Research On Mass Appraisal Of House Prices Based On K-nearest Neighbor Algorithm And Hedonic Price Model In The Background Of Big Data

With the continuous growth of China's real estate industry,the level of urbanization continues to increase,the size of the existing housing market continues to expand,and the demand for real estate evaluation services continues to increase.The increasing demand for evaluation has created new challenges for evaluation activities and also forced the evaluation level of evaluation agencies to improve.High-level real estate assessment can not only provide the basis for government and corporate decision-making,but also reduce transaction costs for both parties in real estate transactions,improve transaction efficiency,and promote the healthy development of the real estate market.For real estate appraisal,real estate is highly heterogeneous.The accuracy of appraisal largely depends on comparable case selection similar to the property to be evaluated,and there are many factors affecting real estate prices,difficulty in collecting factors,and strong subjectivity of appraisal.Assess business development.Mass appraisal technology and big data can solve the above problems.Mass appraisal technology relies on a unified evaluation algorithm to achieve real estate evaluation and improve evaluation efficiency.At present,real estate big data databases established by government departments,geographic information providers,and real estate transaction agencies can provide data support for decision-making.Therefore,fully researching the mass appraisal technology and big data can improve the problems of real estate evaluation,such as difficult data collection,low evaluation efficiency,and large artificial influence,and improve China's real estate evaluation level.This thesis first reviews and evaluates the current status of research at home and abroad.Aiming at too much emphasis on algorithms in domestic mass appraisal research,it ignores the basic evaluation methods and stock price specifications,and the evaluation methods are inconsistent with the characteristics of real estate big data.Combine.The data collection section clearly evaluates the influencing factors and factor quantification methods,and then in accordance with the principles of the market comparison method commonly used in residential real estate valuation,in the case selection process of comparable cases,the traditional manual selection method is replaced by an algorithm.In combination with the selected comparable cases to evaluate the case,the traditional manual evaluation method is replaced,and the mass appraisal system of this thesis is finally established.Through comparative analysis,this thesis uses the K-nearest neighbor algorithm as a comparable case selection algorithm,and the hedonic price model as the valuation model.In the case study part of the thesis,this thesis uses the Nanshan District Shenzhen residential transaction data for case analysis.Through case demonstrations of data collection,case selection,and price evaluation,the feasibility and rationality of the mass appraisal system established in this thesis is verified.Through theoretical and case analysis,the main conclusions of this thesis are as follows: Through a comparative analysis of various algorithms and economic models in batch evaluation,choosing algorithms and models that conform to the characteristics of real estate evaluation can improve the fairness and efficiency problems of traditional single case evaluation.On the premise of meeting real estate specifications,the combination of real estate big data and batch evaluation can enrich the number and dimensions of comparable cases in batch evaluation,and overcome the characteristics of low evaluation efficiency and poor accuracy.