Application Of Regularized Discriminant Analysis To The Social Security Of Residents

In recent years,with the rapid development of information technology and data processing tools,data mining technology is becoming more and more mature.Because of the wide application of discriminant analysis in machine learning,pattern recognition and so on,it has become one of the hot topics in data analysis.Bayesian discriminant analysis is one of the most commonly used classification methods because of its lower misclassification rate.Naive Bayes discriminant analysis is the most commonly used classification models.Due to the assumption of attribute independence of Naive Bayes discriminant analysis,there are some limitations in the actual data analysis.Based on the Naive Bayes discriminant analysis model,this paper improves the algorithm and uses the regularized discriminant analysis to deal with the problem of multi-colinearity between the independent variables.The empirical data of this paper is the analysis of the problem of residents' social security.Because the problem of social security is closely related to the vital interests of the residents,it is of great practical significance to analyze the problems of the residents' social security.At the same time,the residents' social security problems can be identified and analyzed,which can provide favorable information for the residents,the government and the relevant industry organizations,to a certain extent,improve the living standard of the people's livelihood.In this paper,we use both Naive Bayes discriminant analysis and regularized discriminant analysis model to analyze the empirical data,and compare the advantages and disadvantages of the classifier performance evaluation.The results show that the regularized discriminant analysis has higher prediction accuracy while dealing with the multi-colinearity problem.