0
$\begingroup$

Screen shot of dataset are below

I have a dataset(IoT wearable dataset) composed of time-series and integer data; the objective of my task is to use the dataset for classification. Whilst current libraries in sktime accept multivariate time-series data as input, I do not believe they accept integer data. What would be the best practice in this case? I am currently considering converting the integer data time-series by repeating the integer value for the same number of increments as the other multivariate inputs, then applying random noise to the manufactured integer data that I converted to Timeseries- Can anyone comment on whether this is an appropriate method or if there are other more appropriate ways to approach the issue? Thanks in advance.'

$\endgroup$
1
  • 1
    $\begingroup$ Hi, please include a sample of your data and the approach you have tried. This will help the community in answering your question. $\endgroup$ May 6 at 6:11
0
$\begingroup$

You need two backbones: for time series, use LSTM/GRU and for non-time series, use 1D CNN or Linear layers. Once you get the final embeddings from both of them, concatenate the embeddings and finally feed to the classification layer.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.