Student academic performance prediction using random forest algorithm

Department:Computer Science
By:

Klintex001

Project ID: 9343

Rating:

(3.3) votes: 6

Price:₦5000

Abstract

Poor academic performance among students has traditionally been identified through manual, retrospective methods that often flag at-risk students too late for meaningful intervention. This study addresses that gap by developing a web-based Student Academic Performance Prediction System that applies the Random Forest algorithm to forecast whether a student is likely to Pass or Fail. The specific objectives were to collect and preprocess relevant student academic data, identify the key factors influencing academic performance, design and implement a predictive model using Random Forest, and train and test the model using appropriate machine learning techniques. The study adopted an experimental, quantitative, data-driven research design, using retrospective secondary data drawn from institutional student records for the 2024/2025 academic session. Data covering academic history, attendance, and demographic background were anonymized, cleaned, and processed through imputation, normalization, and encoding techniques before being used to train the model. The system was developed using Python, Flask, MySQL, HTML, CSS, and Bootstrap, with the Random Forest classifier implemented through the Scikit-learn library, following an Object-Oriented Analysis and Design (OOAD) methodology combined with iterative prototyping. The resulting application allows administrators to register students, manage academic records, generate performance predictions with an associated confidence value, and maintain a history of past predictions. The dataset was split 80%–20% for training and testing, and the trained model achieved an overall prediction accuracy of 100.00% on the test data, demonstrating the strong potential of the Random Forest algorithm for this classification task. The study concludes that machine learning can be effectively applied to educational data mining to support early identification of at-risk students, and recommends that future work test the model on larger, multi-institutional datasets, incorporate additional predictive factors, and compare Random Forest against other algorithms such as SVM, ANN, and XGBoost...

Preview +

Other Computer Science project topics and materials you might be interested in

Student academic performance prediction using random forest algorithm

Departments

UPLOAD A PROJECT