Statistical Analysis Method And It’s Applications For The Multi-response Categorical Data

This article studies the statistical analysis method for multi-response categorical data andit’s applications.Nowadays, categorical data is a common data form in field investigation and clinical re-search. it characterized by a certain correlation between the response variable in the data, inde-pendently analysis the response variable was not appropriate. Currently, methods of statisticalanalysis on such data are few, but there are many real cases. Therefore, research on statisticalmethods of the multi-response variables categorical data has the most important significanceboth on theory and practice.The data preprocessing in the case study use SAS software, and Generalized EstimatingEquations and Multi-level Logistic Regression Model also use SAS software, and PrincipalComponent Analysis of disaggregated data use SPSS.In this article, to analyze categorical data on multi-response variables, as the issues ofthe multivariate logistic regression. Firstly, parameter estimation and testing methods of themodel are given out by using generalized estimating equations and multi-level logistic modelfor the data’s statistical analysis. On this basis, some more in-depth results are given out byusing the principal component analysis method of disaggregated data to conduct a more in-depth discussion. In the end, we give some application examples for the methods.The main conclusions of this paper as follows: Generalized Estimating Equations andMulti-level Logistic Regression Model fully considered the correlation between more than onedependent variable, it provides a wealth of information close to the actual situation, the ad-vantage of Principal Component Analysis of categorical data can handle all types of data, themethod is more conducive to the depth of the complex correlation between the multiple responseset, and try to ensure the accuracy of the results.