Study On Forecasting Model Of Real Estate Market In Qingdao

The real estate has become one of pillar industries of the national economy.Realestate market’s stable healthy development plays an important role in the nationaleconomy, and let it has a fast and steady development.In recent years, since thedevelopment of the real estate market, many problems have aroused.For the presentexisted problem, all levels of government are finding solutions to theseproblems,Theses obtain the certain result, also appeared some deviations or thenegative effect.Many experts point out that the root causes of macro-control of theproperty is not clear policy results in China takes on the situation of "taking stopgapmeasures" because of lack of a the reasonable and scientific forecast and earlywarning system of the real estate.This article is precisely based on this background,We hope to establish aneffective model to the predictions of real estate market in qingdao,analysis andexplain the data we get. it is a meaningful attempt to build the prediction model ofreal estate market in qingdao.We have observed a strong correlation between the real estate and manyfactors.so building the prediction model of real estate market will not be easy.In orderto combine the advantages of various models and overcome their disadvantages, thispaper proposes a combination model for forecasting based on conditional entropy.sothe model was set up by employing the two kinds of thoughts,It can be viewed fromtwo aspects: trend predictions and factor analysis.The paper firstly constructs a model to carries on the trend estimate, which isused for an improvement gray forecast model;Then it constructs another model byneural network; Finally it set up a combination model for forecasting based onconditional entropy.The model show that the technical measures taken are feasibleand with certain advantages.We hope the method can exactly predict the developmentof real estate market in qingdao.Following the ideas above, This article in predecessor’s research foundation, willcarried from some ways:(1) how to specify predictors of real estate market: Judging from its selecting criteriaand the accuracy and facility of data.the destination of predicting includes underconstruction floor space、real estate development investment and commercialhousing sales area. (2) how to selecte single prediction models：There are many kinds of predictionmethods，which kinds of prediction method is applicable for real estate marketforecasting? we select the grey forecasting and neural network forecasting to setup single prediction models, then construct the combining forecast model basedon conditional entropy. With those modes, the system have very goodcharacteristic.(3) how to selecte the combining forecast model：according to single predictionmodels,Which the combining forecast model is more appropriate. it set up acombination model for forecasting based on conditional entropy and thecombination model set advantages of gray theory and neural network. the modelis more suitable for the development of real estate market in qingdao.