search

mqwfrog / MHCCL

MHCCL: Masked Hierarchical Cluster-wise Contrastive Learning for Multivariate Time Series - a PyTorch Version (AAAI-2023)
33 stars 10 forks source link

Good Jobs！ #6

Closed luojike2333 closed 1 year ago

luojike2333 commented 1 year ago

Good Jobs！ Hello, author.After reading your code, I have a question that the model designed specifically for multivariate time series? Does that mean can't run on a univariate time series.

mqwfrog commented 1 year ago

Hi ~ This model also works with univariate time series, you can set in_channels to 1. As reported in the paper, Epilepsy is a dataset with 1 variable (i.e., univariate time series).

blacksnail789521 commented 1 year ago

Hello, @mqwfrog

I have a question regarding the Epilepsy dataset mentioned in your paper. In the paper, you indicate that the dataset is sourced from UCI and that it has only one feature. However, upon reviewing your code, I found only data_preprocess_wisdm, data_preprocess_HAR, data_preprocess_SHAR, and data_preprocess_UEA methods, but nothing specifically for Epilepsy. This suggests that the Epilepsy dataset might actually be from UEA, not UCI, which is my first point of confusion.

Secondly, while there is indeed an Epilepsy dataset at UEA, it contains four features, not just one as mentioned in your paper. Could you please clarify these discrepancies? Your assistance would be greatly appreciated, and I thank you again for your contribution.

Best regards,

mqwfrog commented 1 year ago

Hi~ The data_preprocess_epilepsy file has been updated. The source of Epilepsy dataset that I used is now unavailable, so the source file is also uploaded in the same folder. The source file is downloaded from https://github.com/emadeldeen24/TS-TCC. You can refer to project TSTCC to get more details.