Abstract: As a tool for data mining,formal concept analysis(FCA) has get rapidly developed and been widely used in machine learning, software engineering and information retrieving. Artificial neutral network(ANN) is a new field of artificial intelligence,whose evolvement is founded on simulating intelligent characteristics of human brain. In fact, the synthesis between FCA and ANN will be greatly advantageous to the further development of such domains as intelligent control, pattern recognition, knowledge processing etc. Based on the enactment toward a BAM neutral network, the paper presented the corresponding relation between formal contexts and Nk-BAM neural network models. Besides, we proved that the stable states of the Nk-BAM neutral network arc corresponding to the concept nodes of concept lattice built on the formal context,which establishes the theoretical foundation for the fusion of FCA and ANN. At the mean time, an algorithm for generating concepts on the models was brought forward and an illustrative example guarantees the availability of the algorithm.

Key words: Formal concept analysis, Formal context, Concept lattice, Neural network, Bidirectional associative memories