Student Retention Intelligence System [SRIS]

Overview

This project aims to predict student dropout risks using machine learning and provide an institution-friendly dashboard and mobile application for real-time monitoring. The system consists of:

• Data Analysis & Preprocessing (Exploratory Data Analysis, cleaning, feature engineering)
• Machine Learning Model Training (Multiple models trained and evaluated)
• Backend API (Node.js-based API for communication)
• Web Dashboard (React frontend for institutional use)
• Model Hosting (Deployed on Hugging Face for inference)

Dataset

Dataset: Predict Students' Dropout and Academic Success (UCI)
This dataset is supported by program SATDAP - Capacitação da Administração Pública under grant POCI-05-5762-FSE-000191, Portugal. Realinho, V., Vieira Martins, M., Machado, J., & Baptista, L. (2021). Predict Students' Dropout and Academic Success [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C5MC89.

Data Processing

import pandas as pd

data = pd.read_csv("student_data.csv")
print(data.info())  # Display dataset structure

1. EDA: Identify missing values, feature distributions, correlations
2. Preprocessing: Handle missing data, normalize numerical values, encode categorical features
3. Feature Engineering: Generate new features, address class imbalances

Machine Learning Model Training

Model Selection

from sklearn.ensemble import RandomForestClassifier

model = RandomForestClassifier(n_estimators=100)
model.fit(X_train, y_train)

1. Baseline Models: Logistic Regression, Decision Trees
2. Advanced Models: Random Forest, XGBoost, Neural Networks

Model Hosting

• Trained model is deployed on Hugging Face
• API endpoint for real-time inference

Backend API (Node.js)

Features

• Handles requests from web dashboard
• Fetches prediction results from Hugging Face model
• Manages institution and student data securely

Sample API Endpoint

app.post("/predict", async (req, res) => {
  const { studentData } = req.body;
  const response = await axios.post("https://huggingface.co/model-endpoint", studentData);
  res.json(response.data);
});

Web Dashboard (React)

Features

• Login system for institutions
• Visualized dropout risk trends
• Notifications for high-risk students

Sample Component

function StudentCard({ student }) {
  return (
    <div className="student-card">
      <h3>{student.name}</h3>
      <p>Dropout Risk: {student.risk}%</p>
    </div>
  );
}

Deployment

Cloud Services

• Model Hosting: Hugging Face
• Backend Hosting: Render/Vercel
• Database: MongoDB Atlas

Deployment Steps

1. Train model and upload to Hugging Face
2. Deploy Node.js API
3. Deploy React web dashboard

Future Improvements

• Integration with more datasets
• Refinement of risk analysis model
• Advanced real-time monitoring tools

"Empowering institutions to improve student retention through AI-driven insights."