With the reform opening policy and the initial establishment of market economy in china, projects had become larger and more complex recently, which led to frequent risk, therefore, it is necessary to improve project management.As construction project needs more time, larger investment, more participations, more complex organization membership, it involves a lot of uncertainty in the whole course. So risk management seems extremely important. While making project decision, a feasible risk evaluation method is required. Hence, it's significance in theory and practice to study risk evaluation method. Firstly, this dissertation introduces the general content of construction project risk management, and emphasis the risk identify and common method in risk evaluation. Some traditional methods of risk evaluation, such as expert analysis questionnaire, Level analytic method, Monte Carlo simulation, fuzzy mathematic method BP neutral network are introduced and their superiorities and shortcomings are discussed. According to these flows, newly development of statically learning theory is introduced, and a project risk evaluation model based on support vector machines (SVM) is proposed.To improve the evaluation efficiency, this paper uses rough sets theory(RS) to preprocess the sample data, construction project risk evaluation model based on RS-SVM. By compared with BP neutral network, the results show that the model is valid and effective. Its superiority and wide application foreground are demonstrated through an example. |