Abstract: There are much of similarity in page structures among Web applications when they have similar functions. Aiming at the current existence of high complexity and low developing efficiency in the process of Web pages’ development, a method of mining existing Web pages’ layout structures which share similar functions is proposed. The technique fully uses the features on Web pages’ layout structures and applies the stage treatment. It firstly uses the page segment algorithms combined with similarity calculation to mine out the code blocks with high degree of similarities from layout structures, then parses style files and matches the stylesheets corresponding to the set of nodes quickly by the way of RoSunday and establishes the tree of document model. The combinations of each submodule can be used to realizethe automatic Web page design. According to the application example, the method can design and generate pages dynamically, it will improve the development efficiency effectively.

Key words: page design, segment algorithms, similarity, layout mining, pattern matching

摘要： 具有相似功能的Web应用，其页面样式和布局往往存在很大的相似性。针对当前Web页面开发复杂度高且效率低的情况，提出一种挖掘现有页面布局结构和样式属性的方法来实现Web页面自动化设计。该方法充分利用Web网页布局结构上的特点，采用分级处理的方式，首先利用页面分块算法思想通过相似度计算挖掘出具有相似性的代码块，其次通过结合RoSunday方法解析样式文件快速匹配出节点集合对应的样式表并建立文档模型树结构，各个子模块之间的相互组合可以实现页面的自动化设计。通过应用实例表明，该方法能动态地设计并生成页面，有效提升Web页面开发效率。

关键词: 页面设计, 分块算法, 相似度, 布局挖掘, 样式匹配