The objective of data mining is the derivation of knowledge from databases. It also utilizes expert knowledge concerning the monotonic relations between the response and predictor variables, which are represented in the form of monotonicity constraints. For classification problems with ordinal attributes very often the class attributes should increase with each or some of the explaining attributes. These are called classification problems with monotonicity constrain. To solve this problem, this work proposed Cascaded Fuzzy Relevance Vector Machine (CFRVM) with monotonicity based inequality constraints. Due to the huge size of data and amount of computation involved in data mining, high-performance computing is an essential component for any successful large-scale data mining application. So, parallelization is done effectively using Map Reduce method and it is integrated with classification method called MR-CFRVM, then feature selection is done by traditional method to reduce the attributes from the database using filter method followed by clustering the segment. In the classification the virtual pair optimization is done by using Ant Colony optimization algorithm (ACO) to improve the accuracy and reducing the error rate. Finally the samples are classified using CFRVM and have comparison on different metrics. The simulation experiment is done by MATLAB environment and proven that the proposed method acquires best result by comparison of different methods.
Volume 11 | 07-Special Issue
Pages: 1477-1491