AN EFFICIENT MULTI-CLASS EXPLORATION OF ENSEMBLE CLASSIFICATION USING GENETIC ALGORITHM

Pattern mining is the process of extracting and identifying patterns in the unstructured data from different sources. The existing methodologies are focused on the single class ensembles with individual parameters. The main objective of the proposed methods is to analyze and identify multiclass ensemble with different perspective of exploration. This multiclass ensemble is classified with multilayered approach in order to achieve high efficiency. The proposed method has been implemented by using evolutionary computing method. The genetic based classification performs more accurate result when compared to existing method of classification. There are various steps of genetic based method are considered such as selection, crossover and mutation process. The ensembles are classified in two layered approach namely local classification and global classification. In local classification the ensembles are maintained in separate clusters whereas global classification performs multiclass classification with single set. This method considered various parameters from diverse sources with high efficiency.