基于PCA-XGBoost联合算法的盗窃犯罪预测模型

PREDICTION MODEL OF THEFT BASED ON PCA-XGBOOST JOINT ALGORITHM

结合大量犯罪数据特征和行为特征,提出一种PCA-XGBoost联合预测模型。采用PCA算法提取数据集的主要特征;应用XGBoost算法提升预测优化和泛化能力,并通过三种检验方法进行准确率检验。此外,经与XGBoost、CART、RF、NB和LR等分类算法模型的预测结果进行对比,表明PCA-XGBoost联合预测模型对盗窃犯罪数量的预测准确度明显高于其他预测模型,具有较高的应用价值。

Combining with a large number of crime data features and behavioral features,this paper proposes a PCA XGBoost joint forecasting model.PCA algorithm was used to extract the main features of the data set,XGBoost algorithm was used to improve the prediction optimization and generalization ability,and the accuracy was tested by three test methods.Compared with the prediction results of XGBoostt,CART,RF,NB,LR classification algorithm models,the results show that the prediction accuracy of PCA-XGBoost joint prediction model on the number of theft is significantly higher than other prediction models,and has higher application value.