基于元搜索引擎的垂直搜索子系统的设计

发布时间：2018-04-24 04:07

本文选题：隐形关键词 + 垂直搜索　；参考：《天津大学》2012年硕士论文

【摘要】：垂直搜索引擎是搜索引擎发展的新阶段。对于搜索引擎的未来发展和具体研究而言，这是一个必然的趋势。当前垂直搜索引擎的系统结构和传统上的全文搜索引擎非常相似，能够较高水平处理专业相关度，不过作为垂直搜索引擎，在某些问题方面和传统上的全文搜索引擎相同，比如较低水平的查全率、消耗的网络资源太多等等。对于存在的这种问题，本文的解决方案为：垂直搜索引擎之具体的系统结构要建立在元搜索之上。借助于这种技术，能够很好的提升查全率，不过对应的专业相关度也呈现较为明显的下降趋势。实验结果告诉我们，这种新的系统功能比较强大，对于垂直搜索引擎期望自己可以达到的相关功能，新系统都可以实现。本研究的主要内容包括了如下几个方面： 1.目前的垂直搜索引擎查全率较低，由于元搜索引擎有较高的查全率，我们设计了一种垂直搜索引擎，采用的方法收集的信息，这是根据meta-search引擎。该系统增加了信息的收集和分析适应需求的垂直搜索引擎。 2.对于搜索引擎来讲，其最基本的基础功能就是进行信息收集。当前的垂直搜索引擎在这个功能角度上存在的主要问题为：较低水平的网络信息覆盖率，，收集到的多为无效的信息等等。据此，本研究提出的解决方法为建立在对用户的具体浏览时间进行统计的基础之上进行信息收集，这种信息收集方法也是建立在元搜索引擎技术之上的，借助于这种信息收集方法，也能够收集到用户给予了较高关注度的信息。这种技术，不但能够提升了信息覆盖率，而且对于被收集的相关信息，也能够提升专业相关度。 3.对于搜索引擎来讲，其核心在于信息检索。在分析收集到的信息的时候，通过将数据挖掘引入其中，本文获得了关键词和较高满意度查询结果之间的具体规则。借助于此，本文提出了一个新的概念，即隐形关键词。经过实验，可以得知隐形关键词的使用，一方面很好的提升了在专业方面，系统查询结果的相关度。 4.对于搜索结果，用户关心最多的是之前的结果，故而作为搜索引擎，一个必须关注、也是必须重视的问题就是对结果如何进行排序。当前，元搜索引擎在进行结果排序的时候，使用的相关信息非常少，也不能对结果相关度给与很好的保证。基于此，本文也进行了改进，本文提出的结果排序方法与系统相契合，同时因为搜索之中将隐性关键词引入，所以对于位置排序算法也进行了很好的改进，同时专业相关度的搜索结果也更为准确。总体来讲，本文的问题解决方案为：建立在元搜索技术之上的垂直搜索引擎，在某种程度上优化了垂直搜索引擎，笔者在本文运用了一种新思路和新方法进行探讨。
[Abstract]:Vertical search engine is a new stage of search engine development. For the future development of search engines and specific research, this is an inevitable trend. The current system structure of vertical search engine is very similar to that of traditional full-text search engine, and it can deal with professional relevance at a high level, but as a vertical search engine, it is the same as traditional full-text search engine in some aspects. For example, the low level of recall, consuming too much network resources and so on. For this problem, the solution of this paper is: the specific system structure of vertical search engine should be based on meta search. With the help of this technique, the recall rate can be improved very well, but the relative professional correlation also shows an obvious downward trend. The experimental results show that the new system is quite powerful and can be implemented for the related functions that the vertical search engine expects it to achieve. The main contents of this study include the following: 1. At present, the vertical search engine has a low recall rate. Because the meta search engine has a high recall rate, we design a vertical search engine, which uses the method to collect information, which is based on the meta-search engine. The system adds information collection and analysis to meet the needs of the vertical search engine. 2. For search engines, its basic function is to collect information. The main problems of the current vertical search engine in this functional angle are: low level of network information coverage, collected mostly invalid information and so on. Therefore, the solution proposed in this study is to collect information on the basis of statistics on the specific browsing time of users, and this information collection method is also based on meta-search engine technology. With the help of this information collection method, users can also collect information with a high degree of attention. This technology can not only improve the information coverage, but also enhance the relevance of the collected information. 3. For search engine, its core is information retrieval. In the analysis of the collected information, by introducing data mining into it, this paper obtains the specific rules between keywords and higher satisfaction query results. With the help of this, this paper proposes a new concept, namely the stealth keyword. Through experiments, we can know the use of hidden keywords, on the one hand, improve the relevance of the system query results in the professional. 4. For search results, users are most concerned about the results before, so as a search engine, a must pay attention to, and must pay attention to is how to sort the results. At present, the meta-search engine uses very little information when sorting the results, and it can not guarantee the relevance of the results. Based on this, this paper also improved, the result sort method proposed in this paper coincides with the system, at the same time, because the hidden key words are introduced in the search, the location sorting algorithm is also improved very well. At the same time professional relevance of the search results are more accurate. In general, the solution of this paper is: the vertical search engine based on meta-search technology, to some extent, optimize the vertical search engine, the author uses a new way of thinking and new method to discuss in this paper.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP391.3

【参考文献】