Study On Hedonic Price Appraisal Method Of Second-hand Housing

Along with the development of the national economy, economic activities related to real estate become more and more frequent, as the demand of property assessment is even greater. From the perspective of market participants or our government to levy taxes, to measure residential prices accurately is an eternal topic.Currently, in appraisal practise three traditional assessment methods such as the market comparison, cost and income approach need excessive reliance on the assessors' experience,but little Mathematical model,so the cost in actual application is high. Dawning on the experience of foreign real estate assessment and base-taxes devolopment, this paper introduced the Hedonic Price Model to study and appraise housing price from the consumers' inherent demand of the property characteristics.In traditional Hedonic Price Model, the most commonly used method is multi-parameter regression methods. The function selection depending on assumption has a great impact on price assessment and always leads to great deviation. This paper introduced the latest research fruits of statistical learning theory called support vector machine to the study of regression and prediction compared with the traditional Hedonic Price Model. On the basis of the actual second-housing transactions data, this paper accessed 320 examples in the study area through surveys and built a Hedonic Price Model.The comparative study of traditional parameters regression and support vector machine regression, showes SVM leads to higher precision in prediction and appraisal, and has some practical value.