Gaussian Fuzzy HSM Algorithm for Classification of BIRADS Dataset using Decision Tree

Classification is one of the methods used in identifying the accuracy of the dataset of multiple classes. This paper has proposed a method named FHSM ( Fuzzy Heterogeneous Node Splitting Measure ) for building decision tree which used Gaussian membership function in assigning fuzzy membership value to the attributes. Whenever decision trees are concerned the size of the tree plays a major role in developing the rules and also shows impact on computations. When huge amount of data is considered we may get larger size decision tree which leads to more number of rules and more computational time, which becomes more critical in handling when dataset is too large. The proposed method tried to reduce the size of the decision tree and also provided more accurate results in classifying the BIRADS data set when compared to most of the other available techniques