Uncertainty Measurement of Interval-valued Decision Information System based on Probability
Interval similarity is the core of interval-valued decision information system,and the related uncertainty measure-ment is of great significance.Aiming at the uncertainty measurement problem of interval-valued decision information system,this paper uses the description of probability to establish high-quality interval similarity and information measurement.Firstly,inter-val similarity based on probability is proposed to establish interval similarity relationship.Then,δ-rough conditional information entropy is proposed by combining roughness and conditional information entropy,and the dual granulation monotonicity of condi-tional attribute and threshold is proved.Finally,numerical experiments verify the effectiveness of correlation measure and monoto-nicity.Based on the new perspective of probability,the new conditional entropy can not only measure the uncertainty caused by the granulation structure,but also can measure the uncertainty caused by upper and lower approximate sets.
Rough setInterval-valued decision systemsProbability similarityRoughnessConditional information entropyδ-rough conditional information entropy