Ensemble Multi-label Feature Selection Using Weighted Harmonic Mean

The volume of data is expanding dramatically due to the rapid growth of technology and the digital revolution. As a result, feature selection has evolved into an important preprocessing task. However, each feature selection process has its own set of advantages and disadvantages. Ensemble approaches have been established to improve the stability and robustness of feature selection algorithms. We present a novel ensemble feature selection approach for multi-label data based on the weighted harmonic mean. To assess its performance, we use seven real-world datasets to compare our proposed method with four existing algorithms. The findings show that our strategy outperforms the others. In addition, we perform a stability analysis to show the robustness of our proposed ensemble technique. This analysis adds to the proof of our method’s efficacy and reliability.