Data classification is the process that finds the common characteristics among a set of objects in a database and classifies them into different classes according to a classification model. Up to the present, the development of classification has made great achievements and various kinds of classification technologies and theories continue to emerge. It has been applied to many areas, such as fraud detection, target marketing, weather forecast, performance prediction, manufacturing, and medical diagnosis. Classification approaches mainly include decision tree, neural network, Bayesian classification, case-based reasoning, genetic algorithms and rough sets method, in which decision tree is one of the most cited approaches. This study describes a model that discovers the fuzzy decision trees from medical databases which are Harbeman’s Survival database and Blood Transfusion Service Center database. The former is to help doctors treat and diagnose a group of patients who show similar prehistoric medical symptoms and the latter is to advise someone the time he/she should donate blood. The proposed data mining procedure consists of two modules. One is a clustering module that is based on a neural network, Adaptive Resonance Theory 2 (ART2), which performs affinity grouping tasks on a large amount of medical records. The other module employs the fuzzy set theory to extract fuzzy decision trees for each homogeneous cluster of data records. Finally, experiments were done to cluster data from two given databases. Simulation results verify the important roles of fuzzy ART2 in finding decision tree rules. Besides, the number of erroneously clustered patterns as well as the computational cost from decision trees can be significantly reduced as compared with the case without using fuzzy ART2.
Data classification is the process that finds the common characteristics among a set of objects in a database and classifies them into different classes according to a classification model. Up to the present, the development of classification has made great achievements and various kinds of classification technologies and theories continue to emerge. It has been applied to many areas, such as fraud detection, target marketing, weather forecast, performance prediction, manufacturing, and medical diagnosis. Classification approaches mainly include decision tree, neural network, Bayesian classification, case-based reasoning, genetic algorithms and rough sets method, in which decision tree is one of the most cited approaches. This study describes a model that discovers the fuzzy decision trees from medical databases which are Harbeman’s Survival database and Blood Transfusion Service Center database. The former is to help doctors treat and diagnose a group of patients who show similar prehistoric medical symptoms and the latter is to advise someone the time he/she should donate blood. The proposed data mining procedure consists of two modules. One is a clustering module that is based on a neural network, Adaptive Resonance Theory 2 (ART2), which performs affinity grouping tasks on a large amount of medical records. The other module employs the fuzzy set theory to extract fuzzy decision trees for each homogeneous cluster of data records. Finally, experiments were done to cluster data from two given databases. Simulation results verify the important roles of fuzzy ART2 in finding decision tree rules. Besides, the number of erroneously clustered patterns as well as the computational cost from decision trees can be significantly reduced as compared with the case without using fuzzy ART2.