Classification method using fuzzy level set subgrouping

We present a new classification system which is based on fuzzy level sets subgrouping. This new classification system allows a fast classification method with quite accurate results. Classification runs were carried out with four different data sets. All four data sets were related to medical diagnostics. Data sets were related to thyroid and diabetes diagnostics, echocardiogram data relates to predicting patients chances of survival after acute heart attack. With lymphography data there are four classes to predict, normal find, metastases, malign lymph and fibrosis, from the existing data. Classification results are compared to some existing results in the literature and the results seem to compare well.