Analyzing and Forecasting the Students Placement Package

Abstract

Nowadays, Educational Institutions are growing at high rates. The goal of all higher education institutions is to provide students with a job through their training and placement office. One of the biggest challenges facing these days is to analyze and increase student performance. The main purpose our developed system is to analyze student’s past historical data includes academic as well as extracurricular activities and forecast the approximate placement package. This system helps both institute and student to plan and improve their performance before actual placement drives by companies. The last 2-year real time dataset having 600+ records are collected from Ramrao Adik Institute of Technology (RAIT) training and placement office from Navi Mumbai, India. Analysis of this dataset is carried out using most popular Random Forest Regression, Multilinear Regression and Decision Tree Regression algorithms. 3 independent models are built, and best performing algorithm is selected, then dumped the model using pickle and deployed the system on cloud. Thus the model with accuracy achieved more than 95% for 10+ trails and experiments has made us clear that it really helps placement cell from time to time to identify potential students and to focus and develop the skills needed for students.