Research on obesity level prediction by the PSO-BP algorithm based on principal component analysis
An obesity level prediction model combining principal component analysis(PCA)and particle swarm optimization BP neural network algorithm(PSO-BP)is proposed.Dimension reduction of the 16 input variables are performed by principal com-ponent analysis,and 11 comprehensive variables which represented principal components are extracted as the input of BP neural network,optimize the weights and thresholds of network by using the PSO algorithm,further improve the ability of network training.Experimental results show that the PSO-BP algorithm based on PCA achieves higher classification prediction power at a shorter time cost.This study provides a scientific basis for the evaluation of individual obesity level,and has important research signifi-cance and application value.
万方数据
