A new classification algorithm combining Choquet integral and logistic regression

Logistic regression algorithm and SVM algorithm are two well-known classification algorithms but when the multi-collinearity between independent variables occurs in above two algorithms, their classifying performance will always be not good. Due to this reason, we firstly proposed a pared-down MLE method in this study to improve the logistic regression algorithm for no needing to group the original data. Secondly, we proposed a novel classification algorithm combining the Choquet integral with respect to the lambda-measure based on gamma-support proposed by our previous work and the improved logistic regression algorithm to further improve the above situation. For evaluating the performances of the SVM, logistic regression and our new algorithm with gamma-support based on lambda-measure and P-support respectively, a real data experiment by using leave-one-out cross-validation accuracy is conducted. Experimental result shows that the proposed classification algorithm combining Choquet integral regression model with gamma-support based on lambda-measure has the best performance.