Classification prediction of the foot disease pattern using decision tree model

Datamining is used to find out desired important and meaningful knowledge in large scale data. The decision tree in classification algorithms has been applied to categorical attributes and numeric attributes in different domains. The purpose of study was to acquire significant information between singular disease groups and biomechanical parameters related with symptoms by developing prediction model. Sample data of 90 patient’s records diagnosed with a singular disease was selected for analysis, in total 2418 data. A dependent variable was composed of 9 singular disease groups. 18 of 32 independent variables closely related to disease were selected and optimized. After object data was divided into training data and test data, C5.0 algorithm was applied for analysis. In conclusion, 10 diagnosis rules were created and major symptom information was verified. On the basis of the study, additional analysis with utilizing other datamining methods will be performed to improve accuracy from now on.