The Research On The Evaluation For Investment Risks Of Freeway Project Based On Support Vector Machines

With the further and fast development of reform & opening and innovation of investment & financing scheme in China, there has been large improvement on freeway projects investment pattern. The investors of freeway construction and management include not only the central government but also the local governments, private investors, foreign capitals, and chartered right management .etc. The freeway construction has come into a grow space period, and at the same time, large numbers of project companies that preside over independent construction, operation and management are founded by varied investors. Such companies will be facing kinds of risks rising at any time; luckily, the potential risks have gained attention from investors and participants such as loan bank .etc.The investing pattern, processes, characters and the risk arising sources, characters on freeway projects were analyzed in fore part of this thesis. The general analysis processes and methods about the investing risks of freeway projects were also introduced in details, and indicated the shortages of traditional assessment method. Thus, according the shortages, a new risk assessment method which based on support vector machine(sSVM) was suggested here by introducing the newest research results of statistics learning theory. Therefore, such a support vector machines assessment mode using for evaluating the freeway investing risks was established in this thesis, which includes the analysis of risks rising factors and related assessment index system, In case study, calculating procedure and basic assessment processes were given out, and the validity of that mode was also verified.The support vector machines mode has a stable performance and needs fewer sample data, which can improve the objectivity and accuracy on the risks assessment of freeway projects investion. That assessment method takes on the advantages of low cost, easy operating and batch analysis ability. and shows a promising application for its varied advantages.