Short-time Classification of Neurodegenerative Diseases Based on Cross-Recurrence Quantification Analysis and Statistical Features

Neurodegenerative diseases refer to progressive disorders of the nervous system that impair motor functions. Using machine learning techniques to analyze and classify gait data can lead to early diagnosis and better management of treatments. This paper aims to classify neurodegenerative patients and healthy individuals by analyzing a three-second gait signal using a combination of effective features from cross-recurrence quantification analysis (CRQA) and statistical analysis. The dataset includes force signals from the left and right feet of ۱۶ healthy individuals (HC), ۱۳ with amyotrophic lateral sclerosis (ALS), ۱۵ with Parkinson's disease (PD), and ۲۰ with Huntington's disease (HD). The CRQA features extracted include recurrence rate, determinism, averaged diagonal length, length of longest diagonal length, entropy of diagonal length, laminarity, trapping time, length of longest vertical line, recurrence time of ۱st type, recurrence time of ۲nd type, recurrence period density entropy, clustering coefficient, and transitivity. Statistical features include mean, variance, skewness, and kurtosis. A sequential feature selection algorithm was used to select effective features. The classification accuracy for the Ensemble (Bagged Trees) classifier was obtained using ۱۰-fold cross-validation for the groups HC vs. PD, HC vs. HD, HC vs. ALS, ALS vs. PD, ALS vs. HD, PD vs. HD, NDD vs. HC, and ALS vs. PD vs. HD vs. HC, with the respective accuracy values of ۹۸.۳%, ۹۴.۸%, ۹۷.۷%, ۹۸.۲%, ۹۸.۴%, ۹۵%, ۹۴%, and ۹۳.۵%. The results indicate that the effective fusion of features and the ability of cross-recurrence quantification analysis to quantify the synergistic relationship of the dynamic movements of the left and right feet provide an effective means of diagnosing diseases during the short ۳-second walking period.

کلیدواژه ها:

Cross-Recurrence Quantification Analysis ، Statistical Features ، Force Signal ، Gait ، Classification