Exploring the Applicability of Formal Concept Analysis on Market Intelligence Data

This paper examines and identifies issues associated with the applicability of FCA on sample data provided by a CUBIST use-case partner. The paper explains the various steps related to the transformation of these data to formal contexts, such as preprocessing, cleansing and simplification, as well as preprocessing and limitation issues, by using two FCA tools currently being developed in CUBIST, FcaBedrock and InClose. The paper demonstrates what is achievable to date, using the above-mentioned tools and what issues need to be considered to achieve more meaningful and intuitive FCA analyses. The paper concludes by suggesting and explaining techniques and features that should be implemented in later iterations of these tools, to deal with the identified barriers. This work has been carried out as a part of the European CUBIST FP7 Project: http://www.cubist-project.eu