基于遗传BP神经网络的房地产投资估算研究-管理科学与工程专业论文.docxVIP

基于遗传 基于遗传 BP 神经网络的房地产投资估算研究 目 目 录 PAGE IV PAGE IV 万方数据 PAGE III PAGE III 万方数据 Abstract The invest ment estimation is the process of predict ing and determining the expenses incurred for construct ion projects refers to the exist ing materia l and specific met hods. Invest ment has influence on the decision- making of real estate company and budget preparation, fina ncing. Along with the maturity of real estate market,not only the real estate development enterprises,but including real estate consultancy, Banks, the rele vant gover nment departments need a objective and real data of invest ment estimation in feasibility stage. However, due to the invest ment estimate of the real estate related to factors of which is complicated nonlinear relation, so to bring the difficulty. Along with the development of computer parallel processing technology, ma ny technology based on moder n mathematics get a good applica tion,a nd the development of the BP neura l network provides a new way for invest ment estimate. However, the BP neura l network is easily into the loca l minimum faults. Based on the influence factors of build ing insta lla tion cost, estimated cost of construct ion and insta lla tion of genet ic BP neura l network model, this model using genet ic algor ithm to optimize the BP neura l network weights and threshold, thus avoid ing the BP neura l network into the loca l minimum value, but also improves the accuracy of estimates. The ma in research work for: Using genet ic algor ithm to optimize the weights and threshold of the BP neura l network, thus avoid ing BP algor ithm into the loca l minimum value. Based on the releva nt data of a company in Beijing, construct ing GA-BP nerve network model of build ings insta lla tion cost, and using MATLAB software. Results show that the genet ic algor ithm to optimize the BP neura l network is feasible. Keywords: real estate project,t he BP neura l network ,genet ic algor ithm, construct ion and insta lla tion