TSFEDL: A Python Library for Time Series Spatio-Temporal Feature Extraction and Prediction using Deep Learning (with Appendices on Detailed Network Architectures and Experimental Cases of Study)

TL;DR

TSFEDL is a Python library that consolidates 20 deep learning methods for extracting spatio-temporal features and predicting time series data, validated through performance tests.

Contribution

It introduces a comprehensive Python library integrating state-of-the-art deep learning models for time series analysis, with detailed architectures and experimental validation.

Findings

Library includes 20 advanced methods for feature extraction and prediction.

Performance validation confirms the effectiveness of the included architectures.

Provides a versatile tool for various time series data mining tasks.

Abstract

The combination of convolutional and recurrent neural networks is a promising framework that allows the extraction of high-quality spatio-temporal features together with its temporal dependencies, which is key for time series prediction problems such as forecasting, classification or anomaly detection, amongst others. In this paper, the TSFEDL library is introduced. It compiles 20 state-of-the-art methods for both time series feature extraction and prediction, employing convolutional and recurrent deep neural networks for its use in several data mining tasks. The library is built upon a set of Tensorflow+Keras and PyTorch modules under the AGPLv3 license. The performance validation of the architectures included in this proposal confirms the usefulness of this Python package.