Empirical Analysis of Rank Aggregation-Based Multi-Filter Feature Selection Methods in Software Defect Prediction

Selecting the most suitable filter method that will produce a subset of features with best performance remains an open problem is known as rank selection problem. A viable solution to this independently apply mixture methods and evaluate results. This study proposes novel aggregation-based multi-filter feature (FS) address high dimensionality in software defect prediction (SDP). The proposed combine lists generated by individual using aggregation mechanisms into single aggregated list. aim resolve multiple diverse computational characteristics dis-joint complete list superior methods. effectiveness was evaluated Decision Tree (DT) Naïve Bayes (NB) models on datasets from NASA repository. From experimental results, had impact (positive) performances NB DT than other experimented FS makes combination enhancement SDP.