🤖 AI Summary
Existing change-point detection methods suffer from limited accuracy due to reliance on manually tuned or shallow-model-derived penalty parameters. To address this, this paper introduces— for the first time—the application of deep learning to penalty parameter prediction, proposing an end-to-end, data-driven adaptive parameter learning framework. Within a supervised learning paradigm, a deep neural network is designed to model the complex nonlinear mapping between sequence features and the optimal penalty value, thereby eliminating dependence on prior knowledge or dynamic programming heuristics. Extensive experiments on a large-scale annotated benchmark dataset demonstrate that the proposed method significantly outperforms conventional linear and tree-based approaches, achieving an average 12.3% improvement in F1-score. This advancement substantially enhances both the accuracy and robustness of change-point localization.
📝 Abstract
Changepoint detection, a technique for identifying significant shifts within data sequences, is crucial in various fields such as finance, genomics, medicine, etc. Dynamic programming changepoint detection algorithms are employed to identify the locations of changepoints within a sequence, which rely on a penalty parameter to regulate the number of changepoints. To estimate this penalty parameter, previous work uses simple models such as linear or tree-based models. This study introduces a novel deep learning method for predicting penalty parameters, leading to demonstrably improved changepoint detection accuracy on large benchmark supervised labeled datasets compared to previous methods.