【摘要】 本文提出了一个混合型多概念获取算法ＨＭＣＡＰ，它将符号学习ＣＡＰ算法的主要思想和ＢＰ神经网络有机结合，以状态在实例空间出现的概率为启发信息，以提供的混合实例集为分类依据，并具有增量学习能力．ＨＭＣＡＰ所依据的实例集既可具有离散属性又可有连续属性，并且能根据用户的要求得到不同精度的结合ＢＰ网的二叉多分类判定树．本文还给出ＨＭＣＡＰ的算法应用实例，ＨＭＣＡＰ可用于自动知识获取系统．更多还原

【Abstract】 Am ulti-concept acquisition algorithm HMCAP is given, which is a hybrid of symbolic and neural network approaches. HMCAP combines the main idea of CAP and BP neural networks,uses the probability of occurence of a state in the example space as its heuristic information, and with the given amalgamated exam’pie set as its foundation of classification.It also has the ability of incremental learn"ing. The example set may have both discrete and continuous attributes. HMCAP can generate binary multi-classcification trees combining BP networks, of various precision according to user’s requirements.The application case of HMCAP is also given. HMCAP can be used in automated knowledge acquisition systems.更多还原