Google和Yahoo的阿拉伯语搜索效果对比分析

发布时间：2018-05-06 11:27

本文选题：Google + Yahoo　；参考：《北京外国语大学》2013年硕士论文

【摘要】：搜索引擎评价是推动搜索引擎发展的关键性技术。评价搜索引擎有多种方法。本文在搜索引擎评测法Cranfield范式评价法的框架上,根据搜索引擎相关性六级评分和DCG方法,对目前阿拉伯国家搜索引擎市场占有率较大的两个搜索引擎Google和Yahoo的阿拉伯语搜索效果进行评测。搜索引擎评价的相关研究多是以满足普通搜索用户的信息需求作为目的的,但由于搜索用户对搜索引擎的满意程度很难被定量地表示,所以目前全世界对搜索引擎的评价通常从搜索引擎的结果质量(Effectiveness)、搜索引擎的效率(Efficiency)和界面(Interface)这三个方面来进行,以此来尽可能将用户对搜索引擎的满意度来量化,从而发现搜索引擎系统中存在的主要问题和缺陷以便解决它们来优化搜索引擎的性能。鉴于每个方面的评测方法、评测过程都非一篇硕士论文所能驾驭,因此本文以搜索引擎的结果质量这一较为基础且核心的点为切入点,对Google和Yahoo的阿拉伯语搜索效果进行评测。对一个互联网产品的性能评测不能割裂地来写,因此本文以Google和Yahoo在阿拉伯国家搜索引擎市场所取得的成就起始并对阿拉伯国家目前本土主要搜索引擎概况做一简述。通过第一章,可以了解到Google和Yahoo搜索引擎在阿拉伯国家受众之广、产品之多。那么,两大搜索引擎在阿拉伯国家所占市场份额是否与其搜索引擎的搜索效果相匹配,为了了解这一点,更好地了解这两大搜索引擎的搜索结果质量究竟如何,第二章和第三章为具体评测部分。由于评测方法比较复杂,所以需要对搜索引擎工作的一些基本原理、本文使用的搜索引擎的评测方法和原理、以及文中所涉及到的其他相关理论知识进行阐述,本文第二章即为相关理论及方法的阐释。第三章为根据第二章所述评测方法对Google和Yahoo阿拉伯语搜索结果质量维度上进行的具体评测和数据分析。通过评测可以发现,Google的阿拉伯语搜索效果比Yahoo好。但是,两者在结果质量相关性上的差距并没有我们之前想象中的那么大。甚至在为数不少的结果中,Yahoo的表现要比Google好。Google和Yahoo在阿拉伯语搜索结果质量上都有可以改进之处。Yahoo主要在低质页面的识别、网站权威度、时效性、页面内容丰富度、无效资源识别上不如Google,而Google虽然在这几方面表现得比Yahoo好,但也不能杜绝这几种情况。Google的主要问题在于过度将部分网站的排名排在前三位结果中,一定程度上影响了用户满意度。通过对Google和Yahoo阿拉伯语搜索结果质量的评测,可以看到Google和Yahoo的阿拉伯语搜索引擎技术还有较大的优化和发展空间。因此,本文最后一章承接前一章这一结论,阐述了阿拉伯语搜索引擎市场的巨大潜力,并对评测过程中发现的阿拉伯网站的特点做一简单总结,希望能够对经常浏览阿拉伯语网页、获取阿拉伯语信息的人士有所助益。
[Abstract]:Search engine evaluation is the key technology to promote the development of search engine. There are many ways to evaluate search engines. In this paper, based on the framework of search engine evaluation method Cranfield normal form evaluation method, according to search engine relevance six grades and DCG method, This paper evaluates the Arabic search results of Google and Yahoo, two search engines with large market share in Arab countries. Most of the research on search engine evaluation is aimed at satisfying the information needs of ordinary search users, but the degree of satisfaction of search engine users is difficult to express quantitatively. So at present, the evaluation of search engines around the world is usually carried out from three aspects: the quality of search engines' results, efficiency of search engines and interfaces, so as to quantify the users' satisfaction with search engines as much as possible. The main problems and defects in search engine system are found in order to solve them to optimize the performance of search engine. In view of the evaluation method of each aspect, the evaluation process is beyond the control of a master's thesis, so this paper evaluates the Arabic search effect of Google and Yahoo based on the basic and core point of search engine result quality. This paper begins with the achievements of Google and Yahoo in the search engine market of Arab countries and gives a brief overview of the main search engines in the Arab countries. Through the first chapter, we can learn that Google and Yahoo search engines have a wide audience and many products in Arab countries. Well, in order to understand whether the market share of the two major search engines in the Arab countries matches the search results of their search engines, in order to better understand the quality of the search results of the two major search engines, The second and third chapters are the specific evaluation part. Because of the complexity of evaluation methods, it is necessary to expound some basic principles of search engine work, the evaluation methods and principles of search engine used in this paper, and other relevant theoretical knowledge involved in this paper. The second chapter is the explanation of relevant theories and methods. The third chapter is the concrete evaluation and data analysis of the quality dimension of Google and Yahoo Arabic search results according to the evaluation methods described in Chapter 2. A review shows that Google's Arabic search results are better than Yahoo's. However, the difference in the quality of the results is not as large as we thought. Even in a number of results, Yahoo performs better than Google. Google and Yahoo can improve the quality of Arabic search results. Yahoo is mainly in the identification of low-quality pages, site authority, timeliness, page content richness. Invalid resource identification is not as good as Google's, and Google does better than Yahoo in several areas, but the main problem with Google is that it over-ranks some sites in the top three results. To some extent, the impact of user satisfaction. By evaluating the quality of Google and Yahoo Arabic search results, we can see that the Arabic search engine technology of Google and Yahoo still has great space for optimization and development. Therefore, the last chapter of this paper follows the conclusion of the previous chapter, expounds the great potential of the Arabic search engine market, and makes a brief summary of the characteristics of the Arab websites found in the course of evaluation. Hope to be able to often visit Arabic pages, access to Arabic information to help.
【学位授予单位】：北京外国语大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：G354;H37

【参考文献】