A Supervised Clustering Algorithm Base on Alternative distance for Using multimedia software to Identify Alzheimer’s disease

Fuzzy algorithms of GG and GK based on Mahalanobis distance can improve those limitations of roller bearing architectural classifications, but Gath and Geva clustering algorithm could only use the data with multi-dependent variable normal distribution. GK-algorithm is limited as it must know the distribution of data. The data set was set by Min-Hwei College of Health Care Management in Taiwan and is a complete test on medical staff and skill of medical care. Hence we can collect the student data in the classes where they have good abilities. This discussion is about supervised classification method constructing the fuzzy clustering by means of putting forward method. Our discussion about an updated essential value not diverging calculation method to promote the prevention of Gath-Geva or Gustafson–Kessel methods, removed the limit of the determinants of common variance of matrices in the Gustafson–Kessel methods, and alternative for the common variance matrix with the statistics of related matrix which exists in the main function. The experimental results of real data sets of using multimedia software to identify Alzheimer’s disease and then to prevent aging prove that the supervised classification method constructing the fuzzy clustering by means of method obtain much better result, when data dispersion between different clusters is overlapping or the shape of clusters is not roller bearing.