CAJ | 학술논문

主题爬虫是收集特定领域资源的网络爬虫。为了保证主题爬虫的查准率，提出一种基于模糊 SVDD（support vector do-main description）监督的 PageRank 爬虫算法，既考虑网页间的链接关系，又使用合适的分类器监督来保证爬虫与主题不偏离。通过与关键词匹配主题爬虫、shark-search 主题爬虫、PageRank 主题爬虫、基于 SVM预测的主题爬虫、普通 SVDD 指导的主题爬虫进行试验对比，验证了该算法具有更高的查准率。
주제파충시수집특정영역자원적망락파충。위료보증주제파충적사준솔，제출일충기우모호 SVDD（support vector do-main description）감독적 PageRank 파충산법，기고필망혈간적련접관계，우사용합괄적분류기감독래보증파충여주제불편리。통과여관건사필배주제파충、shark-search 주제파충、PageRank 주제파충、기우 SVM예측적주제파충、보통 SVDD 지도적주제파충진행시험대비，험증료해산법구유경고적사준솔。
Focused crawler is a web crawler to collect resources from specific fields.In order to ensure focused crawler's precision,the arti-cle proposes a PageRank crawler algorithm based on fuzzy SVDD(support vector domain description)supervision,which not only considers the linking relations among pages,but also uses classifier supervision to prevent crawler from departing from focus.Compared by experiments with keyword matching focused crawler,shark-search focused crawler,PageRank focused crawler,SVMprediction based focused crawler and ordinary SVDD instructed focused crawler,it is validated that the proposed algorithm is more precise.