Explainable AI and Machine Learning Model for California House Price Predictions: Intelligent Model for Homebuyers and Policymakers

Abstract

California's housing and real estate market is one of the most valuable markets in the USA. Many shareholders such as individual homebuyers, home sellers, real estate agents, lenders, and policymakers depend on high-volume information regarding the dynamics at work and their correct estimation. The research project aimed at developing an Explainable AI machine-learning model for California house price predictions. Data on house prices were collected from reliable sources such as California home estate websites, land sites, and public datasets. Features of the data included location, size, number of rooms, area type, availability, sale prices, and oceanic proximity. In this research project, credible, proven, and renowned machine learning algorithms were used most notably, Linear Regression analysis, XG-Boost, and Random Forest. The Random Forest came up quite impressively with a superior accuracy score and low MAE and MSE; thus, it was good for learning the underlying best patterns and relationships that may exist within the data for house price predictions. XG-Boost also did relatively well, showcasing moderately high accuracy and relatively low MSE and MAE, compared to the Linear Regression.