🎓 Admission Prediction with Machine Learning

This project uses machine learning to predict a student’s chance of admission to graduate school based on academic performance and profile features. The dataset contains 500 entries with key factors like GRE Score, TOEFL Score, CGPA, Research Experience, and Recommendation Strength.

Dataset Overview

The dataset (Admission_Predict_Ver1.1.csv) includes the following features:

• GRE Score – Graduate Record Exam results

• TOEFL Score – English proficiency test results

• University Rating – Prestige level of the university (1–5)

• SOP – Statement of Purpose rating

• LOR – Letter of Recommendation strength

• CGPA – Cumulative GPA (out of 10)

• Research – Research experience (0 = No, 1 = Yes)

• Chance of Admit – Probability of admission (target variable)
  
  

Approach

The goal was to compare different regression models and evaluate their accuracy in predicting admission chances.

1. Split dataset into train/test sets

2. Trained multiple regression models:
   
   

   • 📈 SVR (Support Vector Regression)
     
     

   • 📉 Ridge & RidgeCV
     
     

   • 📝 Lasso
     
     

   • 📊 Bayesian Ridge
     
     

   • 📐 ARD Regression
     
     

   • 🔎 TheilSen Regressor
     
     

3. Evaluated performance using R² Score
   
   

Results

Here’s a comparison of model accuracies on the test set:

Model                         Accuracy (R² Score)

SVR                                     0.53

Ridge                                  0.81

RidgeCV                             0.81

Lasso                                  0.66

BayesianRidge                  0.80

ARDRegression                0.80

TheilSenRegressor          0.81 (Best)

Best Performing Model: TheilSen Regressor achieved the highest accuracy (~81%).

Sample Predictions

The model predicts probabilities like this:

Predicted: [0.96, 0.84, 0.86, 0.79, 0.83, 0.50, 0.91, 0.59, 0.78, 0.86]

Actual:    [0.49, 0.91, 0.64, 0.82, 0.90, 0.64, 0.90, 0.89, 0.64, 0.76]

Key Takeaway

Machine learning models like Ridge Regression and TheilSen Regressor can reliably predict admission chances with ~80% accuracy. With more feature engineering (e.g., student profile enhancements), the model could become even more powerful for university admissions counseling or student self-assessment tools.

For Source Code: