Abstract:Information system (IS) is a significant model in the field of artificial intelligence. Information structure is not only a research direction in the field of granular computing (GrC), but also an important method to study an IS. A multiset-valued information system (MVIS) refers to an IS where information values are multisets. A MVIS can be seen as a model that is the result of information fusion of multiple categorical ISs. This model helps deal with missing values in the dataset. This paper studies information structures in a MVIS on the view of GrC and consider their application for uncertainty measurement (UM). First of all, some notions of multisets and probability distribution sets (PDSs) are proposed. Naturally, relationships between multisets and PDSs are researched. Then, the concept of a MVIS based on the notion of multisets is given, and the internal structure of a MVIS is revealed by an incomplete information system (IIS). Furthermore, tolerance relations in a MVIS are defined by using Hellinger distance, and tolerance classes are obtained to construct the information structures of a MVIS. Considering the association of information structures, relationships between information structures are raised from the two aspects of dependence and separation. Moreover, some properties between information structures are provided by using information distance and inclusion degree. Finally, four UMs as the applications of information structures are investigated, and comprehensive experiments on several datasets demonstrate the feasibility and superiority of the proposed measures. These results will be helpful for establishing a framework of GrC in a MVIS and studying UM.

Information structures and uncertainty in an image information system

Information structures and uncertainty measures in a fully fuzzy information system

Information structures and uncertainty measures in a hybrid information system with images

New measures of uncertainty for an interval-valued information system

Uncertainty measurement for a fuzzy relation information system

Information Structures and Uncertainty Measures in an Incomplete Probabilistic Set-Valued Information System

Information Structures and Uncertainty measures in a Hybrid Information System: Gaussian Kernel Method

Measures of uncertainty for a distributed fully fuzzy information system

Uncertainty measurement for a covering information system

Measures of Uncertainty Based on Gaussian Kernel for a Fully Fuzzy Information System.

Uncertainty measurement for a fuzzy set-valued information system

Measures of Uncertainty for an Incomplete Set-Valued Information System With the Optimal Selection of Subsystems: Gaussian Kernel Method

Measures of Uncertainty Based on Gaussian Kernel for Type-2 Fuzzy Information Systems

Information structures in a multiset-valued information system with application to uncertainty measurement

Uncertainty measurement for incomplete lattice-valued information system

Uncertainty measurement method for information system based on fuzzy neighborhood rough set

Uncertainty Measurement for Fuzzy Set-Valued Data

Measures of Uncertainty for a Four-Hybrid Information System and Their Applications

Approximations and uncertainty measures in incomplete information systems

Uncertainty Measures in Ordered Information System Based on Approximation Operators

Information Structures in a Fuzzy Β-Covering Information System