Amortized Neural Clustering of Time Series based on Statistical Features

2026-05-13

AI Summary

This work addresses the limitations of traditional time series clustering methods, which rely on predefined algorithms such as K-means and require manual specification of the number of clusters, thereby struggling to adapt to complex data structures. The authors propose an algorithm-agnostic neural clustering framework that leverages amortized inference to learn clustering rules from simulated data based on statistical features—including autocorrelation and quantile autocorrelation—automatically inferring the number of clusters and constructing affinity structures without explicitly defining cluster shapes or relying on manual hyperparameter tuning. Experimental results demonstrate that the proposed method achieves clustering accuracy comparable to or better than conventional approaches that assume knowledge of the true number of clusters, and it exhibits strong practical utility in real-world applications such as stock return time series analysis.

Abstract

This paper introduces an algorithm-agnostic approach to feature-based time series clustering via amortized neural inference. By training neural networks to approximate the optimal partitioning rule from simulated data, the proposed framework reduces reliance on conventional clustering methods, such as $K$-means, $K$-medoids, or hierarchical clustering, and their associated objective functions and heuristics. Leveraging statistical features, such as autocorrelations and quantile autocorrelations, the approach learns a data-driven affinity structure from which clustering partitions can be recovered, without requiring explicit prior specification of cluster shapes or structures. In addition, one version of the method can automatically determine the number of clusters, avoiding ad-hoc selection procedures. Comprehensive empirical studies show that the proposed framework achieves competitive or superior clustering accuracy relative to traditional methods, even in challenging scenarios where competing techniques are provided with the true number of clusters. An application to financial time series of stock returns illustrates its practical utility. By reducing the need for algorithm selection and calibration, the proposed framework opens new possibilities for automated, adaptive, and data-driven clustering of temporal data across scientific and industrial domains.

Problem

• time series clustering
• statistical features
• cluster number estimation
• algorithm-agnostic clustering
• data-driven partitioning

Innovation

• amortized neural inference
• feature-based clustering
• statistical features
• automatic cluster number selection
• time series clustering