{ "id": 486002, "title": "Better Performance with Online Dataset[TUH Dataset] ", "url": "/competitions/hms-harmful-brain-activity-classification/discussion/486002", "author_name": "Peter", "post_date": "2024-03-23T03:02:36.786000", "votes": 25, "comment_count": 7, "views": 0, "content": "

Hi, Kagglers! πŸ˜€

\n

When using TUH dataset in one stage(low quality) model, it helps my LB score(my model: EfficientNetB0_ns)

\n

Check out Below:
\nπŸ“Œ Processing The TUH EEG Seizure Corpus: https://www.kaggle.com/code/seanbearden/processing-the-tuh-eeg-seizure-corpus
\nπŸ“Œ The TUH 10m EEG Spectrograms Dataset: https://www.kaggle.com/datasets/seanbearden/hms-hba-tuh-tusz-seizures

\n

πŸ“Œ My Custom Notebook: https://www.kaggle.com/code/seoyunje/hms-tuh-eeg-spectrogram
\nπŸ“Œ My Custom 50s Dataset: https://www.kaggle.com/datasets/seoyunje/tuh-eeg-spectrogram
\nπŸ“Œ My Custom 10m Dataset: https://www.kaggle.com/datasets/seoyunje/tuh-10m-spectrogram

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
My Strategy
fIl_na<=0.3
total_evalutors1
Seizure1.0
", "messages": [ { "id": 2711718, "postDate": "2024-03-23T03:02:36.787Z", "content": "

Hi, Kagglers! πŸ˜€

\n

When using TUH dataset in one stage(low quality) model, it helps my LB score(my model: EfficientNetB0_ns)

\n

Check out Below:
\nπŸ“Œ Processing The TUH EEG Seizure Corpus: https://www.kaggle.com/code/seanbearden/processing-the-tuh-eeg-seizure-corpus
\nπŸ“Œ The TUH 10m EEG Spectrograms Dataset: https://www.kaggle.com/datasets/seanbearden/hms-hba-tuh-tusz-seizures

\n

πŸ“Œ My Custom Notebook: https://www.kaggle.com/code/seoyunje/hms-tuh-eeg-spectrogram
\nπŸ“Œ My Custom 50s Dataset: https://www.kaggle.com/datasets/seoyunje/tuh-eeg-spectrogram
\nπŸ“Œ My Custom 10m Dataset: https://www.kaggle.com/datasets/seoyunje/tuh-10m-spectrogram

\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
My Strategy
fIl_na<=0.3
total_evalutors1
Seizure1.0
", "rawMarkdown": "Hi, Kagglers! πŸ˜€\n\n**When using TUH dataset in one stage(low quality) model, it helps my LB score(my model: EfficientNetB0_ns)**\n\nCheck out Below: \nπŸ“Œ Processing The TUH EEG Seizure Corpus: https://www.kaggle.com/code/seanbearden/processing-the-tuh-eeg-seizure-corpus\nπŸ“Œ The TUH 10m EEG Spectrograms Dataset: https://www.kaggle.com/datasets/seanbearden/hms-hba-tuh-tusz-seizures\n\nπŸ“Œ My Custom Notebook: https://www.kaggle.com/code/seoyunje/hms-tuh-eeg-spectrogram \nπŸ“Œ My Custom 50s Dataset: https://www.kaggle.com/datasets/seoyunje/tuh-eeg-spectrogram\nπŸ“Œ My Custom 10m Dataset: https://www.kaggle.com/datasets/seoyunje/tuh-10m-spectrogram\n\n| My Strategy | |\n| --- | --- |\n| fIl_na | <=0.3 |\n| total_evalutors | 1 |\n| Seizure | 1.0 |\n\n\n", "votes": 25 }, { "id": 2723514, "postDate": "2024-03-30T09:40:01.790Z", "content": "

Would like to spend some time on this dataset, but is this dataset allowed in this competition? Seems we still could not find the answer.

", "rawMarkdown": "Would like to spend some time on this dataset, but is this dataset allowed in this competition? Seems we still could not find the answer.", "votes": 1 }, { "id": 2719828, "postDate": "2024-03-28T00:48:22.107Z", "content": "

hi Peter

\n

May I know how the 10m dataset was generated, i checked this url https://www.kaggle.com/code/seoyunje/hms-tuh-10min-spectrogram which was not available

", "rawMarkdown": "hi Peter \n\nMay I know how the 10m dataset was generated, i checked this url https://www.kaggle.com/code/seoyunje/hms-tuh-10min-spectrogram which was not available ", "votes": 1, "replies": [ { "id": 2719975, "postDate": "2024-03-28T03:39:27.713Z", "content": "

Sorry, I forgot to make from private to public. from now on you can access it. And then 10m dataset is also same with 50s dataset. you can customize too.

