Classification of precise or point valued data is the regular practice in traditional approaches. If the data varies from point valued to a range bound values makes the classification complicated. In this scenario of uncertain data usual practice is to match the range bound values to single values by considering their mean or average. This approach however sacrifices the accuracy of the classifying the tuples to their associated class labels. The classical decision tree technique can be extended to handle the range bound values by considering the probability density function (pdf) over complete information of a feature or attribute.