Fuzzy Cognitive Maps and Hidden Markov Models: Comparative Analysis of Efficiency within the Confines of the Time Series Classification Task

TL;DR

This paper compares the efficiency of Fuzzy Cognitive Maps and Hidden Markov Models in time series classification, highlighting the impact of model configuration and dataset dependence on accuracy.

Contribution

It introduces a comparative analysis of HMM and FCM models with different configurations for time series classification.

Findings

One-model-per-series approach outperforms other configurations.

Model choice effectiveness depends on dataset characteristics.

Hyperparameters significantly influence classification accuracy.

Abstract

Time series classification is one of the very popular machine learning tasks. In this paper, we explore the application of Hidden Markov Model (HMM) for time series classification. We distinguish between two modes of HMM application. The first, in which a single model is built for each class. The second, in which one HMM is built for each time series. We then transfer both approaches for classifier construction to the domain of Fuzzy Cognitive Maps. The identified four models, HMM NN (HMM, one per series), HMM 1C (HMM, one per class), FCM NN, and FCM 1C are then studied in a series of experiments. We compare the performance of different models and investigate the impact of their hyperparameters on the time series classification accuracy. The empirical evaluation shows a clear advantage of the one-model-per-series approach. The results show that the choice between HMM and FCM should be dataset-dependent.