基于改进PSO-BP神经网络算法在一般盗窃犯罪预测中的应用

针对BP神经网络对初始权重敏感,容易陷入局部最优解的问题,引入粒子群优化算法(PSO),对网络权重进行全局搜索,同时采用BP神经网络权重更新方法对PSO搜索到的权重和阈值进行进一步的更新,构建改进后的PSO-BP神经网络模型,对一般盗窃犯罪数量进行预测。应用美国芝加哥市2015年-2017年盗窃犯罪数据以及总人口数、房价中位数、本科率等11个影响因子数据,对改进前后的模型进行了预测对比实验。结果表明,改进后的PSO-BP神经网络模型成功克服了BP模型的缺陷,相对误差由4.68%降低到1.635%。

APPLICATION OF IMPROVED PSO-BP NEURAL NETWORK ALGORITHM IN THE PREDICTION OF THEFT CRIME

Aiming at the problem that BP neural network is sensitive to the initial weight and easy to fall into local optimal solution,we introduce PSO to search the network weight globally,and BP neural network weight update method was used to further update the weights and thresholds of PSO.We constructed an improved PSO-BP neural network model to predict the number of general theft crime.We applied the data of theft crime in Chicago in the United States from 2015 to 2017,as well as the data of total population,median housing price and undergraduate rate and other 11 influencing factors,and conducted a predictive comparison experiment on the model before and after improvement.The result shows that the improved PSO-BP neural network model successfully overcomes the defects of the BP model,and the relative error is reduced from 4.68%to 1.635%.