Study On The Qualification Evaluation Of Construction Contracts Based On The Bp Neural Network

Today, we have lots of problems in qualification management of construction contracts, such as too much details in provisions of qualification management, too much clauses to define qualification level, lack of unify of different standards, illegal affiliated relationship, unable to adjust the structure of construction industry properly, unable to improve the qualification of people who engage in construction industry. On the base of study on resent situation of qualification management in construction contracts domestic and abroad, the thesis tries to construct a model to define a construction contract's qualification by using mathematic method. This model is supposed to be helpful, such as be helpful on implement of related laws and regulations, matins good order of construction market, improves global competitiveness of construction contracts comprehensively, makes better of the system of qualification evaluation.The content of the thesis concludes four aspects. Firstly, the thesis makes clear of the problems of qualification management in construction contracts in China by doing questionnaire survey. Besides, the thesis studies on the resent situation of qualification management in construction industry by doing comparative analysis. All of the research provides an approach to establish an index system which is useful for qualification evaluation. Secondly, the thesis establishes an index system which contains three levels by doing fish bone analysis based on the research of the resent situation of qualification management domestic and abroad. Then the thesis filters the content of the index system by doing questionnaire survey twice and statistical analysis. Finally, the thesis establishes an index system which contains 19 independent indexes reflecting the situation of a construction contract's funding, finance, the quality of personnel and the technological level. Thirdly, the thesis establishes a quantitative model to define a construction contract's qualification by using BP neural network, and the thesis chooses the learning function and parameters by doing a lot of tests. The model uses MATLAB software to calculate, and gets a result of a construction contract's score. And the thesis defines the level of qualification into 5 levels in according to the score. The 5 levels are excellent, good, fine, poor, and bad. Fourthly, the thesis collects information of a construction contract in Tianjin as an example by visiting and surfing on the internet to make a list of qualification evaluation indexes. Then, evaluate the construction contract's qualification level by using the model. At last, we can get the score and the level of the construction contract's qualification.