tsuniverse 0.0.10

A library for efficiently processing a time series universe to determine causal features.

tsuniverse

A library for efficiently processing a time series universe to determine causal features.

Dependencies :globe_with_meridians:

Python 3.11.6:

• pandas
• pyarrow
• scikit-learn
• timeseries-features
• scipy
• dcor
• numpy
• pyHSICLasso

Raison D'être :thought_balloon:

tsuniverse aims take a universe of time series and figure out features from that universe that can be used to predict a single time series.

Architecture :triangular_ruler:

tsuniverse is a functional library, meaning that each phase of the feature extraction goes through functions without side-effects. It attempts to do as much multiprocessing as it can to make this process quicker. Each feature extraction is done in different phases, those phases are:

1. Pearson Correlations.
2. Mutual Information.
3. Spearmans Rho.
4. Kendalls Tau.
5. Distance Correlations.
6. HSIC.

Installation :inbox_tray:

This is a python package hosted on pypi, so to install simply run the following command:

pip install tsuniverse

or install using this local repository:

python setup.py install --old-and-unmanageable

Usage example :eyes:

The use of tsuniverse is entirely through code due to it being a library. It attempts to hide most of its complexity from the user, so it only has a few functions of relevance in its outward API.

Generating Features

To generate features:

import datetime

import pandas as pd

from tsuniverse.process import process

df = ... # Your timeseries dataframe
features = process(df)

This will produce a list of features that you can produce with timeseries-features.

License :memo:

The project is available under the MIT License.