Design for conceptual knowledge processing: case studies in applied formal concept analysis

Conceptual Knowledge Processing (CKP) is a knowledge management and data analysis technique that makes use of conceptual structures. Formal Concept Analysis (FCA) is a CKP methodology that uses lattice theory to represent units of thought, or concepts. When FCA is used in software applications, it makes use of a process called Mixed Initiative. Mixed Initiative breaks down the roles of user and machine, allowing each to play to their strengths. This process allows the computer, which can process vast amounts of data, to produce interaction options from which the user can select. A human can interpret semantic knowledge contained within the data that a computer cannot. This synergy of user and computer allows complex tasks to be performed. Wille [Wil99] proposed ten atomic tasks of CKP which are combined to make these more complex tasks. The ten tasks are exploration, search, recognition, identification, analysis, investigation, decision, improvement, restructuring and memorisation. Individually, these tasks represent facets of interaction with conceptual systems. This thesis uses the ten tasks of Conceptual Knowledge Processing as a framework for experimentation with applications that use Formal Concept Analysis. The applications used for this analysis are MailSleuth, SurfMachine, DSift, ImageSleuth and SearchSleuth. These applications approach various problems, using FCA as the primary knowledge structure and interaction framework. Each application uses various interface components and varying degrees and types of exposure to the FCA structures on which they are based. The connection between CKP tasks and interface exposure is then explored and reported.