A Cuckoo search-based optimized ensemble model (CSOEM) for the analysis of human gait

The use of mobile devices has contributed greatly to the expansion of Human Activity Recognition (HAR) studies in recent years. Researchers find it attractive because of its versatility, low cost, compact size, ease of usage, and wide range of possible applications. Conventional, biological, and control-based systems are just some of the methods that have been created for humanoid robot movement in recent years. This article specifically targeted improvement in the proposed method, which is different from previous papers. This is being done with the use of the publicly available Human Activity Gait (HAG) data set, which documents a wide range of different types of activities. IMU sensors were used to collect this data set. Several experiments were conducted using different machine-learning strategies, each with its own set of hyper-parameters, to determine how best to utilize these data. In our proposed model Cuckoo Search Optimization is being used for optimum feature selection. On this data set, we have tested a number of machine learning models, including LR, KNN, DT, and proposed CSOEM (Cuckoo Search-Based Optimized Ensemble Model). The simulation suggests that the proposed model CSOEM achieves an impressive accuracy of 98%. This CSOEM is built by combining the feature selection strategy of Cuckoo Search Optimizations with the ensembling of the LR, KNN, and DT.