,
Research Scholar, Department of Computer Science, Karpagam Academy of Higher Education , Coimbatore, Tamil Nadu , India
Assistant Professor, Department of Computer Science, Karpagam Academy of Higher Education , Coimbatore, Tamil Nadu , India
The innovation of Machine Learning (ML) techniques is evolving from basic techniques to optimized techniques, considerably improving the performance of prediction models. In the proposed work, the study primarily explores fundamental ML classification methods to classify banking customers based on their credit information. The classification of customers targets five categories: Outstanding, Excellent, Good, Satisfactory, and Bad. The aim is to assess the profitable customer categories and gain successful business by offering resources. The basic classification algorithms used in the proposed work are K-Nearest Neighbour (K-NN), Support Vector Machines (SVM), Decision Tree (DT), and Random Forest (RF) Classifiers. Using standard evaluation metrics, the performance of the classifiers are evaluated. Based on the metrics the comparative analysis is conducted, and comprehended the performance metrics need to be elevated. To manipulate this, Hyperparameter GridSearchCV (HGSCV) optimization is adopted, which is putative for its exhaustive search capabilities. However, the present accuracy scores of algorithms could be slightly improved while applying the HGSCV. Subsequently, the analysis moves on to an advanced optimization meta-heuristic optimized technique known as Particle Swarm Optimization (PSO). In this approach, the GlobalBestPSO method is implemented to tune the classifiers. The performance of the optimized classifiers such as GlobalBestPSO-SVM (gbestPSO-SVM), GlobalBestPSO-KNN (gbestPSO-KNN), GlobalBestPSO-DT (gbestPSO-DT), and GlobalBestPSO-RF (gbestPSO-RF) classifier are evaluated by analyzing the chosen set of parameters. The comparison of test results demonstrates the outstanding performance metrics in the optimized method, with accuracy outperformed with exceeding 0.95 score. The proposed hybrid model, integrating GlobalBestPSO with basic classifiers, superiors both traditional classifiers and tuned model HGSCV. The analysis is concluded to figure out the performance metrics of boosted classifiers, which optimized with the GlobalBestPSO, offers superior performance than others beyond all metrics.
This is an open access article distributed under the Creative Commons Attribution Non-Commercial License (CC BY-NC) License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
0
The statements, opinions and data contained in the journal are solely those of the individual authors and contributors and not of the publisher and the editor(s). We stay neutral with regard to jurisdictional claims in published maps and institutional affiliations.