Formal Concept Analysis Reduction Method Based on Modified Hamming Distance

One of the approaches used in the field of data analysis is related to the theory of concept lattices, also known as Formal Concept Analysis (FCA). This family of methods is able to process and analyze object-attribute input data models. The output of these methods is represented as concept lattice, which is hierarchically organized structure of concepts. The concept represents groups of objects based on the presence of their shared attributes. One of the main problems of FCA methods is a large number of generated concepts. In this paper, we aim to reduce the number of concepts generated by a method called Generalized One-Sided Concept Lattice (GOSCL), which main advantage, compared to other FCA methods, is its capability to work with different types of attributes used in input data table. We propose reduction method, which is able to work with such heterogeneous input data table and is based on Modified Hamming Distance.