Identifying the Machine Learning Techniques for Classification of Target Datasets

Abstract

Given the dynamic and convoluted nature of numerous datasets, the necessity of enhancing performance outcomes and handling multiple datasets has become more challenging. To handle these issues effectively and improve the quality of multiple approaches, the capabilities of various Machine Learning techniques such as K-Nearest Neighbor (KNN), Logistic Regression (LR), Naive Bayes(NB) and Support Vector Machine (SVM) have been utilized in this study. In this paper, the binary classification method using five different datasets and many predictor variables have been utilized. Moreover, this research has mainly focused on determining the classification of data into the subsets that share the standard designs. In this regard, many approaches had been studied extensively and used to achieve better yields from the existing literature; however, they were inadequate to provide efficient outcomes. By applying four Supervised ML classification algorithms along with the UCI Datasets of ML Repository, the robustness of the method is progressed. The proposed mechanism is assessed by adopting five performance criteria concerning the accuracy, AUC (Area Under Curve), precision, recall and F-measure values. The current study experimental results revealed that there is a significant improvement in the confusion matrix rate compared with a similar study and this method can also be used for machine learning problems such as binary classification.

Keywords

• Machine Learning
• Data Mining
• K-Nearest Neighbor
• Logistic Regression
• Naive Bayes