Comparative Analysis of Housing Price Prediction of Dhaka City Using Machine Learning

EasyChair Preprint 12258, version 1

Abstract

The housing market is rapidly expanding, making it crucial to forecast housing prices for both company and individual consumers. However, several factors affect house price variations. As Bangladesh is an overpopulated country, the sale price of real estate is influenced by several interrelated factors. The size, location, and amenities of the property are important variables that could determine the price. We examined about 8000 houses in the Dhaka and Chittagong Region in Bangladesh as a case study and discussed how the increase in housing prices could vary by each of the contributing components. In this research, we conduct an extensive analysis and investigation of twelve machine learning methods for housing price prediction. Our study offers a comprehensive study for assessing the effectiveness and reliability of machine learning models for predicting home prices. The results showed LGBM to be the second best model with an R2 equal to 85% and XGB to be the best model with an R2 equal to 94%.

Keyphrases: Housing, KNN, Random Forest, machine learning, price prediction

@booklet{EasyChair:12258,
  author    = {Md Saiful Islam Sajol and Mohammad Nizam Uddin},
  title     = {Comparative Analysis of Housing Price Prediction of Dhaka City Using Machine Learning},
  howpublished = {EasyChair Preprint 12258},
  year      = {EasyChair, 2024}}