MINIROCKET: A Very Fast (Almost) Deterministic Transform for Time Series Classification

TL;DR

MINIROCKET is a highly efficient, nearly deterministic variant of ROCKET that drastically reduces training time for time series classification while maintaining state-of-the-art accuracy, enabling rapid analysis of large datasets.

Contribution

This paper introduces MINIROCKET, a faster, nearly deterministic reformulation of ROCKET that preserves accuracy and significantly reduces computational costs.

Findings

MINIROCKET is up to 75 times faster than ROCKET on large datasets.

It maintains the same accuracy as ROCKET across 109 datasets.

MINIROCKET outperforms other methods with similar computational expense.

Abstract

Until recently, the most accurate methods for time series classification were limited by high computational complexity. ROCKET achieves state-of-the-art accuracy with a fraction of the computational expense of most existing methods by transforming input time series using random convolutional kernels, and using the transformed features to train a linear classifier. We reformulate ROCKET into a new method, MINIROCKET, making it up to 75 times faster on larger datasets, and making it almost deterministic (and optionally, with additional computational expense, fully deterministic), while maintaining essentially the same accuracy. Using this method, it is possible to train and test a classifier on all of 109 datasets from the UCR archive to state-of-the-art accuracy in less than 10 minutes. MINIROCKET is significantly faster than any other method of comparable accuracy (including ROCKET), and significantly more accurate than any other method of even roughly-similar computational expense. As such, we suggest that MINIROCKET should now be considered and used as the default variant of ROCKET.