一种基于案例属性特征的区间相似度改进算法研究

基于案例推理的关键是确定问题案例与案例库中的案例属性特征的相似度，针对相似度的算法问题，在分析比较多种已有算法的基础上，提出了一种解决感知信息和案例库中的案例属性特征均为不确定或者不完整信息的区间相似度计算模型，论证了该模型的计算结果适用于特定区间内微小区间相似的问题。结合某型艇抗沉辅助决策系统的预案推理问题，对已有的算法求解结果进行比较，证实了该模型提高了案例相似度计算结果灵敏度。

A Research for Mended Range Attributes Similarity Calculation Models of Case-based Reasoning

Ascertaining the attributes similarity between the problem case and cases in the case base is the key of case-based reasoning. Considered the problem of similarity calculation models, analyzed several existent models and brought out a mended range similarity calculation model. This model can deal with the problem that perceptive information and attributes information of cases in case base are both uncertain and incomplete. Demonstrated the calculation model be the same with tiny ranges similarity in the certain range. United the plan reasoning problem of a certain submarine confront flooding assistant decision-making system, compared the calculation results deduced by several existent models, proved that the calculation model can enhance the case similarity sensitivity.