The Research Of Establishing A Construction Cost Information Prediction Platform Based On Artificial Neural Network

At present, most big cities of our country have their own building cost of projects information management website, but, have not built the entire system's information platform.The construction cost information provided is one display-like way, the data exchange are few, can only satisfy the project need of oneself provinces or cities scope, But in the same platform in the same region to deal with non-project, and called other parts of the project cost information, but to predict some certain information or engineering works for information analysis are hard to achieve. Therefore, based on such a reality, we attempt to create a more system-wide information platform, prominent platform function, use neural network works on certain information forecast.This article is the establishment research of projects information forecasts platform based on the artificial neural networks. The first consideration is the establishment of a more system-wide project cost information platform, Information platform includes as far as possible the project cost information of various regions of the country, Forms the system completeness, as far as possible to achieve data exchange. Manifests the platform in the complete information platform's foundation the superiority, establishes an information forecast module using the neural network, carries on the building cost of projects information forecast fully using the platform rich database resources. Simultaneously the platform also increased certain construction cost index analysis function, carries on the corresponding index analysis to some stage or some single project, thus gives the construction cost personnel or the construction cost unit comprehensive project information.Article main research content①Artificial neural network theory researchElaborated the artificial neural networks' history, the concept, the characteristic as well as artificial neural networks' structure and the study way, emphatically the analytical of BP neural network's basic principle and it's study way, lay the foundation for the platform forecast module.②Platform functional analysisThis platform's function divides into two levels: Basic function and major function. The basic function is to manifest this platform is an information platform, may carry on the inquiry and the consultation through the basic function to various local construction cost information, namely the general platform function we usually see at present. But the major function is manifests the difference of this platform, the platform has fruitful database resources, and the functional module technology is novel, as to use the rich database resources and the advanced forecasting technology fully, we can use the member function, forecast the building cost of projects information.③Platform construction researchWith theoretical foundation and constructing a platform for clear thinking, how carry such a theory into reality involves many aspects. Here, start from the platform forecast principle, a modular platform to run on the platform, then to how to achieve platform, a stage of a phase of analysis solution, and ultimately achieving platform functional and simple, smooth realization of the objective.④Case studyBy collecting the actual cost information of those projects that had been completed of Chongqing, carries on to the platform forecast function trains and simulates the examination, carries on to the platform forecast function tries,optimization, and ultimately achieving the smooth operation platform to ensure that the platform for the accuracy of projections.