Identifying a group of subject experts using formal concept analysis

• Published in: 2016 IEEE 8th International Conference on Intelligent Systems (IS) pp. 464 - 469
• Main Authors: Boeva, Veselka, Angelova, Milena, Boneva, Lilyana, Tsiporkova, Elena
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.09.2016
• Subjects: Context; Data mining; Electronic mail; expert finding; Formal concept analysis; health science; knowledge management; Lattices; LinkedIn; Semantics
• DOI: 10.1109/IS.2016.7737462

Expertise retrieval has already gained significant interest in the area of information retrieval due to multitude of concrete application contexts where search for specific experts is required. In this paper, we introduce a formal concept analysis approach for clustering of a group of experts with respect to given subject areas. Initially, the domain of interest is presented at some level of abstraction by partitioning the domain to a number of subject areas. Next each expert is represented by a vector of contributions of the expert to the different areas. This defines an overlapping partition, which is further analyzed and refined into a disjoint one by applying Formal Concept Analysis. In this way a hierarchical conceptual structure is automatically derived. It consists of concepts where each one represents a subset of experts belonging to a number of subject areas. The set of all derived concepts partitions the experts into a set of disjoint expert areas.