Performance Comparison of Data Mining Methods in Diagnosis of Complex Diseases

The data used in Genome Wide Association studies is vast in amount and high dimensional. Therefore, different data mining methods are used in order to find the relations between profiles and diseases. These methods are then used for diagnostic models. In this study two different data sets were used. The melonoma data set consists of 1025 cases and 531 controls. The multi ethnic prostate cancer data set consists of 2325 cases and 2350 controls. The underlying SNPs were searched by different data mining methods such as Decision Trees, Naive Bayes and Support Vector Machines. For both diseases support vector machine presented the best performance results. This method presented 75.68% of accuracy for prostate cancer data and 78.6% of accuracy for melonoma.