Design Of Forecasting Plan For The Price Trend Of Commercial Housing In Shanghai

Today,with the rapid development of network and information technology,search engines play an important role in information search for home buyers.In a sense,search engines actually act as a bridge between potential homebuyers,real estate developers,and websites that offer a variety of property information.Therefore,from the analysis of the network search data,we can find the psychological expectations and behavior patterns of all parties involved in the real estate market,and these factors will directly affect the real estate price.In Shanghai,housing prices have long been inseparable from finance.Shanghai's housing prices have become among the international high-price cities.Commercial housing is not only a consumer product,but also a financial product.This paper studies and predicts the price index of commercial housing in Shanghai based on text mining and SVR.The data used in this paper comes from three categories: one is the network news and forum question and answer data for text mining to determine the price-related keywords,and the other is the Baidu index network search data after the keyword is determined for modeling learning prediction.Both types of data are automatically obtained using the R language web crawler,and the data is authentic and reliable.The process of establishing the keyword lexicon is to combine the dictionary-based word segmentation and the statistical-based maximum probability segmentation with the machine learning hidden Markov model to perform text segmentation and establish the final keyword lexicon through multiple iterations.After the keyword lexicon is established,the Baidu index data of each keyword is obtained through the web crawler.After the data is preprocessed,the SVR model is selected by the particle swarm optimization algorithm(PSO)after the kernel function is selected with the help of the R language tool.The parameter optimization is performed,and finally an optimal parameter model is obtained,which is obviously superior to the model before the unoptimized parameter.In addition,by using macroeconomic data and using Baidu search recommendation to determine the keyword data to predict the commodity housing price index,and text-based mining to determine the price-related keywords,to obtain the keyword Baidu index search data,according to the Shanghai commercial housing price The index performs PSO-SVR model learning prediction and BP neural network for regression prediction.The results show that text-based mining and Baidu index use PSO-SVR model to predict the housing price index more effectively.