Random Forests for Diabetes Diagnosis

Random forest is one of the most recent successful research findings for decision tree learning. It is widely used in the medical field, particularly for diabetes diagnosis. Diabetes is reaching epidemic proportions in many developing and newly industrialized countries. Thus, random forest should be exploited to treat diabetes diagnosis. In this paper, we exploit the principle of random forests for the implementation of a powerful model for the diagnosis of diabetes. The experiments were carried out on the Pima Indians Diabetes data set selected from the UCI repository. First, several random forests have been developed with different numbers of trees in order to define the optimum size of the forest. Then, random forests were compared with other machine learning methods. The results of the experiments show that our approach based on random forest has proved to be more efficient in comparison with other methods of machine learning.