Predicting Startup Profit using Multiple Linear Regression

Objective:

The goal of this project is to predict the profit of startups based on how much they spend on:

• Research & Development (R&D)

• Administration

• Marketing

• And the State they operate in

Using this, we aim to help businesses make informed investment decisions.

Dataset Description: 50_Startups.csv

This dataset contains 50 records of startups, each with the following attributes:

2. Splitting the Data

• Defined the independent variables (X) and dependent variable (y = Profit)

• Split the dataset into training and testing sets (80/20 ratio)

pythonCopyEditfrom sklearn.model_selection import train_test_split
X = df.drop("Profit", axis=1)
y = df["Profit"]
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

---

3. Model Training

• Used Multiple Linear Regression to fit the training data

pythonCopyEditfrom sklearn.linear_model import LinearRegression
model = LinearRegression()
model.fit(X_train, y_train)

• The model tries to fit a linear equation:
  Profit = b0 + b1*(R&D Spend) + b2*(Admin) + b3*(Marketing) + b4*(State_X) + ... + error

---

4. Model Evaluation

• Evaluated using:

  • R² Score: Tells how well the model explains the variation in profit

  • Mean Squared Error (MSE): Measures average squared difference between predicted and actual profits

pythonCopyEditfrom sklearn.metrics import r2_score, mean_squared_error

y_pred = model.predict(X_test)
print("R² Score:", r2_score(y_test, y_pred))
print("MSE:", mean_squared_error(y_test, y_pred))

• Also visualized predicted vs actual profits using matplotlib or seaborn:

pythonCopyEditimport matplotlib.pyplot as plt
plt.scatter(y_test, y_pred)
plt.xlabel("Actual Profit")
plt.ylabel("Predicted Profit")
plt.title("Actual vs Predicted Profit")
plt.show()

---

5. Feature Optimization (Optional)

• Used Backward Elimination with statsmodels to find which features are statistically significant.

pythonCopyEditimport statsmodels.api as sm
X_opt = sm.add_constant(X)  # Adds intercept
model_ols = sm.OLS(y, X_opt).fit()
print(model_ols.summary())

• Based on p-values, we can iteratively remove insignificant features to simplify the model without losing accuracy.

---

Insights & Learnings

• R&D Spend had the most significant positive impact on profit.

• Marketing Spend also played a role, but to a lesser extent.

• Administration costs and State had very little influence.

• The project gave me hands-on exposure to:

  • Feature engineering

  • Dummy variable trap avoidance

  • Model building and evaluation

  • Statistical analysis with statsmodels

---

🔧 Tools & Libraries Used

• Python

• Pandas & NumPy

• Scikit-learn

• Statsmodels

• Matplotlib & Seaborn

• Google Colab

---

Conclusion

This project is a great real-world example of how machine learning models can help in business decision making. It allowed me to explore both technical ML implementation and business reasoning.

---