Penalty Learning for Optimal Partitioning using Multilayer Perceptron

TL;DR

This paper introduces a multilayer perceptron to predict penalties in changepoint detection, improving accuracy over traditional linear and tree-based models, especially in genomic data analysis.

Contribution

It proposes a novel MLP-based approach for predicting penalties in optimal partitioning, enhancing non-linear modeling capabilities and continuous output predictions.

Findings

Improved accuracy and F1 score on genomic datasets

Better handling of non-linearity compared to linear models

Enhanced predictive performance over tree-based models

Abstract

Changepoint detection is a technique used to identify significant shifts in sequences and is widely used in fields such as finance, genomics, and medicine. To identify the changepoints, dynamic programming (DP) algorithms, particularly Optimal Partitioning (OP) family, are widely used. To control the changepoints count, these algorithms use a fixed penalty to penalize the changepoints presence. To predict the optimal value of that penalty, existing methods used simple models such as linear or tree-based, which may limit predictive performance. To address this issue, this study proposes using a multilayer perceptron (MLP) with a ReLU activation function to predict the penalty. The proposed model generates continuous predictions -- as opposed to the stepwise ones in tree-based models -- and handles non-linearity better than linear models. Experiments on large benchmark genomic datasets demonstrate that the proposed model improves accuracy and F1 score compared to existing models.