Car Price Prediction

Predicting Used Car Prices

Goal

• Utilize publicly available data to develop models that accurately predict used car prices.
• Compare various predictive modeling techniques to determine the best-performing model based on predictive power (Lowest Mean Squared Error - MSE).

Description

• Exploratory Data Analysis (EDA):
  • Data cleaning, handling missing values, feature engineering.
  • Visualizing distributions, correlations, and key insights.
• Predictive Modeling: Comparing multiple machine learning methods:
  • Decision Trees
  • Neural Networks (NN)
  • Regression (LASSO/Ridge)
  • Gradient Boosting (XGBoost)
• Implementation: Developed in R, utilizing the following packages:
  • tidyr, ggplot2 (Data wrangling & visualization)
  • tree (Decision Trees)
  • glmnet (Regression)
  • nnet (Neural Networks)
  • XGBoost (Gradient Boosting)

Progress

Completed in 2023.