Competent multi relational classifier using filter based feature selection method on CrossMine algorithm

Multi Relational Classifier deals with structured data which is stored in relational databases. When working with MRC (Multi relational Classifier) no need to convert multiple relations in single universal relation. CrossMine algorithm works with relational dataset without doing propositionalization. CrossMine algorithm use Tuple ID Propagation provides the way to directly search pattern from multiple tables with the use of virtual join. Accuracy is the major issue that should be considered while designing any classifier. In this paper, we have implemented CrossMine algorithm and then applied filter based feature selection method in order to increase the classifier accuracy.