ZumraMNF/student-scholarship-predictor

Student Scholarship Predictor

This Space predicts whether a student is likely to receive a scholarship based on their academic and demographic profile.

Dataset

The model is trained on Student_performance_dataset.csv, which contains the following columns:

• —Student_ID, Student_Name
• —Math_Score, Science_Score, English_Score
• —Attendance_Percentage
• —Age
• —Gender
• —Department
• —Scholarship (target: Yes / No)

The single row with a missing Scholarship value is dropped before training, along with any rows missing required feature values.

Model

A RandomForestClassifier (class_weight="balanced", n_estimators=200, max_depth=5, min_samples_leaf=10, random_state=42) is trained on an 80/20 stratified train/test split of the encoded features:

• —Math_Score, Science_Score, English_Score, Attendance_Percentage, Age
• —Label-encoded Gender
• —Label-encoded Department

The target (Scholarship) is also label-encoded. Test accuracy is computed at startup and displayed live in the app header.

App

A Gradio Blocks interface (red theme) lets you enter a student's scores, attendance, age, gender, and department to get a Yes/No scholarship prediction with a confidence score and full probability breakdown.