The Establishment And Evaluation Of Logistic Regression Analysis Model In Excel

Objective This study was determined to form a logistic regression analysis model for pariedand grouped data with Excel in order to provide an easy-to-learn new calculating method forstatistical analysis.Methods First, according to the mathematic principle of logistic regression, chose variableswith forward stepwise regression, then used newton's method to get the function parameters ofmaximum likelihood ratio. After that, the programming and trial was done in Excel using IF andCount Statements, ABS Functions, Matrix Disposal, Logic analysis and Iterative Calcualtion etcand the logistic regression model was also established.Results The result showed that the screening factors of the eastalished logistic regressionanalysis model for paied data and grouped data in Excel were consistent with the one in SPSS.The absolute error of the comparision among parameters was no more than 0.0078. The relativeerror was no more than 0.60%.Conclusions The established logistic regression model in Excel could be used in logisticregression analysis for 1～3 factors of paied data and grouped data. The model was characteristicof simple, easy-to-learn and practical.