", "rawMarkdown": "Sorry, I forgot to make from private to public. from now on you can access it. And then 10m dataset is also same with 50s dataset. you can customize too.", "votes": 1 } ] }, { "id": 2716560, "postDate": "2024-03-26T05:19:01.680Z", "content": "

thanks for sharing, for deep learning models, more input data often means better scoresπŸ₯³

", "rawMarkdown": "thanks for sharing, for deep learning models, more input data often means better scoresπŸ₯³", "votes": 1 }, { "id": 2711976, "postDate": "2024-03-23T08:14:27.437Z", "content": "

Awesome work Peter! This is super helpful

", "rawMarkdown": "Awesome work Peter! This is super helpful", "votes": 2 }, { "id": 2716709, "postDate": "2024-03-26T07:01:14.217Z", "rawMarkdown": "", "isDeleted": true }, { "id": 2737209, "postDate": "2024-04-05T16:34:59.470Z", "content": "

Thank you for your work!

", "rawMarkdown": "Thank you for your work!", "votes": 1 } ], "comments": [ { "id": 2723514, "author_name": "gezi", "author_url": "", "post_date": "2024-03-30T09:40:01.790000", "content": "

Would like to spend some time on this dataset, but is this dataset allowed in this competition? Seems we still could not find the answer.

", "votes": 1, "replies": [] }, { "id": 2719828, "author_name": "HZM", "author_url": "", "post_date": "2024-03-28T00:48:22.107000", "content": "

hi Peter

\n

May I know how the 10m dataset was generated, i checked this url https://www.kaggle.com/code/seoyunje/hms-tuh-10min-spectrogram which was not available

", "votes": 1, "replies": [ { "id": 2719975, "author_name": "Peter", "author_url": "", "post_date": "2024-03-28T03:39:27.713000", "content": "

Sorry, I forgot to make from private to public. from now on you can access it. And then 10m dataset is also same with 50s dataset. you can customize too.

", "votes": 1, "replies": [] } ] }, { "id": 2716560, "author_name": "RogerOcean", "author_url": "", "post_date": "2024-03-26T05:19:01.680000", "content": "

thanks for sharing, for deep learning models, more input data often means better scoresπŸ₯³

", "votes": 1, "replies": [] }, { "id": 2711976, "author_name": "snehal", "author_url": "", "post_date": "2024-03-23T08:14:27.437000", "content": "

Awesome work Peter! This is super helpful

", "votes": 2, "replies": [] }, { "id": 2716709, "author_name": "", "author_url": "", "post_date": "2024-03-26T07:01:14.217000", "content": "", "votes": 0, "replies": [] }, { "id": 2737209, "author_name": "Danial Zakaria", "author_url": "", "post_date": "2024-04-05T16:34:59.470000", "content": "

Thank you for your work!

", "votes": 1, "replies": [] } ], "raw_markdown_by_id": { "2711718": "Hi, Kagglers! πŸ˜€\n\n**When using TUH dataset in one stage(low quality) model, it helps my LB score(my model: EfficientNetB0_ns)**\n\nCheck out Below: \nπŸ“Œ Processing The TUH EEG Seizure Corpus: https://www.kaggle.com/code/seanbearden/processing-the-tuh-eeg-seizure-corpus\nπŸ“Œ The TUH 10m EEG Spectrograms Dataset: https://www.kaggle.com/datasets/seanbearden/hms-hba-tuh-tusz-seizures\n\nπŸ“Œ My Custom Notebook: https://www.kaggle.com/code/seoyunje/hms-tuh-eeg-spectrogram \nπŸ“Œ My Custom 50s Dataset: https://www.kaggle.com/datasets/seoyunje/tuh-eeg-spectrogram\nπŸ“Œ My Custom 10m Dataset: https://www.kaggle.com/datasets/seoyunje/tuh-10m-spectrogram\n\n| My Strategy | |\n| --- | --- |\n| fIl_na | <=0.3 |\n| total_evalutors | 1 |\n| Seizure | 1.0 |\n\n\n", "2723514": "Would like to spend some time on this dataset, but is this dataset allowed in this competition? Seems we still could not find the answer.", "2719828": "hi Peter \n\nMay I know how the 10m dataset was generated, i checked this url https://www.kaggle.com/code/seoyunje/hms-tuh-10min-spectrogram which was not available ", "2716560": "thanks for sharing, for deep learning models, more input data often means better scoresπŸ₯³", "2711976": "Awesome work Peter! This is super helpful", "2716709": "", "2737209": "Thank you for your work!" } }