A novel approach to predict diabetes by Cascading Clustering and Classification

Knowledge of incidence and prevalence of a disease is vital in Community Medicine to control a disease. It is important in Internal Medicine for clinical diagnosis and presumptive treatment on a probability model. Prevalence informs the total case load at a given time. Incidence yields a pointer to extent of attention required and choice of measures. Initially K-means clustering is used to group the disease related data into clusters and assigns classes to clusters. Subsequently multiple different classification algorithms are trained on the result set to build the final classifier model based on K-fold cross validation method. This methodology is evaluated using 768 raw diabetes data obtained from a city hospital. The best accuracy for the given dataset is achieved in bagging algorithm compared to other classifiers. The proposed approach helps doctors in their diagnosis decisions and also in their treatment planning procedures for different categories.