🚗 Car Price Prediction – Regression Model Building

A supervised machine learning project focused on implementing and comparing multiple regression algorithms to predict car prices in the American automobile market.

---

🚀 Run Notebook in Google Colab

Click below to open the notebook:

---

📘 Project Overview

This project is based on a business problem where an automobile company plans to enter the US market and aims to understand key factors influencing car prices.

The complete pipeline includes:

• Data Cleaning

• Exploratory Data Analysis (EDA)

• Feature Engineering

• Feature Scaling

• Model Building

• Model Evaluation

• Hyperparameter Tuning

• Model Comparison

The goal is to build an accurate regression model to predict car prices and support data-driven pricing strategies.

---

🎯 Objective

The main objectives of this project are:

🔹 Identify variables affecting car price

🔹 Analyze relationships between features and price

🔹 Build multiple regression models

🔹 Compare model performances

🔹 Optimize the best model using hyperparameter tuning

🔹 Provide business insights for pricing strategy

---

📂 Dataset Description

Component	Description
Records	205 cars
Features	25+ independent variables
Target Variable	price
Data Type	Numerical and Categorical

Key Features:

• Engine size

• Horsepower

• Fuel type

• Drive wheel type

• Car dimensions

• Brand

• Mileage

• Technical specifications

---

🧹 Data Preprocessing

The following preprocessing steps were performed:

✔ Loaded dataset using Pandas

✔ Initial exploration (shape, info, summary statistics)

✔ Handled missing values

✔ Removed duplicate records

✔ Detected and handled outliers

✔ Feature engineering (brand extraction)

✔ Encoded categorical variables

✔ Split dataset into Training (80%) and Testing (20%)

✔ Scaled features using StandardScaler

---

📊 Exploratory Data Analysis (EDA)

Visualizations used:

• Histogram

• Box Plot

• Heatmap Correlation

• Scatter Plot

Insights:

• Identified strong relationships between engine size, horsepower, and price

• Detected outliers affecting model performance

• Understood feature distributions

---

🤖 Regression Models Implemented

The following models were trained and evaluated:

• Linear Regression

• Decision Tree Regressor

• Random Forest Regressor

• Gradient Boosting Regressor

• Support Vector Regressor (SVR)

• Pruned Decision Tree

---

📊 Model Performance (Before Tuning)

Model	R² Score	MSE	MAE
Random Forest Regressor	0.913	0.0136	0.0882
Gradient Boosting Regressor	0.908	0.0143	0.0838
Linear Regression	0.879	0.0189	0.1054
Decision Tree Regressor	0.860	0.0219	0.1062
Support Vector Regressor	0.831	0.0263	0.1051
Pruned Decision Tree	0.831	0.0263	0.0958

✅ Best Model: Random Forest Regressor

❌ Worst Model: Pruned Decision Tree

---

⚙️ Hyperparameter Tuning

• Applied GridSearchCV

• Used Pipeline to prevent data leakage

• Performed 5-Fold Cross Validation

Parameters Tuned:

• n_estimators

• max_depth

• min_samples_split

---

📊 Model Performance (After Tuning)

Model	R² Score	MSE	MAE
Random Forest (Untuned)	0.9129	0.0136	0.0882
Random Forest (Tuned)	0.9131	0.0136	0.0847

🏆 Best Model: Tuned Random Forest Regressor

Final Metrics:

• R² Score: 0.9131

• MSE: 0.0136

• MAE: 0.0847

✅ Model performance improved after hyperparameter tuning

---

🏆 Final Model

✔ Random Forest Regressor selected

✔ Optimized using hyperparameter tuning

✔ Provides stable and accurate predictions

---

🔍 Feature Importance Analysis

Important features influencing car price:

• Engine size

• Curb weight

• Horsepower

• Car width

• highwaympg

These features play a major role in determining car pricing strategy.

---

📊 Evaluation Metrics

• R² Score

• Mean Squared Error (MSE)

• Mean Absolute Error (MAE)

---

⚠️ Limitations

• Dataset size is relatively small (205 records)

• Limited features may not capture all real-world pricing factors

• Market dynamics and external economic factors are not included

• Model may not generalize well to different regions or time periods

• Performance may vary for unseen or highly diverse car categories

---

🛠 Tech Stack

Tool	Purpose
Python	Programming language
Pandas	Data handling
NumPy	Numerical computation
Matplotlib / Seaborn	Visualization
Scikit-learn	ML models & preprocessing
Google Colab	Development

---

📁 Repository Structure

car-price-prediction/

│

├── CarPrice_Assignment.csv

├── Car_Price_Prediction.ipynb

├── README.md

---

🚀 How to Run the Project

1️⃣ Open Notebook

Click the Google Colab link above

---

2️⃣ Install Dependencies

pip install pandas numpy matplotlib seaborn scikit-learn

---

3️⃣ Run the Notebook

• Execute all cells step-by-step

• Analyze model performance

---

📌 Business Insight

This project helps:

• Identify key drivers of car pricing

• Optimize product features

• Support strategic pricing decisions

• Enable data-driven business planning

---

📌 Academic Submission

This project was created as part of a Machine Learning & Data Science program, showcasing end-to-end regression modeling, including data preprocessing, EDA, feature engineering, model comparison, and optimization for car price prediction.

---

👤 Author

Name: Laya Mary Joy

Organization: Entri Elevate

Date: February 14, 2026

---

⭐ Acknowledgment

Thanks to Entri Elevate for guidance and support.

---

📌 Future Improvements

• Use larger and more diverse datasets

• Include real-time market data

• Try advanced models (XGBoost, LightGBM)

• Deploy as a web application

---