🎓 Admission Prediction using Machine Learning & Prolog

📌 Project Overview

This project predicts the likelihood of a student getting admitted into a university based on academic and profile features such as GRE score, TOEFL score, CGPA, and more. It utilizes both:

• A regression-based machine learning approach for predicting the chance of admission as a continuous probability.
• A Prolog-based rule system implementing a decision tree logic to classify admission as Yes or No based on input thresholds.

---

🎯 Objectives

• Analyze key factors influencing graduate school admissions.
• Build a regression model to predict the chance of admission.
• Implement a symbolic logic system to classify admission decisions using Prolog.
• Evaluate and optimize model performance using appropriate metrics.
• Provide an interactive Prolog interface for rule-based admission prediction.

---

📂 Dataset Description

The dataset includes various student attributes:

Feature	Description
GRE Score	Graduate Record Exam score (out of 340)
TOEFL Score	Test of English as a Foreign Language score
University Rating	Rating of university (1 to 5)
SOP	Statement of Purpose strength (1 to 5)
LOR	Letter of Recommendation strength (1 to 5)
CGPA	Undergraduate GPA (out of 10)
Research	Research experience (0 = No, 1 = Yes)
Chance of Admit	Target variable (probability from 0 to 1)

---

📊 Exploratory Data Analysis (ML Model)

• Visualized features against Chance of Admit.
• Identified CGPA, GRE, TOEFL, and Research as strong predictors via correlation heatmaps.

---

🧹 Data Preprocessing (ML Model)

• Handled missing values and outliers.
• Scaled numeric features using StandardScaler.
• No categorical encoding required for binary research feature.

---

📈 Model Building (ML Model)

Models Tried:

• Linear Regression (baseline)
• Ridge Regression
• Lasso Regression
• Random Forest Regressor

Evaluation Metrics:

• R² Score
• Mean Absolute Error (MAE)
• Mean Squared Error (MSE)
• Root Mean Squared Error (RMSE)

Results:

• Random Forest performed best, achieving R² > 0.85.
• Ridge and Lasso helped reduce overfitting.
• Linear Regression provided a solid baseline.

---

🧪 Hyperparameter Tuning (ML Model)

• Used GridSearchCV to optimize Ridge alpha and Random Forest parameters.
• Best results with tuned Random Forest.

---

🤖 Prolog-Based Admission Prediction

• Implements a decision tree logic in Prolog using dynamic predicates.
• Accepts user inputs interactively: GRE, TOEFL, CGPA, SOP, LOR, University Rating, and Research experience.
• Applies defined admission rules based on thresholds to classify admission as Yes or No.
• Useful for explainable, rule-based decision making alongside machine learning predictions.
• Provides an educational example of symbolic AI in admission prediction.

---

📊 Summary of Model Performance

Model	R² Score	RMSE	MAE
Linear Regression	~0.80	~0.05	~0.04
Random Forest	~0.85+	~0.03	~0.025
Ridge/Lasso	~0.78–0.82	~0.04	~0.035

---

📂 Project Structure

• Data Analysis and Machine Learning code (Python notebooks/scripts)
• Prolog knowledge base and user interaction script for rule-based admission
• Documentation and reports summarizing insights and model evaluations

👨‍💻 Author Krunal Patel https://github.com/Krunalscorp