Skip to content

Repository files navigation

KCET Rank Prediction and College Recommendation System 🚀

Table of Contents Project Overview Dataset Acquisition and Preparation Exploratory Data Analysis Machine Learning Methodology Models and Hyperparameter Tuning Evaluation Metrics Streamlit Web Application Deployment

Running Locally

Future Enhancements

Disclaimer

Project Overview

This project delivers a robust ML solution for predicting KCET ranks and providing personalized college recommendations. Students can estimate competitive ranks before official results and explore colleges based on predicted ranks and branch preferences.

Technologies used: Python, Pandas, Scikit-learn, Streamlit

Example of KCET Rank Predictor & College Finder UI

Dataset Acquisition and Preparation

Sources: Official KCET datasets from KEA for years 2023–2025.

Data Cleaning:

Unified Excel format, handled missing values

Normalized column names

Standardized course and college names

Integration: Merged datasets to form a comprehensive foundation for model training.

Exploratory Data Analysis (EDA)

Distribution plots to visualize marks and rank spread

Correlation heatmaps for feature relationships

Statistical summaries to detect anomalies and biases

Machine Learning Methodology Models and Hyperparameter Tuning Model Best R² RMSE Key Hyperparameters Example Gradient Boosting 0.9996 1169.60 subsample=0.8, n_estimators=500 Random Forest 0.9996 1170.07 n_estimators=100, max_depth=30 Decision Tree 0.9996 1170.01 min_samples_split=8, max_depth=40 Extra Trees 0.9994 1345.15 n_estimators=100, max_depth=50 Ridge Regression 0.9214 15872.43 alpha=0.01, solver='lsqr'

Gradient Boosting performed best.

Linear models provided strong baselines.

Evaluation Metrics

R² Score – Explained variance

RMSE – Average prediction error in rank units

Streamlit Web Application

Features: Interactive input for KCET marks, board marks, and exam year

Real-time rank prediction College recommendations filtered by rank, branch(es), location, and type

Multi-branch selection with robust matching (ignores “engineering/engg”)

Clickable college websites Professional gradient background and readable UI Deployment

The application is deployed on Render:

Try the live KCET Rank Predictor & College Finder https://kcet-rank-prediction-college-tolm.onrender.com Running Locally

Clone the repository:

git clone https://github.com/yourusername/kcet-rank-prediction.git cd kcet-rank-prediction

Install dependencies:

pip install -r requirements.txt

Run the Streamlit app:

streamlit run streamlit_app.py

Future Enhancements

Category-wise seat reservations & quota-aware recommendations

Extended dataset with branch-specific cutoff trends

Interpretability using SHAP

Multi-modal input options (voice, image)

CSV/Excel download of eligible colleges

Disclaimer ⚠️

The KCET Rank Predictor is trained on limited historical data and is intended for preliminary guidance only. Predicted ranks may differ from official results. Users should verify recommendations with official college sources.

About

This project presents a robust machine learning solution to predict KCET (Karnataka Common Entrance Test) ranks based on marks and provide personalized college recommendations. The system aids students in estimating their competitive rank prior to official results and assists in selecting suitable colleges based on predicted ranks and branch.

Topics

Resources

Stars

3 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages