Determination of a Min-Max Neural Network Clustering Fuzzy Approach In Comparison with the Classical Statistical Clustering Methods in Medical Studies along wtih its Appicotion in Identifying Richets

Determination of a pattern for categorization of the observations has been one of the main issues in many of the recent studies. It is a fact that the health status of the subjects (normal or health vs sick), is the main variable in most of the studies in the field of medical sciences. Therefore, it is crucial to employ appropriate categorization methods in order to minimize the risk of error in these studies. Categorization of the data is one of the main issues in multivariate statistical analyses. In the present study, we tried to introduce a method that uses a clustering fuzzy approach, on the basis of neural network, for categorization of the subjects into "health" and "sick" groups. The results were also compared with the classical statistical methods.108 students participating in this study were tested for 3 variables: calcium, phosphorus, and alkaline phosphatase in order to identify whether they were suffering from richests, The results indicated that categorization of the subject on the basis of fuzzy neural network enjoyed 92% accuracy, while cluster analysis and discriminant analysis demonstrated 81% and 88.8% accuracy respectively.