Learning Penalty for Optimal Partitioning via Automatic Feature Extraction

TL;DR

This paper introduces a novel method using recurrent networks to automatically learn the penalty parameter for optimal changepoint detection from raw data, improving accuracy over traditional feature-based approaches.

Contribution

It proposes a new approach that leverages recurrent neural networks to learn the penalty parameter directly from raw sequences, eliminating manual feature extraction.

Findings

Outperforms traditional methods in genomic datasets

Improves changepoint detection accuracy

Demonstrates effectiveness across multiple datasets

Abstract

Changepoint detection identifies significant shifts in data sequences, making it important in areas like finance, genetics, and healthcare. The Optimal Partitioning algorithms efficiently detect these changes, using a penalty parameter to limit the changepoints count. Determining the optimal value for this penalty can be challenging. Traditionally, this process involved manually extracting statistical features, such as sequence length or variance to make the prediction. This study proposes a novel approach that uses recurrent networks to learn this penalty directly from raw sequences by automatically extracting features. Experiments conducted on 20 benchmark genomic datasets show that this novel method generally outperforms traditional ones in changepoint detection accuracy.