Research On Prediction Model Of Construction Cost Based On FPR-ANN

As China entering the WTO, the construction industry is in an increasingly competitive market. To foothold in the competition, scientific cost management for construction enterprises is critical. Project cost forecast is an important part of project cost management. It is an important part of strengthening cost management and the first part of the dynamic controlling and a significant manifestation of intensification scientific management. Reasonable forecast of project cost is a key to construction enterprises'survival and development and it is important for construction enterprises'bidding. The author proposes to use fuzzy pattern recognition combined with BP neural network for forecasting project cost in the current situation. The content of the thesis are as follows:1.According to the completed thesis, the other authors use ANN for forecasting project cost lacking a reasonable basis for selecting factors. With literature review of thesis of the project cost forecasting, the author extracts some factors by analysis and classification. On the basis of this, author works up a project cost forecasting system. It is the input of BP neural network, so is a basic work for BP neural network..2.Another problem is selecting sample projects. The author put forward to using fuzzy pattern recognition to determine the similar sample projects. The selecting principal is selecting projects of high similarity with tested project as sample projects.3.BP neural network is a more mature than other ANN in the development and relatively simple to use. So the author applies it to establish a forecasting system for project cost. It is proved that this approach is feasible by an case.