因素空间理论下多目标因果分析的降维算法

Dimension reduction algorithm of multiple targets causal analysis based on factor space theory

为降低利用多目标因果分析法进行因果规则提取的计算繁杂性,给出多目标因素条件下降维映射和钻入决定度降维的定义,对目标因素进行合取,根据选定目标因素的钻入决定度判断结果,进行目标因素的约简,得到了目标因素的降维算法;在目标因素的降维算法的基础上,研究了目标因素之间的关联性,并给出了关联指标定义,结合多目标因果分析法,提出多目标因果分析降维算法.研究结果表明:多目标因果分析降维算法具有实用性和有效性.研究结论为多目标优化带来一些新空间.

In order to reduce the complexity of the calculation of causal rule extraction using multiple targets causality analysis method,the definition of multiple targets factor condition reduction dimension mapping and drill in decision degree reduction dimension are given.Target factors are combined.Target factors are reducted according to drilling determination of selected target factors to judge results,and the target factor reduction algorithm is obtained.On the basis of the dimension reduction algorithm of the objective factors,the relevance between the objective factors is studied,the definition of the correlation index is given,and the dimension reduction algorithm of the multiple targets causal analysis is obtained by combining the multiple targets causal analysis method.The research results show that the dimension reduction algorithm of the multiple targets causal analysis is practical and effective.The research results bring some new space for the multiple targets optimization.