Enhancing Classification Power: Tree Strength-Infused Enriched Random Forest
Vikas Jain, Tej Bahadur Chandra, Atul Kumar Srivastava · 2024
Over the last two decades, Random forest (RF) has been one of the well-known and most exploited ensemble-based machine learning approaches. It has been used for computer vision, pattern recognition, medical imaging, and many other fields. However, it has always been a challenging task to come up with an approach that will lead to constructing a more reliable RF. In this study, we used a concept of tree strength to construct the RF, called as Enriched Random Forest (ERF). As a feature extraction, Bag of Visual Words (BoVW) is used, and Grey Wolf Optimization (GWO) as a feature selection during the construction of the ERF. The suggested method has been evaluated using the MNSIT, Caltech-101, Caltech-256, and UCI repositories—all of which are well-known and openly accessible. The outcomes of the experiments indicate that the ERF has achieved better results than other RF-based approaches.