Classification of wisconsin breast cancer database

Statistical data analysis includes developing methods for classification of variousdatabases. These databases may have data about lots of sectors for example financial,industrial, food, biomedical or etc. The main aim is to get a result by making aclassification for a product, patient or something.In this study, we used classification methods for biomedical analysis. Our sampledatabase has breast cancer data which is one of the most cause of cancer, becausedetection of this cancer is very important. Database has 9 attributes which is used forclassification. These attributes are numerical numbers. Explanations of attributes aregiven in detail in following chapters. After having numerical numbers via FNAprocedure, class labels can be given both by classical examinations and byclassification algorithms. Our database includes a class column which real diagnosisexists. This study aims to consider classification algorithms results carefully.Different methods and algorithms have been used; classification accuracies havebeen given depending on real values. Results are compared and some ideas can arisefor using software programs to classify sickness instances. Computer supporteddiagnosis can be used more commonly.Keywords: Biomedical data classification, KNN rule, linear discriminant, neuralnetwork, support vector machine, breast cancer.