Cognivia- Student Burnout Predictor

A beginner-friendly machine learning project that predicts the risk of student burnout based on lifestyle and behavioral patterns such as sleep, study hours, screen time, exercise, and mood.

This project demonstrates the complete basic machine learning workflow including dataset generation, model training, evaluation, and prediction using Python.

---

Project Overview

Student burnout is a common issue caused by excessive workload, lack of sleep, stress, and unhealthy study habits. Cognivia attempts to model these factors and predict the likelihood of burnout using a machine learning algorithm.

The system analyzes several lifestyle variables and predicts the burnout level as:

• Low
• Medium
• High

The model is trained on a synthetic dataset generated using Python and evaluated using a Random Forest classifier.

---

Features

• Synthetic dataset generation for student lifestyle patterns
• Machine learning model training using Random Forest
• Burnout risk prediction from user input
• Terminal-based interactive prediction system
• Organized machine learning project structure
• Beginner-friendly implementation

---

Machine Learning Pipeline

The project follows a typical machine learning workflow:

Dataset Generation → Data Preparation → Model Training → Model Evaluation → Prediction System

---

Dataset Description

The dataset contains 500 simulated student records generated programmatically.

Each record includes the following features:

Feature	Description
sleep_hours	Number of hours the student sleeps
study_hours	Hours spent studying per day
screen_time	Hours spent on phone/computer
exercise	Whether the student exercised (0 = No, 1 = Yes)
mood	Self-reported mood level (1–5 scale)
burnout	Burnout risk category (Low, Medium, High)

Example dataset entry:

sleep_hours: 4

study_hours: 9

screen_time: 10

exercise: 0

mood: 2

burnout: High

---

Model Used

The project uses the Random Forest Classifier from Scikit-Learn.

Random Forest works by creating multiple decision trees and combining their predictions to improve accuracy and reduce overfitting.

This algorithm was chosen because it:

• Handles structured data well
• Is beginner friendly
• Provides strong prediction performance

---

Model Accuracy

After training and testing on the dataset:

Model Accuracy: 0.95 (95%)

This means the model correctly predicted burnout levels for approximately 95% of the test data.

Note: The dataset is synthetically generated, which results in clearer patterns and higher accuracy compared to real-world datasets.

---

Installation

Clone the repository:

git clone https://github.com/angelabera/Cognivia
cd Cognivia

Create a virtual environment:

python -m venv venv

Activate the Virtual Environment

Windows (PowerShell / Command Prompt)

venv\Scripts\activate

macOS / Linux

source venv/bin/activate

Install required dependencies:

pip install -r requirements.txt

---

Running the Project

1. Generate the Dataset

Run the following command:

python ml/generate_dataset.py

This will create the dataset:

data/burnout_dataset.csv

---

2. Train the Machine Learning Model

python ml/train_model.py

This will train the Random Forest model and save it as:

ml/burnout_model.pkl

---

3. Run Burnout Prediction

python ml/predict.py

Example interaction:

• Enter sleep hours: 4
• Enter study hours: 9
• Enter screen time: 10
• Exercise (0=no,1=yes): 0
• Mood (1–5): 2

Output:

Predicted Burnout Level: High

---

Data visualization (optional)

If you'd like to explore the generated data visually, run the visualization script after creating the dataset.

python ml/visualize_data.py

The script loads data/burnout_dataset.csv and opens a scatter plot showing sleep_hours vs study_hours. See the code at ml/visualize_data.py.

Note: Ensure data/burnout_dataset.csv exists (run the dataset generator first).

---

Technologies Used

• Python
• Pandas
• NumPy
• Scikit-Learn
• Matplotlib
• Machine Learning

---

Learning Outcomes

This project demonstrates key machine learning concepts such as:

• Dataset generation and handling
• Feature selection
• Supervised learning
• Model training and evaluation
• Prediction systems
• Project structuring for machine learning workflows

---

Author

Angela Bera