Study on the Quantitative Assessment Method of Fall Risk Based on Multidimensional Gait Features

This study developed a new model for predicting fall risk, employing the Immune Particle Swarm Optimization (IPSO) algorithm to optimize the LightGBM algorithm, with gait feature data as input for predicting fall risk values and calculating the Mean Square Error (MSE). This method aims to enhance the accuracy of fall risk prediction. According to the experimental results, the use of the IPSO-optimized LightGBM algorithm model significantly improved the prediction accuracy for fall risk and demonstrated an excellent model fitting capability.