论文部分内容阅读
针对目标案例与历史案例的非一致属性集合问题,基于匹配度生成子案例库,使用基于核密度的群体序关系分析法对目标案例各属性进行赋权.利用符号型属性和数值型属性的特征,通过不同相似度计算方法分别计算各属性下的相似度,集结各属性相似度信息和权重信息得到非一致属性集合的历史案例与目标案例的相似度,最终得到应急方案的推荐结果.最后通过算例表明了该方法的有效性和可行性.
Aiming at the non-uniform attribute set problem of the target case and the historical case, a sub-case library is generated based on the matching degree, and the group case relation analysis method based on kernel density is used to empower each attribute of the target case. By using the characteristics of the symbolic attribute and the numeric attribute , The similarity of each attribute is calculated respectively by different similarity calculation methods, the similarities between the historical case and the target case are obtained by assembling the similarity information and the weight information of each attribute, and finally the recommended result of the emergency scheme is obtained finally. The example shows the effectiveness and feasibility of the method.