Prediction and Analysis of House Prices in Boston Based on Regression Model

Abstract

In artificial intelligence learning, Boston housing price forecast analysis problem is a classic regression problem. Based on the housing price information collected by the U.S. Census Bureau in Boston, Massachusetts. this paper divides the housing price data set of Boston and builds the regression model linear regression, decision tree regression and support vector machine regression SVR and trains the data set, so as to obtain the relationship between different data related to Boston house price, and use this relationship to connect all data, it can finally predict the future house price trend in Boston and display it through visual operation. Through three regression model prediction value, respectively compared with the actual value, the trend of overall and actual and estimated values of the same, but there is a certain error, especially when spot prices higher or lower, often cannot get accurate forecast, so the data for the selection of the characteristic value still exists space for improvement, future study needs to get more data and the characteristics of abundant data.