基于含熵期望曲线的云模型相关性度量方法

Pertinence measurement of cloud model by using expectation-entropy curves

针对现有基于云模型的相关性度量方法缺少必要约束条件的问题,提出一种基于含熵期望曲线的云模型相关性度量方法.将云模型中超熵期望曲线的与区域和或区域的面积比作为相关性的度量基准,解决云模型的区间约束以及半云度量问题.利用3熵(3σ)约束增加云滴的聚集,减少计算开销.将超熵纳入计算,考虑云滴厚度对云模型相关性的影响.本方法克服了面向随机云滴的距离度量方法和数字特征变换方法中存在的计算复杂度高、结果不稳定的问题,同时满足了三类约束的实际计算需求.实验表明该方法能够客观有效地对云模型相关性进行度量,并在基于云模型的系统评价任务中得到了验证.

To improve pertinence measurement of cloud model because of its lack of necessary condi- tion restrictions, an improved approach based on expectation-entropy curves was proposed. The ratio of ＂And Area/Or Area＂ of cloud model was used as benchmark to solve the interval restriction and semi-cloud measurement problems. Using restriction on three times of entropy restriction to increase assemble of droplets, calculating cost was decreased. Considering the emphasis of cloud thickness, hyper-entropy was introduced to compute the pertinence. Comparing with distance measurement method that oriented random droplets and digital characteristics conversion method, this approach e- liminates high time complexity and high instability of outcomes. At the same time, it meets practical requirements under three types of restrictions. Test shows that this approach can measure pertinence of cloud model objectively and effectively, and can be validated in system evaluation tasks by using cloud model.