Figure 7 shows system overview to realize the implementation model shown in Fig. 6. This system consists of a database with 6 tables and two file groups. A database is accessed with SqlTool and Web application. Both file groups are accessed with Web application. One file group of “Drug Candidate” is also accessed with Linux command. It is freely accessed to search for new interaction pairs by Gene ID and chemicalID. Using this database, the clustering is created by statistical regression analysis and discriminant analysis, and SVM creates the learning model with the clustering. When the data is input in incidence prediction, its incidence is calculated with those learning model files using SVM.