📈 Multiple Linear Regression

Predict startup profitability using Multiple Linear Regression with Backward Elimination — implemented in both Python and R.

What It Does

Models how R&D Spend, Administration, Marketing Spend, and State influence a startup's Profit using the 50 Startups dataset.

1. Loads and preprocesses data (one-hot encodes the categorical State column)
2. Splits into 80/20 train/test sets
3. Fits a Multiple Linear Regression model and reports R² / RMSE
4. Performs Backward Elimination (statsmodels OLS) to find the optimal predictor subset

Dataset

50_Startups.csv — 50 records with columns:

Column	Description
R&D Spend	Research & development expenditure
Administration	Administrative costs
Marketing Spend	Marketing expenditure
State	New York, California, or Florida
Profit	Target variable

🛠 Tech Stack

	Language	Libraries
🐍	Python 3.10+	numpy · pandas · matplotlib · scikit-learn · statsmodels
📊	R	caTools

Getting Started

Python

pip install numpy pandas matplotlib scikit-learn statsmodels
python multiple_linear_regression.py

R

install.packages("caTools")   # first time only
source("multiple_linear_regression.R")

Both scripts expect 50_Startups.csv in the same directory.

⚠️ Known Issues

• No cross-validation or hyperparameter tuning (simple demonstration project).
• The R script encodes State as numeric factor levels (1, 2, 3), which may imply ordinality — acceptable for lm() with factor types but worth noting.

License

MIT © 2018 Kaustabh Ganguly