Government Investment Project Risk Analysis And Evaluation

Project risk assessment is a key step in risk management,which is a process to make a reasonable risk assessment of the overall level of project. As the government-funded projects with the construction of large-scale, varietied investment, uncertainties, long construction cycle, and that social featureof far-reaching impact, the projects of government requires more effective risk assessment than that of the average project risk assessment. The research on the government-funded projects has both the theoretical and realistic meanings.In this paper, combining the characteristics of the government-funded projects, the sources of risk from the construction agent and the relative affecting factors that are forecasted at the early stages of decision-making in the project were identified and analyzed to establish a government-funded project risk index evaluation system. The article is based on the principles of artificial neural network to establish a three-layer BP neural network model using the Matlab tool to program.In the new net,sample data is for the network training, and testing, finally it caculates a comprehensive assessment of the value of the risk,and achieve the practical application. This model,which is based on historical data rather than a professional experience in risk decision-making method, provided a more scientific basis for the latter part of the data.The risk assessment model of government investment projects based on BP neural network, runs on a fixed learning algorithm, compares the network training effect, and achieves the most suitable number of hidden layer nodes. Secondly, by analysising and researching eight different BP learning algorithms it determines the optimal learning function of this model. Togethered with other relevant structural parameters, the best model is building. Test results are basically consistent with the actual expectations, and the results are satisfactory. At the same time, the article also proposed to improve the BP neural network method, by the introduction of the genetic algorithm. Weights and thresholds of the BP neural network are pre-adjusted to avoid the BP neural network into a local minimum, and thus to achieve better network of fitting effect.Research in this practice proved the feasibility of risk assessment using artificial neural networks for government investment projects. It is provided new method or idea that is about a combination of modern analysis of BP algorithm in the risk assessment.