Student Performance Index Prediction

This project demonstrates the implementation of multiple variable linear regression from scratch using Python. The goal is to predict a student's performance index based on the following factors:

• Hours Studied
• Previous Scores
• Extracurricular Activities
• Sleep Hours
• Sample Question Papers Practiced

Dataset

The dataset used for this project is available on Kaggle: Student Performance Dataset.

---

Implementation Overview

Key Features:

1. Z-Score Normalization
   Standardizes the data to center it around 0 and scale it to unit variance.

2. Custom Gradient Descent
   Performs weight and bias updates iteratively to minimize the Mean Squared Error (MSE).

3. Cost Function
   Evaluates the model's performance using MSE.

4. Feature Engineering
   Adds polynomial features for better model accuracy.

5. Visualization
   Visualizes actual vs. predicted values for better understanding of the model's fit.

6. R² Score Calculation
   Measures the goodness of fit for the regression model.

7. Custom Predictions
   Predicts the performance index for new input data.

---

Getting Started

Prerequisites:

• Python 3.x
• Libraries: numpy, pandas, matplotlib

Installation:

1. Clone this repository.

   git clone https://github.com/KartikAg13/student_performance_prediction.git
   cd student_performance_prediction

2. Download the dataset from the Kaggle link and place it in the root folder.

---

How to Use

1. Run the Notebook
   Open the Python notebook file in Jupyter or any compatible environment and execute the cells step-by-step.

2. Make Predictions
   Input new data in the format [Hours Studied, Previous Scores, Extracurricular Activities, Sleep Hours, Sample Question Papers Practiced] to predict the Performance Index.

---

Results

• Initial Cost: Evaluates the model's cost before training.
• Final Cost: Reduced cost after applying gradient descent.
• R² Score: Indicates how well the regression model fits the data.

---

Project Structure

• main.ipynb: Main implementation notebook.
• README.md: Project documentation.

---

Example Prediction

Input:

x_predict = np.array([6, 71, 1, 8, 2])

Output:

Predicted Performance Index: 85.43

---

Contributing

Contributions are welcome! Please fork this repository and submit a pull request with your changes.

---

Acknowledgments

• Kaggle for providing the dataset.
• Coursera for inspiration through the Machine Learning course.