Linear classification techniques in statistical learning and an application to diabetes data

In parallel with current technological developments, interest in topics such as data mining, statistical learning and machine learning is gradually increasing. Statistical learning plays an important role particularly in areas such as engineering and health sciences with finance and industry. In this study, firstly general information about statistical learning theory is given and then the theoretical information about discriminant analysis and logistic regression analysis which are the statistical classification methods are examined in detail. It is aimed to compare the classification results of the models obtained using these methods. It can be interpreted how the changes in the independent variables affect the response variable by means of these models. Data set of 768 Pima Native female individuals aged 21-81 years are used in the application which is performed for the purpose of the study. The classification successes of the models for discriminant and logistic regression analysis, which are set up for this data set, are evaluated and the results are interpreted. Keywords: Statistical learning, Discriminant analysis, Logistic regression analysis, Classification