tsfresh - extract time series features with ease- 1 min
Have you ever worked with time series data or sequences?
Think of server logs, medical data of EEGs, production data in Industrie 4.0 environments, …
Extracting features and finding patterns to be used for classification or regression can be a quite challenging and time consuming task. This is especially true, as this needs much manual work from experts with a lot of domain knowledge.
Our developed package tsfresh frees your time spend on feature extraction by using a large catalog of automatically extracted features, known to be useful in time series machine learning tasks. Additionally, it can rank them by their significance and throw out features without useful information.
Using tsfresh, you can automize your time series classification, find anomalies in your timely annotated data easily and also have some fun with financial data…
I will cover all this from a data science as well as a python developer point of view in this blog in the next time. Stay tuned