Developed Modified Particle Swarm Optimization For Feature Selection On Learning Based Big Data In Cloud Computing

Abstract

The recent years, Machine Learning (ML) and Educational Data Mining (EDM) methods received significant attention for several aspects. Data Mining Methods can be employed for extracting meaningful data for the educational datasets. CMT (Cloud Mapping Tables), a result based mostly on utilization of information storage repository techniques for cloud-based Big - data analysis, can provide not only improved information management but also endurance with multidimensional information processing capacities. It could be utilized for a prediction of the current educational system, Student Academic Performance (SAP) has been used to appropriate corrective measures. . Data Storage is one of the best techniques to handle Big data. The information is frequently not even in the correct format to assist a company's decision-making process. In this paper, the issue addressed was applying a feature selection-based classification model for SAP. Initially, an adjusted Particle Swarm Optimization (MPSO) algorithm was applied effectively to select the set of features. For information categorization, the Multi-Layer Perceptron using Stochastic Gradient Descent (MLP-SGD) framework has been used. The suggested students ' performance approaches were combined with the Radial Basis Function (RBF) approach to reduce the miscategorized cases in the collected data, resulting in improved classification results. To verify the efficiency of the suggested, sets of data on the academic performance of students was obtained from secondary schools in Tanjore, Tamil Nadu, India. The modeled results demonstrated that the proposed significantly outperform the previous techniques.