Question 1

How do I convert my time series data to tslearn format?

Accepted Answer

Use the to_time_series_dataset utility function from tslearn.utils, or convert from other toolkits like pandas as described in the integration guide. The data must be a 3D numpy array with dimensions (n_ts, max_sz, d).

Question 2

tslearn vs tsfresh for time series analysis?

Accepted Answer

tslearn focuses on end-to-end machine learning algorithms like classification and clustering, while tsfresh is specialized for extracting features from time series to use with general ML models. Choose tslearn for direct modeling and tsfresh for feature engineering.

Question 3

Can tslearn handle multivariate time series?

Accepted Answer

Yes, tslearn supports multivariate time series through the 3D array format where the third dimension represents the number of dimensions, allowing for analysis of multi-channel data as shown in the examples.

Question 4

How to perform time series clustering with tslearn?

Accepted Answer

Use algorithms like TimeSeriesKMeans or KShape from the clustering module. First, preprocess your data with scaling or resampling, then fit the model similar to scikit-learn, as demonstrated in the getting started examples.

Question 5

What distance metrics does tslearn support for similarity measurement?

Accepted Answer

tslearn includes metrics like Dynamic Time Warping (DTW) and Global Alignment Kernel (GAK) for measuring similarity between time series, which are essential for tasks like nearest neighbors classification and clustering.

Question 6

Is tslearn good for time series forecasting?

Accepted Answer

tslearn offers regression models like KNN regressor and MLP for predicting continuous values, which can be adapted for forecasting, but it lacks specialized forecasting models like ARIMA, so it may not be the best for pure prediction tasks.

tslearn

What is tslearn?

Overview

Use Cases

Best For

Related Projects

Found a gem we're missing?

Not Ideal For

Pros & Cons

Pros

Cons

Frequently Asked Questions