https://github.com/nickenshidqia/predict_price_of_used_car_using_linear_regression

Build a machine learning model that can predict the price of used car
https://github.com/nickenshidqia/predict_price_of_used_car_using_linear_regression

car-price-prediction data-science linear-regression machine-learning prediction

Last synced: 11 days ago
JSON representation

Build a machine learning model that can predict the price of used car

• Host: GitHub
• URL: https://github.com/nickenshidqia/predict_price_of_used_car_using_linear_regression
• Owner: nickenshidqia
• Created: 2023-12-18T14:52:35.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2023-12-18T15:42:11.000Z (about 1 year ago)
• Last Synced: 2024-11-05T21:47:18.247Z (about 2 months ago)
• Topics: car-price-prediction, data-science, linear-regression, machine-learning, prediction
• Language: Jupyter Notebook
• Homepage:
• Size: 1.91 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Machine Learning Project Using Linear Regression to Predict Price of Used Car

## Project Description

**Problem :**  

Buying or selling a used car can be a complex process, and determining a fair market value for a used car is often subjective. This project addresses the challenge of predicting the price of a used car based on its specifications.

**Challenges :**  

Build a machine learning model that can predict the price of used car

## Project Goal

By leveraging machine learning techniques and historical data, the goal is to develop a model that provides accurate and reliable estimates of a used car's market value, taking into account various features and attributes.

## Tools & Library Used

[](https://www.python.org/)  

[](https://jupyter.org/)  

## Project Result

[Click here to get full code](https://github.com/nickenshidqia/Predict_Price_of_Used_Car_Using_Linear_Regression/blob/648fc30d7b33b1878e4afb321b29416d5d5ccf30/Used%20car%20price%20prediction.ipynb)

### Dataset

    

 

- Brand ==> BMW is generally more expensive than Toyota

- Mileage ==> the more car is driven, the cheaper it should be

- EngineV ==> sports car have larger engines than economy cars

- Year of production ==> the older the car, the cheaper it is, with exception of vintage vehicles  

 

### Exploring the PDF's (Probability Density Function)  

  

  

**Price**  

- this is dataset about used car, therefore $300,000 is an excessive price

- Ouliers are great issue for OLS ==> remove top 1% of distributions !!  

 

**Engine Volume**  

- The situation with engine volume is very strange

- From google search, Car engine volumes are usually below 6.5l, so we're gonna select data below 6.5l  

  

**Mileage & Year**  

- Mileage has positive skewed distribution

- Year has negative skewed distribution

- We're gonna handling the outliers  

  

### Check linearity  

  

  

- OLS assumption of linear regression is linear, but from the plot, price is exponentially distributed

- good transformation in that case is a log transformation  

 

### Check Multicollienarity    



- Year has the highest VIF, drop 'Year'

- VIF 'Year' = 10.35 > 10 ==> unacceptable

### Linear Regression Model



- For high prices, we have a higher concentration of values around the 45 degree line ==> Our model is very good at predicting higher prices.



- Our model is explaining 75% of the variablity of the data, relatively good result

  

  

**Weights Interpretation :**  

- A positive weight shows that as a feature increases in value, so do log_price and 'Price' respectively

- Example : EngineV, the bigger the Engine volume, the higher the price

- A negative weight shows that as a feature increases in value, log_price and 'Price' decrease

- Example : Mileage, the more a car is being driven, the lower the price gets

- Dummies are compared with the benchamrk dummy

- A positive weight shows that the respective category is more expensive than the benchmark

- Example : respective category (BMV) is more expensive than the benchmark (Audi)

- A negative weight shows that the respective category is less expensive than the benchmark

- Example : respective category (Mitsubishi) is more expensive than the benchmark (Audi)

- The bigger the weights the bigger the impact

- Mileage is te most prominent feature in the regression. It is more than twice as important as Engine Volume