House price prediction using polynomial regression with Particle Swarm Optimization

Abstract

This paper combines the characteristics of the real Washington DC house price estate market and the theoretical model of housing prices to conduct an empirical analysis and comparison of the current mainstream housing price index prediction models. It is found that the current mainstream model only studies the trend of the housing price index itself, and is not sensitive to the characteristics of the house itself. Therefore, this paper uses a multiple regression model to integrate the advantages of external factors and use the improved housing price composition establishing a multiple regression prediction model with the particle swarm optimization. It not only makes up for the disadvantages of poorly determined housing price regression indicators and lack of statistical data in multiple regression prediction, but also enables the model to reflect the inflection point of housing prices in advance. PSO is used for selection of affect variables and regression analysis is used to determine the optimal coefficient in prediction.