A Pattern Mining Approach for Classifying Multivariate Temporal Data

TLDR

Learning classification models from complex multivariate temporal data in electronic health records is challenging because temporal pattern mining generates many patterns, most of which are irrelevant to the classification task. The study aims to define effective temporal features by introducing a minimal predictive temporal patterns framework that selects a small set of predictive, non‑spurious patterns. The method uses temporal abstractions and pattern mining to extract classification features and applies this framework to predict patients at risk of heparin‑induced thrombocytopenia. Results show that this approach yields more accurate classifiers, advancing the development of intelligent clinical monitoring systems.

Abstract

We study the problem of learning classification models from complex multivariate temporal data encountered in electronic health record systems. The challenge is to define a good set of features that are able to represent well the temporal aspect of the data. Our method relies on temporal abstractions and temporal pattern mining to extract the classification features. Temporal pattern mining usually returns a large number of temporal patterns, most of which may be irrelevant to the classification task. To address this problem, we present the minimal predictive temporal patterns framework to generate a small set of predictive and non-spurious patterns. We apply our approach to the real-world clinical task of predicting patients who are at risk of developing heparin induced thrombocytopenia. The results demonstrate the benefit of our approach in learning accurate classifiers, which is a key step for developing intelligent clinical monitoring systems.