Combined Utilization of STFT, CWT & DFA in Classification of Young & Old Gait Using Machine Learning Techniques

Numerous studies on the stride intervals of young and elderly individuals indicate that elderly individuals tend to have slower stride intervals, take shorter steps, and exhibit longer stride intervals. Research on these differences is crucial as it helps predict fall risk in the elderly based on their stride intervals, potentially reducing injuries and gait-related accidents. A classification system for analyzing the gait of young and elderly individuals through stride intervals using various machine learning techniques can provide a smart approach to preventing falls. The features of stride intervals can be extracted using techniques like Short-Time Fourier Transform (STFT), Continuous Wavelet Transform (CWT), and Detrended Fluctuation Analysis (DFA), each with its own advantages and disadvantages. This paper explores the combined utilization of STFT, CWT, and DFA. Features are collected from the stride intervals of young and elderly individuals and analyzed using DFA, STFT, and CWT. These features are then compiled and fed into Random Forest, Support Vector Machine (SVM), and K-Nearest Neighbor (KNN) classifiers for stride interval classification. The machine learning techniques are compared in terms of precision, recall, and F1-score. The results demonstrate that the combined use of STFT, CWT, and DFA can achieve 100% precision, recall, and F1-score when using SVM, with the shortest processing time of 2.6902 seconds compared to other machine learning techniques utilized in this paper. This paper contributes by proving that the combined use of STFT, CWT, and DFA can help overcome the limitations of each individual technique.