Prediction of the Computer Science Department's Educational Performance Through Machine Learning Model by Analyzing Students' Academic Statements

Abstract

In this digital world, technology is changing every second. For this reason, tech-based educational departments should be aware of their strong and weak areas. By doing so, they can analyze their performance and emphasis the respected fields to prepare their students for the challenging job market and higher studies. In this research, a novel machine learning model is proposed to predict the performance of computer science department students by analyzing their academic statements. The academic areas of the departments are divided into fifteen fields, each of which is categorized as Excellent, Very Good, Good, Average, and Bad. One thousand students’ academic statements were used as the dataset. The prediction outputs are labeled into multiple categories, with multiclass data in each. Different models are developed using multiclass-multioutput classification algorithms like Decision Tree (Tree), Extra Tree (Tree), Extra Trees (Ensemble), K-Neighbors (Neighbors), Radius Neighbors (Neighbors), and Random Forest (Ensemble) Classifier. From the developed models, the model that used "Random Forest Classifier (Ensemble)" provided the best results, and the accuracy was almost 94%. Finally, a comparison of our research with previous and existing different developed models is presented to show the novelty of this research.