摘要：Second-hand housing market is the barometer of the real estate market since the buyers of second-hand houses usually are those who really want to live, whereas financial investment and speculation are not their goals. So the price and determinants of second-hand houses reflect the real demand of housing market. In this paper KNN related algorithms are applied to study the problems associated with price of second-hand house. It includes using KNN and weighted-KNN algorithms to predict the price, using cross validation method to compute average deviation of prediction algorithm and compare KNN's prediction effect with weighted-KNN's, and using stimulated annealing optimization algorithm to compute the weight values of house attributes and evaluate the relative importances of them. Through the analysis of attribute importance it can show the influences of different house attributes on house price and the main concerns of buyers. All these results can give valuable information for manages, decision makers and appraisers of real estate.
