A Layered Recursive Construction Algorithm and a visualization method for concept lattice

As an efficient tool for data analysis, Formal Concept Analysis (FCA) is widely used in domains of software engineering and machine learning. The construction of concept lattice is the core step of FCA. A Layered Recursive Construction Algorithm (LRCA) is proposed in this paper. It creates direct child concepts for every concept and adopts the method of logical operation of attribute vectors to guarantee only direct child concepts can be produced. And it uses layered examining method to guarantee creating child concepts for each parent only once. A simulation experiment is done for performance evaluation. The experimental result shows that LRCA can greatly raise the processing ability by more than 6 times. In addition, a heuristic method is given for concept lattice visualization. It can automatically draw distinct and neat concept lattice diagram, and make the crosses of edges as few as possible.