Machine Learning Case Studies

Overview This repository contains machine learning case studies developed as part of my learning journey in Data Analytics and Machine Learning. The projects focus on understanding the complete machine learning workflow. The objective of these case studies is not only to build predictive models but also to understand the reasoning behind model selection, statistical significance, and business decision-making.

Skills Demonstrated

• Python
• Pandas
• NumPy
• Data Visualization
• Exploratory Data Analysis (EDA)
• Statistics
• Correlation Analysis
• Hypothesis Testing
• Simple Linear Regression (SLR)
• Multiple Linear Regression (MLR)
• Model Evaluation Metrics
• Statistical Significance Testing
• Variance Inflation Factor (VIF)
• Residual Analysis
• Business Insights Generation

Case Studies Included

Newspaper Circulation Prediction Using Simple Linear Regression

• Business Problem Can Daily Newspaper Circulation be used to predict Sunday Newspaper Circulation?

-- Techniques Used

• Data Exploration
• Correlation Analysis
• Simple Linear Regression
• Best Fit Line Analysis
• R² Score
• MAE
• MSE
• RMSE
• P-value
• T-value
• OLS Regression Analysis

-- Key Findings

• Strong positive relationship between Daily and Sunday circulation.
• Daily circulation was found to be a statistically significant predictor.
• The model explained a substantial portion of the variation in Sunday circulation.

Advertising Sales Prediction Using Multiple Linear Regression

-- Business Problem Which advertising channels contribute most to product sales?

-- Features

• TV Advertising Budget
• Radio Advertising Budget
• Newspaper Advertising Budget -- Target
• Sales

Techniques Used

• Correlation Analysis
• Heatmap Visualization
• Multiple Linear Regression
• Model Evaluation Adjusted R² P-value Analysis T-value Analysis F-statistic Analysis Variance Inflation Factor (VIF) Residual Analysis

Key Findings

• TV advertising showed the strongest impact on sales.
• Radio advertising contributed significantly to sales growth.
• Newspaper advertising was not statistically significant.
• The model achieved excellent predictive performance with an R² score above 0.90.

Conclusion

This repository showcases my practical learning journey in Machine Learning through hands-on case studies. The projects demonstrate the complete machine learning workflow, including data exploration, statistical analysis, model building, evaluation, and business insights. As I continue learning new concepts and algorithms, this repository will grow with additional case studies and real-world applications, reflecting my progress in Data Analytics and Machine Learning.