/Car Price Prediction using ML

Car Price Prediction Using Machine Learning

1. Introduction

Car Price Prediction is a beginner-friendly Machine Learning project where the goal is to predict the selling price of a car based on different characteristics of the vehicle.

The price of a used car can depend on many factors. Some important factors include the age of the car, engine size, total mileage, and the number of previous owners. Machine Learning can analyze historical car information and learn the relationship between these features and the selling price.

This is a Supervised Machine Learning problem because the model is trained using historical car data where the actual selling prices are already known.

The target variable in this project is Price, which is a continuous numerical value. Therefore, this is a Regression problem.

In this project, we use Linear Regression, a popular and beginner-friendly regression algorithm for predicting continuous numerical values.

2. Objective

The main objective of this project is to build a Machine Learning model that learns the relationship between car features and selling price, then uses the learned relationship to predict the price of a new car.

The project demonstrates important Machine Learning concepts such as:

  • Creating a car dataset

  • Selecting input features

  • Selecting a continuous target variable

  • Regression

  • Linear Regression

  • Model training

  • Making predictions

  • Interpreting numerical results

3. Features Used

The model uses four input features.

Car Age

Car_Age represents the age of the car in years. Generally, older cars may have a lower market value compared with newer vehicles.

Engine Size

Engine_Size represents the approximate engine capacity in cubic centimeters (cc). Engine size can be associated with the characteristics and value of a vehicle.

Mileage

Mileage represents the total distance traveled by the car in kilometers. Cars with higher mileage may have lower resale prices.

Number of Owners

Owners represents the number of previous owners of the car. The number of previous owners can be considered as one of the factors when estimating a used car's value.

The target variable is:

Price – The estimated selling price of the car.

4. Technologies Used

The following technologies are used:

  • Python – Main programming language.

  • Pandas – Used to create and manage the car dataset.

  • Scikit-learn – Used to implement the Machine Learning algorithm.

  • Linear Regression – Used to predict the car's selling price.

5. Dataset

The project uses a small sample dataset containing seven cars.

Each record contains:

  • Car age

  • Engine size

  • Mileage

  • Number of previous owners

  • Selling price

Dataset

   Car_Age  Engine_Size  Mileage  Owners  Price
0        1         1500    10000       1  25000
1        2         1600    20000       1  23000
2        3         1800    30000       2  21000
3        5         2000    50000       2  17000
4        6         2200    60000       3  15000
5        8         2500    80000       3  12000
6       10         3000   100000       4   9000

The model learns patterns from these historical car records.

6. Working Principle

The Car Price Prediction system follows a basic supervised Machine Learning workflow.

First, a sample car dataset is created using Python and Pandas. The dataset contains car characteristics along with their known selling prices.

Next, the four input features—Car_Age, Engine_Size, Mileage, and Owners—are selected and stored in X.

The Price column is selected as the target variable and stored in y.

A Linear Regression model is then created using Scikit-learn.

The model is trained using the historical car data. During training, the algorithm learns the relationship between the car features and its price.

After training, information about a new car is provided to the model. The trained model uses the learned relationship to estimate the selling price of that car.

7. Python Implementation

The following Python program implements the Car Price Prediction system.

# Import required libraries

import pandas as pd
from sklearn.linear_model import LinearRegression

# --------------------------------------------
# Step 1: Create the dataset
# --------------------------------------------

# Sample car data

data = {
    "Car_Age": [1, 2, 3, 5, 6, 8, 10],
    "Engine_Size": [1500, 1600, 1800, 2000, 2200, 2500, 3000],
    "Mileage": [10000, 20000, 30000, 50000, 60000, 80000, 100000],
    "Owners": [1, 1, 2, 2, 3, 3, 4],
    "Price": [25000, 23000, 21000, 17000, 15000, 12000, 9000]
}

# Convert dictionary into DataFrame

df = pd.DataFrame(data)

# Display dataset

print("Dataset:")
print(df)

# --------------------------------------------
# Step 2: Select input features (X)
# --------------------------------------------

X = df[
    [
        "Car_Age",
        "Engine_Size",
        "Mileage",
        "Owners"
    ]
]

# --------------------------------------------
# Step 3: Select target variable (y)
# --------------------------------------------

y = df["Price"]

# --------------------------------------------
# Step 4: Create Linear Regression model
# --------------------------------------------

model = LinearRegression()

# --------------------------------------------
# Step 5: Train the model
# --------------------------------------------

model.fit(X, y)

# --------------------------------------------
# Step 6: Predict car price
# --------------------------------------------

# New car details:
# Age = 4 years
# Engine Size = 1900cc
# Mileage = 40000 km
# Owners = 2

new_car = [[4, 1900, 40000, 2]]

predicted_price = model.predict(new_car)

# --------------------------------------------
# Step 7: Display prediction
# --------------------------------------------

print("\nPredicted Car Price:")
print(f"${predicted_price[0]:,.2f}")

8. Explanation of the Python Program

The program begins by importing Pandas and Linear Regression from Scikit-learn.

Pandas is used to create the sample car dataset, while Linear Regression is used to predict the continuous price value.

Step 1: Create the Dataset

A Python dictionary is created containing information about seven cars.

