形式背景的BAM神经网络模型及模型上的概念生成

BAM Models of Formal Contexts and Generating Concepts on the Models

形式概念分析是近年来发展较为迅速的一种数据挖掘工具,它已被广泛地应用于机器学习、软件配置、信息获取等领域,而神经网络是基于模拟人脑的智能特点而发展起来的一门新兴学科。它们之间的融合将有利于智能控制、模式识别、知识处理等学科的进一步发展。通过对BAM神经网络的设定,建立了形式背景和NK-BAM神经网络之间的对应关系,论证了NK-BAM模型的稳定状态与形式背景的概念格的概念结点之间的对应,为概念分析和神经网络的进一步研究奠定了理论基础。同时给出了一个基于神经网络的概念生成算法,并通过实例验证了算法的有效性。

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 are 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.