Actual laboratory data analysis with data mining techniques

Data mining techniques are widely used in studies where vast amount of data is being produced. Since it is possible that human perception can miss some details about the diagnose process, data mining techniques can aid to make early diagnosis and cure diseases. In this research, models has been developed in order to determine parameters which has significant roles in diagnosing anemia and anemia types, with laboratory data which consists of 6 months of measurements. A data mining research in the medicine sector has been made in order not to miss the diagnosis in the plentifulness of medical measurements. Open source data mining software (WEKA) has been used in modeling process. Classification algorithms were applied to the data and it has been seen that rules.PART and trees.J48 algorithms had the highest classification success. In clustering analyses it has been seen that FilteredClusterer algorithm has 100% concurrency with data set. Association rules has been produced with Tertius algorithm and they have been verified by a specialist. It has been concluded that FilteredAssociator association rules and rules from decisions trees can't be used practically in diagnose processes.