The Application Of RBF Neural Network Prediction Algorithm Based On Principal Component Analysis

Real estate, from the narrow level of understanding, includes real estate occupiedland, aboveground and underground parts of various buildings; and real estate refersto a property and it includes real estate itself, wealth, property or bringing their owncause various interests, as well as a variety of business activities in the real estatefrom the broad level of understanding, which named the level of understanding of theeconomic benefits.,.Under the socialist market economy, China’s real estate industry has a reasonableoperating system, and it reflected the thriving prosperity. The real estate industry is acomprehensive industry, which includes the following individual process industry, thefirst step is land development, the second step is the construction of houses, the thirdstep is housing maintenance, and the fourth step is a housing management. It is alsorelated to other industries, such as the conversion of land use rights paid renamed, thesale of various industrial housing home ownership, rental housing ownership, housingmortgage real estate loans caused by the real estate and so on.Real estate prices have been the focus of issues of public concern, and real estateis a commodity which has prices, and prices fluctuate with the value. Real estate is aspecial commodity, real estate prices continued to rise, it is the result of many factorsproduced. Real estate is demanded, then there is the market, and the price is againdetermined by demand and supply simultaneously. Real estate is different from othercommodities, it also affected by other factors, such as the government’s macro-control,inflation, exchange rate changes, and so on.Data of this study came from "China Statistical Yearbook" which has the realestate-related statistics. This paper established the real estate forecasting model to study the relationshipamong prices and more than one real estate-related indicators of their cities, whichcan. forecast the next year’s price the index data relating to real estate prices in thisyear’s.This paper studies the theory of algorithms, which includes principal componentanalysis, multiple linear regression algorithm and RBF neural network algorithm.Principal component analysis algorithm, means using lower-dimensional thinking, itwill be converted into a small number of multiple process variables main ingredient.RBF neural network that radial basis function (Radial Basis Function) neural network,a good performance front network. It has the best approximation performance, itstraining method is fast and easy, and local optimization problem does not exist.Using these algorithms, the paper proposed RBF neural network predictionmethod based on principal component analysis,. Firstly the algorithm made principalcomponent analysis of the data to the purpose of dimensionality reduction; secondly itpredicted the data based on RBF neural network. The algorithm combines theadvantages of principal component analysis and RBF neural network algorithm, theuse of principal component analysis achieved the dimension reduction, so that theinput node RBF neural network was reduced, and mainly information wasconcentrated after eliminating redundant information.The multiple linear regression algorithm, RBF neural network algorithm and theRBF neural network prediction algorithm based on principal component analysis wereall used to establish a predictive model of real estate prices. Uing the three predictivemodeling, it obtained comparative analysis of the results, and it found that thepredicted results of RBF neural network prediction algorithm based on principalcomponent analysis in this article proposed were better.