论文部分内容阅读
为解决定性与定量属性数据混合及属性值缺失的应急案例决策问题,建立一种基于证据推理的相似度测算方法。通过给出目标案例与历史案例的属性距离定义,借助证据推理制定属性距离的基本置信度评价和融合规则,并得到案例的置信度分布形式。将计算案例的平均效用作为案例相似度的评估值,提取相似度满足阈值且最大的历史案例,将其应急方案作为当前突发事件的最优备选方案。最后,将该相似度测算方法运用到火灾和地震应急案例中,求解与当前突发事件最相似的历史案例。结果表明:利用区间距离方法计算属性距离,能有效表示不确定属性的距离;用基于证据推理的相似度测算方法,能有效度量复杂数据情形下的应急案例相似度。
In order to solve the emergency case decision-making problem of mixture of qualitative and quantitative attribute data and missing attribute value, a method of similarity calculation based on evidence reasoning is established. By giving the definition of the attribute distance between the target case and the historical case, the basic confidence evaluation and fusion rules of attribute distance are established by means of evidence reasoning, and the distribution form of the confidence of the case is obtained. The average utility of the calculated case is taken as the evaluation value of the case similarity, and the historical case whose similarity satisfies the threshold and the maximum is extracted, and the emergency plan is taken as the optimal alternative for the current emergency. Finally, the method of calculating the similarity is applied to fire and earthquake emergency cases to solve the historical case which is the most similar to the current emergency. The results show that the method of interval distance can be used to calculate the distance of the attribute, which can effectively represent the distance of the uncertain attribute. The method of similarity based on evidence reasoning can effectively measure the similarity of emergency cases in complex data.