Logistic regression analysis and machine learning applications

This study, prepared under the title of "Logistic Regression and Machine Learning Applications", consists of five parts. In the first part the purpose of the thesis and the literature research is included. The statements that is namely a random variable that is required to generate binomial and multivariate logistic regression models, probability distribution, distribution function, probability density function, binomial distribution, maximum likelihood estimation, Newton Raphson Method, odss ratio, creating model and model evaluation are included in the second part. In the third part the expression regression analysis is given and binary and multiple logistic regression are obtained by using mathematical proofs. In tha fourth part binary logistic regression model for the first data set and multivariate logistic regression model for the second data set was built. The β coefficients of created models are calculated algebraically and also the written matlab code in two different ways after then the model was assessed success of the models was measured. In conclusion in the fifth part it is shown that the β coefficients in the logistic regression models created or the data sets in the fourth part are the same values as a result of both algebraic operations and running matlab code. The average successes of two models are shown for two models in consequence of k-fold cross validation and comments were made about the model successes. Key Words: Machine Learning, Supervised Learning, Classification Algorithms Maximum Likelihood Estimation, Logistic Regression