Modeling real estate dynamics using survival analysis

摘要

This article introduces an adapted version of survival analysis for predicting the period of time a property will stay on market from the listing date to the sale agreement. Survival analysis is a method developed for medical research, in which the dependent variable is the survival time of a patient. Generalizing, the method can be applied in most problems where the dependent variable is time - in our case, the time a property stays on market before selling. Experimental results show that survival analysis brings some advantages when compared to regression analysis on our problem, not only in terms of prediction accuracy: survival curves offer descriptive quantitative views on the influence specific house features have on the variable of interest - the time on market.