A price prediction method In real estate market

Abstract

Current housing price prediction usually employs hedonic or repeat-sales models. The objective is to build a statistical model which is more focused on statistic methods. Neither ordinary nor regularized regression model haven been applied to the field of real estate, even though they are rather well-known statistical procedures. This thesis concludes lots of ordinary and regularized regression models. A theoretical review was performed for these models, and Boston Housing data was used to evaluate their performance. The results were found to be reasonable, from a statistical perspective.