The study on web data mining based on belief rough set classification

In this paper, a new approach of classification system based on rough sets named BRSC have been applied to generate a classification model from uncertain data consisting of web usage. The uncertainty appears only in decision attributes and is handled by the TBM, one interpretation of the belief function theory. The feature selection step used to construct the BRSC is based on the calculation of dynamic core to extract more relevant and stable features for the classification process. In experimentations, three evaluation criteria have been chosen to judge the performance of the BRSC applied to the web usage mining dataset.