Font Size: a A A

Study On House Price Predict Based On BP Neural Network By Genetic Algorithm

Posted on:2020-06-19Degree:MasterType:Thesis
Country:ChinaCandidate:W GaoFull Text:PDF
GTID:2428330596478558Subject:Computational Mathematics
Abstract/Summary:PDF Full Text Request
With the development of the real estate industry,the state,society and individuals are paying more attention to housing prices.There are many factors affecting housing prices.In addition to the regulation of the state,there are still some uncontrollable factors.Among them,lighting,households,and environment all have a certain degree of influence on housing prices,resulting in a large change in the real estate market.Therefore,it is of practical significance to study a model with high precision forecasting house price,and neural network has become a new type of prediction method.Based on the theory of BP neural network and MATLAB toolbox function,506 sets of data related to Boston house prices were selected to normalize the data.406 sets of data were randomly selected as training data,parameters were adjusted,BP neural network model was established,and the house price was predicted by the trained network.However,although the traditional BP neural network has strong nonlinear mapping ability,it takes a long time in the prediction process,the prediction effect is poor,and it is easy to fall into local optimal.The traditional genetic algorithm has poor search ability and is easy to fall into local optimal.Aiming at the above problems,the genetic algorithm is improved by immune cloning algorithm and niche technology.The improved genetic algorithm can optimize the initial weight and threshold of BP neural network,so as to optimize BP neural network.Under this condition,the GA-BP network model,the NGA-BP neural network model,and the CGA-BP neural network model were established to predict house prices.Through the function in the MATLAB toolbox,406 groups in the source data are randomly selected as input data to train the network.According to the designed network model,the remaining 100 sets of data are tested and analyzed.The experimental results show that compared with the traditional BP neural network model,the GA-BP network model has less error,higher prediction accuracy and better training times.The BP neural network with improved genetic algorithm is better than the GA-BP neural network model.The prediction accuracy of the NGA-BP neural network model is high,and the simulation time of the CGA-BP neural network model is small.
Keywords/Search Tags:Housing Price Forecasting, BP Neural Network, Genetic Algorithm, Improved Genetic Algorithm, Error Analysis
PDF Full Text Request
Related items