Simple-Linear-Regression-model

Week 1 Machine Learning Project

Predicting Student Exam Performance Using Linear Regression

Overview

This project explores the key factors that influence students’ exam performance. The goal was to analyze study-related behaviors, identify which factors most strongly impact exam scores, and build a simple predictive model using linear regression.

Tools Used

• Python
• Pandas
• NumPy
• Matplotlib
• Scikit-Learn (sklearn)
• Jupyter Notebook

Loading and Exploring the Dataset

The dataset includes the following variables:

• Hours Studied
• Attendance Rate
• Exam Score

Initial exploration was performed using df.info() and df.describe() to understand the dataset structure, data types, and statistical summaries.

Visualizing Relationships

Scatter plots were used to explore relationships between the variables and exam performance:

• Hours Studied vs Exam Score
• Attendance Rate vs Exam Score

Insights

• Hours studied showed a strong positive correlation with exam scores.
• Attendance rate showed a weaker but noticeable influence.

Correlation Results

Variable Pair	Correlation
Hours Studied vs Exam Score	0.8146
Attendance Rate vs Exam Score	0.2177
Hours Studied vs Attendance	-0.0874

Hours Studied was the strongest predictor, with a Pearson correlation of 0.8146 (p-value ≈ 1e-48), indicating a strong and statistically significant relationship.

Building the Linear Regression Model

A linear regression model was created using:

• Predictor (X): Hours Studied
• Target (y): Exam Score

The regression line was plotted over the scatter data to visualize the model fit.

The dataset was split using train_test_split():

• 80% training data
• 20% testing data
• random_state=42 for reproducibility

Model Evaluation

A custom evaluation function calculated the following metrics:

• R-squared (R²)
• Mean Absolute Error (MAE)
• Mean Squared Error (MSE)
• Root Mean Squared Error (RMSE)

Training Performance

• R²: 0.6593
• MAE: 4.6463
• MSE: 34.2023
• RMSE: 5.8483

Testing Performance

• R²: 0.6753
• MAE: 4.7044
• MSE: 35.1104
• RMSE: 5.9254

The similar training and testing results indicate that the model generalizes well.
Study hours remain a strong and reliable predictor of exam performance.

What I Learned

• How to inspect and clean datasets using Pandas
• How to visualize relationships with Matplotlib
• How to interpret correlation using Pearson’s coefficient
• How to build and evaluate a Linear Regression model
• How to calculate R², MAE, MSE, and RMSE
• Why study hours strongly predict academic performance