Improved Lion Optimization based Enhanced Computation Analysis and Prediction Strategy for Dropout and Placement Performance Using Big Data

Abstract

Background: Predicting the undergraduate’s placement performance is vital as it impacts the credibility of educational institutions. Hence, it is significant to predict their performance based on placement in the early days of degree program.

Objectives: The study intends to predict the undergraduate’s placement performance through the introduced ANN-R (Artificial Neural Network based Regression) as it is able to handle fault tolerance. For efficient prediction, relevant feature selection is needed that is performed by the proposed ILO (Improved Lion Optimization) algorithm as it has the ability to find nearest probable optimal solution.

Methodology: Initially, the parameters and population are initialised. Subsequently, first best-agent is stated in accordance with fitness function. Subsequently, position of present search agent is updated. This iteration continues until all the features are selected and optimized result is attained. Here best score is computed using the proposed ILO for feature selection. Finally, the dropout analysis and placement performance of students is predicted using the introduced ANN-R through a train and test split.

Results/Conclusion: Performance of the proposed system is analysed in accordance with loss metrics. Additionally, internal comparison is performed to find the extent to which the actual and predicted values correlate with one another during prediction using the existing and proposed system. The outcomes revealed that the proposed system has the ability to predict the student’s placement performance along with domain of interest with minimum errors than the traditional system. This makes the proposed system to be highly suitable for predicting student’s performance.