Prediction System for Student’s Academic Performance to increase University Admission System and Cumulative Grade Point Average Credits

Abstract

Education sector is a big boon for society and it is utmost important to strengthen the university admission system by constructing basic eligibility criteria in order to maintain consistent results and to analyze students’ performance in the forthcoming semesters. This research work incorporates two prediction systems in which prediction system 1 consists of supervised machine learning classification algorithms such as Support Vector Machine, Random Forest, Naïve Bayes, Artificial Neural Networks Multi-Layer Perceptron and prediction system 2 is feeded with unsupervised clustering algorithms such as KNN, K-Means, DBSCAN and Agglomerative hierarchical clustering algorithms that have been trained with students’ academic and personal details. It is found that 98% of detection accuracy is yielded as the result of supervised classification algorithms. Data is an important asset for every organization and hence this article is proposed to secure data from common breaches in software defined network. In this article, hybrid cipher model is proposed to safeguard the communication of data transmitted among the layers in software defined networks. The logic of hybrid cipher model is incorporated in software defined controller which encrypts open flow request and response messages. Software Defined Network is adapted for implementing hybrid cipher model as the network provides customizable platform and act as a unmanned security featured software controller. The proposed Hybrid Diagonal Transposition algorithm is incorporated with software defined wireless sensing node for encrypting user’s data. Hence the unmanned security featured wireless sensing node is situation-aware, it detects malicious traffic flows and encrypts user’s data. Hybrid Diagonal Transposition algorithm prevents data breaches in Software Defined Networks. Results are interpreted for various network and sensor metrics such as routing hops, participating node temperature, battery voltage, humidity, lights, received packets per node, number of network hops, power consumption, radio duty cycle, temperature of sensors, beacon interval, network hops, routing metric and the same work will be extended in future for comparative results.