be of potential use here are Temporal Convolutional Networks designed for sequence modeling. Here is a link to a notebook that was used in a previous competition with temporal data: \"Temporal Convolutional Network using Keras-TCN\" using the Keras TCN package.

\n

As in this competition we have no access to the internet, to insall, attach the kaggle dataset keras-tcn (3.5.0) whl to your notebook, then install keras-TCN with

\n
!pip install -q /kaggle//keras-tcn--whl/keras_tcn--py3-none-.whl\n tcn  TCN, tcn_full_summary\n
\n

which should work with the internet off.

\n

All the best,
\ncarl

", "messages": [ { "id": "2425636", "postDate": "09/06/2023 04:45:16", "content": "

Having taken a cursory look at the data, one possible technique that could be of potential use here are Temporal Convolutional Networks designed for sequence modeling. Here is a link to a notebook that was used in a previous competition with temporal data: \"Temporal Convolutional Network using Keras-TCN\" using the Keras TCN package.

\n

As in this competition we have no access to the internet, to insall, attach the kaggle dataset keras-tcn (3.5.0) whl to your notebook, then install keras-TCN with

\n
!pip install -q /kaggle//keras-tcn--whl/keras_tcn--py3-none-.whl\n tcn  TCN, tcn_full_summary\n
\n

which should work with the internet off.

\n

All the best,
\ncarl

", "rawMarkdown": "Having taken a cursory look at the data, one possible technique that could be of potential use here are Temporal Convolutional Networks designed for sequence modeling. Here is a link to a notebook that was used in a previous competition with temporal data: [\"Temporal Convolutional Network using Keras-TCN\"](https://www.kaggle.com/code/carlmcbrideellis/temporal-convolutional-network-using-keras-tcn) using the [Keras TCN](https://github.com/philipperemy/keras-tcn) package.\n\n\nAs in this competition we have no access to the internet, to insall, attach the kaggle dataset [keras-tcn (3.5.0) whl](https://www.kaggle.com/datasets/carlmcbrideellis/keras-tcn-350-whl) to your notebook, then install keras-TCN with\n\n```python\n!pip install -q /kaggle/input/keras-tcn-350-whl/keras_tcn-3.5.0-py3-none-any.whl\nfrom tcn import TCN, tcn_full_summary\n```\n\nwhich should work with the internet off.\n\n\nAll the best,\ncarl", "votes": null }, { "id": "2425681", "postDate": "09/06/2023 05:37:43", "content": "

Will check it out ty

", "rawMarkdown": "Will check it out ty", "votes": null } ], "comments": [ { "id": 2425681, "author_name": "thomcat", "author_url": "", "post_date": "09/06/2023 05:37:43", "content": "

Will check it out ty

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "2425636": "Having taken a cursory look at the data, one possible technique that could be of potential use here are Temporal Convolutional Networks designed for sequence modeling. Here is a link to a notebook that was used in a previous competition with temporal data: [\"Temporal Convolutional Network using Keras-TCN\"](https://www.kaggle.com/code/carlmcbrideellis/temporal-convolutional-network-using-keras-tcn) using the [Keras TCN](https://github.com/philipperemy/keras-tcn) package.\n\n\nAs in this competition we have no access to the internet, to insall, attach the kaggle dataset [keras-tcn (3.5.0) whl](https://www.kaggle.com/datasets/carlmcbrideellis/keras-tcn-350-whl) to your notebook, then install keras-TCN with\n\n```python\n!pip install -q /kaggle/input/keras-tcn-350-whl/keras_tcn-3.5.0-py3-none-any.whl\nfrom tcn import TCN, tcn_full_summary\n```\n\nwhich should work with the internet off.\n\n\nAll the best,\ncarl", "2425681": "Will check it out ty" }, "source": "meta" }