Combining Page Group Structure and Content for Roughly Filtering Researchers' Homepages with High Recall (特集:情報融合) Combining Page Group Structure and Content for Roughly Filtering Researchers' HomePages with High Recall(Special Issue on Fusions of Information Contents and Information Technologies)

This paper proposes a method for gathering researchers' homepages (or entry pages) by applying new simple and effective page group models for exploiting the mutual relations between the structure and content of a page group, aiming at narrowing down the candidates with a very high recall. First, 12 property-based keyword lists that correspond to researchers' common properties are created and are assigned either organization-related or other. Next, several page group models (PGMs) are introduced taking into consideration the link structure and URL hierarchy. Although the application of PGMs generally causes a lot of noises, modified PGMs with two original techniques are introduced to reduce these noises. Then based on the PGMs, the keywords are propagated to a potential entry page from its surrounding pages, composing a virtual entry page. Finally, the virtual entry pages that score at least a threshold number are selected. The effectiveness of the method is shown by comparing it to a single-page-based method through experiments using a 100GB web data set and a manually created sample data set.