基于量化限制容差关系的数据填充算法研究

RESEARCH ON DATA FILLING ALGORITHM BASED ON QUANTITATIVE CONSTRAINED TOLERANCE RELATION

在对粗糙集模型中的基于容差关系的ROUSTIDA算法和基于量化容差关系的VTRIDA算法进行分析的基础上,提出一种综合量化容差关系和限制容差关系的数据填充方法 VLTA(Valued and Limited Tolerance Algorithm)。VLTA算法克服了ROUSTIDA算法在无差别对象属性值发生冲突情况下无法对相同属性进行补齐的缺陷,改进了VTRIDA算法对容差关系不合理的量化定义。实验表明,VLTA算法填充结果准确率更高,更符合实际。

Based on the analysis of the ROUSTIDA algorithm based on the tolerance relation and the VTRIDA algorithm based on the quantitative tolerance relation in the rough set model, a data filling method based on the integrated quantitative tolerance relation and the tolerance relation was proposed. VLTA （Valued and Limited Tolerance Algorithm） overcame the shortcomings of ROUSTIDA which could not fill the same attribute under the conflict of non- discriminant object attribute value, and improved the quantization definition of VTRIDA to the unreasonable tolerance relation. Experiments showed that the VLTA algorithm had higher accuracy and more realistic results.