一种具有相依关系的二维云推理方法及其在预测中的应用

基于云模型的定性知识推理通过构造规则发生器实现不确定性推理。然而传统的二维云推理方法并未考虑条件云之间的相互影响,故此对传统方法进行改进,通过条件云云滴的乘积作用,构造相应的规则发生器,提出一种基于相依条件云的二维云推理方法。将该方法应用于美国GDP预测,预测结果表明该方法的效果明显优于传统方法,进而表明该方法可行且有效,并在一定程度上弥补了传统方法的不足。

Uncertainty Reasoning Based on Related Planar Cloud and Application in Prediction

Qualitative reasoning based on cloud model implements uncertainty reasoning by constructing cloud regulation generators.But the traditional two-dimensional cloud reasoning methods do not take notice of the interaction between the condition clouds,as an improvement on the traditional method ,a new approach to planar cloud reasoning method based on related condition clouds were suggested in this paper,which was followed by the focus on the product of cloud droplets and the construction of corresponded regulation generators.Then this new approach was applied in the prediction of GDP of America.The prediction results show that the new approach outperforms the traditional one,which indicates that the new method is feasible and effective,and to some extent remedies the shortcomings of traditional way.