基于解释学习的包装器生成

Wrapper Generation Based on Explanation Learning

开发Web信息抽取系统的核心是为各个Web信息源构造包装器,而构造包装器的关键在于规则学习器。鉴于传统的规则学习器一般都基于单一的学习策略,结合归纳学习和分析学习的优点,提出了基于解释学习的规则学习器,以此为核心生成包装器,并将其应用到了实际的包装器生成系统中去。

The key componem in the web information extraction system is the wrapper constructed for each web data source, and the kernel of a wrapper is a rule learner. In view of the traditional wrapper based on single - strategy in a general way, combining the advantage of induction learning and analysis learning, this paper brings forward an explanation- based learning rule learner and wrapper generation method .Furthermore, it applies the method to the instance of wrapper generation system.