Project Abstract

Car prices are influenced by a multitude of intricate factors, making the process of purchasing a dream car daunting, especially for rarer and more exclusive models. In light of economic uncertainty and volatility, the imperative for well-informed decisions has become increasingly urgent in recent years. Leveraging machine learning techniques, artificial intelligence offers a promising solution to empower customers with the confidence to make informed decisions when buying their dream cars. This study aims to assess the feasibility and effectiveness of regression machine learning techniques for predicting car prices. In this study, the machine learning models employed included Random Forest, XGBoost, LightGBM, Gradient Boosting Machines, and a custom implementation of Random Forest. By providing a comparative analysis of these five models, this study uncovers key insights into their predictive capabilities. Additionally, a user-friendly front end interface was developed using the Python library Tkinter to allow interaction with the models. The industry sponsor generously provided a large dataset for the models to analyse and derive insights from. Data pre-processing was carried out to ensure the models were trained on accurate data. The main findings reveal that XGBoost is the most effective model at predicting car prices with a remarkably high r-squared evaluation score of 0.9679. Random forest was the second best with a high r-squared evaluation score similar to XGBoost of 0.9530. The custom implementation performed the poorest, with a notably low score of 0.1720. The findings confirm the feasibility and effectiveness of using regression machine learning techniques for predicting dream car prices. This study emphasises the practical significance of machine learning models in real-world applications of artificial intelligence.

Keywords: Ai, Machine Learning, Predictive Modelling

 Conference Details

Session: Poster Session B at Poster Stand 16

Location: Sir Stanley Clarke Auditorium at Wednesday 8th 09:00 – 12:30

Markers: Matt Roach, Simon Robinson

Course: BSc Computer Science, 3rd Year

Future Plans: I’m looking for work