Research On Biding Of Construction Projects Based On Rs And Svm

After China's accession to WTO, the construction market has been standardized gradually, the competition environment has been more and more fierce, and the tension which the contractors have to face up to has increased by degrees. Meanwhile, the going out strategy and the model of promoting economic development by investing infrastructure bring the construction companies tremendous growth opportunities, which means that both challenges and opportunities exist for the companies. Whether the company could develop depends on the ideal projects that the contractor could obtain. The capability to fetch the ideal projects depends on not only the outstanding acting and public relation ability but also the suitable biding strategy, which is the strategy to determine the make-up. The biding decision of construction project which is also called make-up decision is studied by the method of rough set and SVM(support vector machine). The research is practicable for the contractor.After literature review, the index system of the biding decision was constructed. In accordance with the principle of entirety, science and maneuverability, 19 factors which have a significant influence on the make-up are determined. The factors are classified into 5 categories, which are location factor, economic factor, experience factor, entrepreneur factor and project factor. The accurate description and scientific sampling are given at last.The model of biding decision is constructed based on rough set and SVM, which is to solve the problem of make-up decision of the construction projects for the contractor. Firstly, the sample data is calculated by discrete treatment by SOM net method. Secondly, the factors was decreased by the method of decreasing the attributes of the factors with rough set, and kernel function and parameter choice of the SVM is determined by LIBSVM2.89. Last but not the lest, the outcomes of the biding model based on rough set-SVM and rough set-neural network are compared with each other, which proved the method of rough set-SVM is better than the method of rough set-neural network in accuracy.