Explainable, Multi-Region Price Prediction

Today, Machine Learning has risen to prominence in a variety of fields, from medical diagnosis to stock market trading. Working in the same line, we have focused our paper on one such application of machine learning. In this paper, we have implemented machine learning algorithms to develop prediction forms for house prices based on actual data of real estate in Boston and California. A comparative study of house prices, this paper first introduces the learning algorithms used. Then, multiple regression models and their optimized parameter values are discussed. Basis the prediction results, the paper first compares the performance of Machine Learning models through quantitative metrics, evaluates the importance of attributes influencing the valuation of houses in Boston and California through SHAP values, and finally draws a conclusion on the real estate prices using economic reasoning.