Dataset for Human Activity Recognition Using Wearable Motion Sensors and LSTM-Based Classification

This repository contains wearable inertial sensor recordings collected during multiple activities of daily living and exported from the Xsens software environment. The dataset includes raw XLSX recordings, activity annotation metadata, and a sample Python analysis script implementing leave-one-subject-out evaluation using classical machine learning methods and LSTM-based temporal neural networks. The provided script supports feature extraction, temporal sequence generation, baseline benchmarking (Logistic Regression, Random Forest, and SVM), and confusion-matrix generation for reproducible human activity recognition experiments.