ALRC: A Novel Adaptive Linear Regression Based Classification for Grade based Student Learning using Radio Frequency Identification

Abstract

Radio Frequency based Identification is a prominent enabler for many real-world applications. One viable application is the performance monitoring of university students. The data received from the radio frequency identification tags can be collected and crunched for making quick, accurate real-time decision. An attempt is made to classify the students based on their academic performance using a few path-breaking machine learning algorithms. The accurate classification of students’ performance under different categories is bound to facilitate grade-based teaching and training for specific students. To achieve this, a novel classification mechanism based on adaptive linear regression algorithm is proposed and implemented using Java. To verify and validate the performance of the proposed adaptive linear regression based classification, the performance metrics of the existing classifiers such as Naive Bayes, K-Nearest Neighbourhood, Ada Boost, decision table, decision trees, random forest, linear regression and multi class classifier algorithm are compared. From the results, it is confirmed that the metrics such as correlation co-efficient, precision, F1 score, mean absolute error, root mean square error, relative absolute error, root relative squared error, recall, receiver operating characteristic curve and precision recall curves are found to be better for the proposed adaptive linear regression based classification. That is, the results obtained have clearly illustrated the efficiency of the proposal.