提出了一种基于LCS的特征树最大相似性匹配网页去噪算法。通过将目标网页和相似网页转化为特征树,并将特征树映射为一个特征节点序列,利用LCS算法能获得最长子序列全局最优解的特点,找出两棵特征树之间的不同节点作为候选集,并对候选集进行聚集评分找出网页重要内容块。给出了算法的原型系统,并对每一个模块的实现做了详尽的描述。 A LCS-based matching algorithm for the maximum similarity of feature pages is proposed. By transforming the target web pages and similar web pages into feature trees and mapping the feature tree into a feature node sequence, the LCS algorithm can obtain the global optimal solution of the longest sub-sequence and find out the different nodes between the two feature trees as Candidate set, and the collection of the candidate score to find the page important piece of content. A prototype system of the algorithm is given, and a detailed description of the implementation of each module is given.