首页 | 本学科首页   官方微博 | 高级检索  
     

实时垂直搜索引擎对象缓存优化策略
引用本文:周佳庆,吴羽,江锦华,陈刚,董轶. 实时垂直搜索引擎对象缓存优化策略[J]. 浙江大学学报(工学版), 2011, 45(1): 14-19. DOI: 10.3785/j.issn.1008-973X.2011.01.003
作者姓名:周佳庆  吴羽  江锦华  陈刚  董轶
作者单位:1.浙江大学 计算机科学与技术学院, 浙江 杭州 310027; 2.工商银行浙江省分行, 浙江 杭州 310009
基金项目:国家自然科学基金资助项目(60603044, 60803003);浙江省科技计划重大科技攻关项目(2006c11108).
摘    要:针对实时垂直搜索引擎搜索对象热门度多变和数据抓取由查询驱动等问题,提出一种全新的实时垂直搜索引擎对象缓存优化策略.基于对象及属性间的关联设计热门对象预测模型,预测热门对象的变化趋势;基于用户查询及对象变化符合泊松过程的特点,推导最大化数据新鲜度的计算方法,从理论上给出资源分配和动态平衡的最优策略.大量的对比实验验证了新的缓存优化策略在较少开销增长的前提下,用户查询结果平均新鲜度和准确率均明显优于传统固定频率的缓存策略.

关 键 词:缓存策略  实时搜索  垂直搜索  搜索引擎

Object cache optimization strategy for real-time vertical search engine
ZHOU Jia-qing,WU Yu,JIANG Jin-hua,CHEN Gang,DONG Yi. Object cache optimization strategy for real-time vertical search engine[J]. Journal of Zhejiang University(Engineering Science), 2011, 45(1): 14-19. DOI: 10.3785/j.issn.1008-973X.2011.01.003
Authors:ZHOU Jia-qing  WU Yu  JIANG Jin-hua  CHEN Gang  DONG Yi
Affiliation:1.College of Computer Science and Technology, Zhejiang University, Hangzhou 310027, China;2. Zhejiang Branch of Industrial and Commercial Bank of China, Hangzhou 310009, China
Abstract:A new vertical search engine object cache optimization strategy was proposed to address the challenges like the changeful of popular objects, the property of query triggered data crawl and so on. A popular object prediction model was proposed based on relationships between objects and their properties in order to predict the tendency of popular object distribution. Since user query and data changed by  Poisson process, a procedure to maximize the data freshness and an optimal strategy to distribute and balance resource were proposed. Experimental results show that  the increase in time complexity is relative limited, while the average freshness of user query result and the query precision ratio preceded traditional fixed-rate cache strategy.
Keywords:cache strategy  real-time search  vertical search  search engine
点击此处可从《浙江大学学报(工学版)》浏览原始摘要信息
点击此处可从《浙江大学学报(工学版)》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号