Research on Human Activity Behavior Recognition Technology Based on LGBM Model

Human activity behavior recognition technology is widely used in medical health, fall detection and exercise training. This paper constructs a model to solve the problem of classifying human activity behavior recognition by miniature inertial sensors at different positions of the body. First, after data processing and analysis, we constructed the LGBM model. Later, we verified the ability to generalize the excellent results of the model to identify the input data by adjusting the training data. Finally, we study and address the problem of possible overfitting of the model.