【作者】 刘雪梅；

【导师】 李献业；

【作者基本信息】 东北师范大学 ， 软件工程， 2016， 硕士

【摘要】 搜索引擎作为人类获取信息的首要工具,它的优劣直接决定了人们获取信息的准确性、速度和成本。目前针对于搜索引擎时效性方面的研究已经进行了很多年,但是存在一些问题,一是研究内容很少,二是研究方法没有详细的介绍。现如今人们对于信息的时效性要求越来越高,特别的对于突发事件,人们更加关注事件的发展状态,渴望了解事件的每一步进展,一个好的搜索引擎,对于突发事件的响应必须具有很好的策略才能满足用户的需要。本文对搜索引擎时效性进行深入研究,采用基于统计语言模型的query获取方法,提出了一种高效可行的时效性kpi的计算算法。首先,阐述了本文的理论基础,介绍了时效性评估的相关方法,分析了基于突发新闻事件时效性评估的意义及其重要性。其次,采用基于统计语言模型突发query挖掘的方法挖掘出突发事件中的突发query,获得相同query的不同搜索引擎结果页面,对结果页面进行分析,获取展现结果的位置和时间因子信息,利用时效性kpi计算方法计算出不同搜索引擎的时效性kpi指标。最后,设计并开发基于突发事件的搜索引擎时效性评估模型系统。本文所提出的时效性kpi计算方法已经在商业上得到应用,能够为公司搜索引擎功能和服务的改进提供一定的方向性建议,理论方面提出了一套简单易行、适合集群运算的突发事件时效性评估方法,可供相关研究者参考。更多还原

【Abstract】 Search engines as a primary tool for human access to information, it will directly determine people’s access to information accuracy, speed and cost. Current research directed to the search engine in terms of timeliness has been carried out for many years, but there are some problems, one is little research, two research methods not described in detail. Nowadays people for time-critical information is increasingly high, especially for emergencies, people are more concerned about the state of development of events, eager to learn every step of the progress of events, a good search engine, inclusion and for emergencies show must have a good strategy in order to meet the needs of users.In this paper, the search engine timeliness of in-depth study, using statistical language model extraction based query methods, propose an efficient and feasible timeliness kpi calculation algorithm. Firstly, the theoretical basis of this paper introduces the method of aging assessment, analyzes the significance and importance of breaking news events based on the timeliness of assessments. Secondly, the use of excavated emergencies sudden burst query query statistical language model based on mining methods, different search engine results pages get the same query, and analyze the results page to get the results to show the location and the time factor information, calculate the timeliness kpi indicators using different search engines timeliness kpi calculation. Finally, the design and development of the timeliness of the search engine evaluation model based monitoring system emergencies.Kpi timeliness of the proposed calculation method has been applied commercially, can provide some directional suggestions to improve search engine and improved services, the theory put forward a set of simple, suitable for emergency operation of the cluster timeliness monitoring method, researchers for the relevant reference.更多还原