Application of PCA-LSTM model in human behavior recognition

Abstract

For the problem that the recognition rate of human upper extremity movements is not high and existing models are prone to "dimensional disaster", a pattern recognition algorithm based on PCA-LSTM neural network is studied. Collect and process the surface electromyography signal (sEMG) of the human upper limbs, and put it into the PCA model for data dimensionality reduction, put the dimensionality-reduced data into the LSTM neural network model to classify the human behavior, count the classification efficiency and recognition rate.Comparing PCA-LSTM with the traditional classification algorithm SVM and random forest, the calculation efficiency is improved by 31% and 53%, the recognition rate is increased by 3.9% and 4.6% and the classification effect is significant.