This project demonstrates the use of a Decision Tree classifier (ID3 algorithm) to predict student performance based on their study hours, attendance, and previous grades.
Dataset: The project uses a CSV file (student_data.csv) containing historical data on student study hours, attendance, previous grades, and their corresponding performance. Model: A Decision Tree classifier is trained on this data to learn the patterns and relationships between these features and performance. Prediction: The model can take new student data as input and predict their likely performance. Feedback Loop: The project also adds the new student data and predicted performance back to the original dataset, creating a continuous learning loop for the model.
study_hours (float) attendance (float) previous_grades (float) performance (categorical, e.g., "High", "Medium", "Low") #Dependencies: Install the required libraries: pip install pandas scikit-learn python student_performance_prediction.py
Enter the student's study hours, attendance percentage, and previous grades when prompted. The model will predict their performance. Project Structure student_data.csv: The dataset student_performance_prediction.py: The main Python script