For example:

Car Age = 1 year
Engine Size = 1500cc
Mileage = 10,000 km
Owners = 1
Price = $25,000

The dictionary is converted into a Pandas DataFrame.

Step 2: Select Input Features

The four car-related features are selected:

X = df[
    [
        "Car_Age",
        "Engine_Size",
        "Mileage",
        "Owners"
    ]
]

These features are used by the model to predict the car price.

Step 3: Select Target Variable

The Price column is selected as the target:

y = df["Price"]

The target is a continuous numerical value, making this a regression problem.

Step 4: Create the Model

A Linear Regression model is created:

model = LinearRegression()

Linear Regression is suitable because the objective is to predict a numerical value.

Step 5: Train the Model

The model is trained using:

model.fit(X, y)

During training, the model learns the relationship between car age, engine size, mileage, number of owners, and price.

Step 6: Predict a New Car's Price

The new car has the following details:

Car Age = 4 years
Engine Size = 1900cc
Mileage = 40,000 km
Owners = 2

The information is provided to the trained model:

new_car = [[4, 1900, 40000, 2]]

predicted_price = model.predict(new_car)

The model then calculates the estimated selling price.

9. Prediction

Input

Car Age = 4 years
Engine Size = 1900cc
Mileage = 40,000 km
Owners = 2

Output

Predicted Car Price:

$18,500.00

Based on the patterns learned from the sample dataset, the model predicts that the new car has an estimated price of $18,500.00.

This is only an estimated value generated by the Machine Learning model and does not represent an actual market price.

10. How Linear Regression Works

Linear Regression is a supervised Machine Learning algorithm used to predict continuous numerical values.

In this project, the target variable is the car price.

The model attempts to find a mathematical relationship between the input features and the price.

The relationship can be represented conceptually as:

Price =
    (Car Age × Weight)
  + (Engine Size × Weight)
  + (Mileage × Weight)
  + (Owners × Weight)
  + Intercept

During training, the algorithm calculates suitable coefficients for the input features.

The model learns how changes in the car's characteristics are associated with changes in its price.

When a new car is provided, the learned coefficients are used to calculate an estimated selling price.

11. Advantages

The Car Price Prediction project has several advantages:

  • Simple and beginner-friendly.

  • Demonstrates regression concepts.

  • Uses multiple input features.

  • Easy to implement using Python.

  • Demonstrates continuous value prediction.

  • Linear Regression is computationally efficient.

  • Easy to understand and interpret.

  • Can be extended with additional vehicle features.

12. Limitations

The biggest limitation of this project is the very small sample dataset. Only seven cars are used for training, which is not sufficient for reliable real-world price prediction.

Actual car prices depend on many additional factors that are not included in this example.

These factors may include:

  • Car brand

  • Car model

  • Manufacturing year

  • Vehicle condition

  • Fuel type

  • Transmission type

  • Service history

  • Accident history

  • Location

  • Market demand

  • Optional features

  • Current market conditions

Engine size may also have different effects depending on the car's brand and model.

The model also has not been evaluated using a separate test dataset.

13. Future Improvements

The project can be improved by using a larger and more diverse car dataset containing thousands of real-world vehicle records.

Additional features can be included, such as:

  • Brand

  • Model

  • Manufacturing year

  • Fuel type

  • Transmission

  • Vehicle condition

  • Service history

  • Number of previous owners

  • Location

Categorical features such as brand and fuel type can be converted into numerical representations using appropriate encoding techniques.

The dataset can also be divided into training and testing datasets to measure model performance.

Regression metrics can be used to evaluate the model, including:

  • Mean Absolute Error (MAE)

  • Mean Squared Error (MSE)

  • Root Mean Squared Error (RMSE)

  • R² Score

Other regression algorithms can also be compared with Linear Regression, including:

  • Decision Tree Regression

  • Random Forest Regression

  • Gradient Boosting Regression

  • Support Vector Regression

A more advanced system could also use current market data to provide more realistic price estimates.

14. Real-World Applications

Car price prediction systems can be useful in several areas.

Used Car Marketplaces

Online car marketplaces can estimate the expected selling price of a vehicle based on its characteristics.

Car Dealerships

Dealerships can use historical sales data to help estimate suitable prices for used vehicles.

Vehicle Valuation

Financial and automotive businesses can use Machine Learning models to support vehicle valuation.

Buying and Selling Decisions

A prediction system can provide buyers and sellers with an estimated market value to help them make more informed decisions.

Price Comparison

Users can compare the predicted value of a vehicle with its listed selling price.

However, real-world pricing should consider current market conditions and professional valuation rather than relying only on a Machine Learning prediction.

15. Conclusion

The Car Price Prediction Using Machine Learning project demonstrates how supervised Machine Learning can be used to predict the selling price of a vehicle.

The project uses Linear Regression with four input features: car age, engine size, mileage, and number of previous owners. The model learns from historical car data where the actual prices are already known.

For the new car with:

Car Age = 4 years
Engine Size = 1900cc
Mileage = 40,000 km
Owners = 2

the model produces the following prediction:

Predicted Car Price:

$18,500.00

This project provides a simple introduction to regression, feature selection, model training, and numerical prediction. With a larger real-world dataset, additional vehicle features, proper model evaluation, and more advanced algorithms, the system can be developed into a more accurate car valuation and price prediction application.