基于功效系数法与BP神经网络的造船业风险预警研究

将功效系数法和BP神经网络相结合,从汇率、利率、出口退税、钢材综合价格指数、行业平均工人工资、设备价格、新船价格指数、BDI指数(八个对造船业影响最大的指标)来综合衡量中国造船业的风险;运用功效系数法得出历史数据的警情,用历史数据训练BP神经网络预警模型;通过一组数据进行检验,并对2011年前的造船业风险进行了预警.研究表明:该方法能够对造船业风险进行测评并预报警情,对防范我国造船业风险具有一定的现实指导意义.

Research on Early-warning of China's Shipbuilding Industry Risk Based on Efficacy Coefficient Method and BP Neural Network

In this paper the efficacy coefficient method and BP neural network is combined, then China's shipbuilding industry risk is measuned comprehensively from the most important eight aspects: exchange rates, interest rates, export tax rebates, the composite steel price index, industry average wages, equipment price, price index for new ships, BDI index. What's more, the historical data situation is drawn by the use of efficacy coefficient method, and an early-warning BP neural network is trained with the historical data.Then the model is tested through a set of data. Finally the early warning to the shipbuilding industry by the end of 2010 is given. It is showed that present method is able to carry out an assessment of shipbuilding industry risk and forecasts an alert, so it has certain practical significance to prevent industry risk for China's shipbuilding industry.