最大和搜索结果多样性问题及其贪婪算法分析
发布时间：2019-02-13 06:33
【摘要】：研究互联网搜索结果的最优多样性问题.给定用户搜索关键词较少,以及关键词本身的多义性,同时由于搜索系统一次呈现结果存在数量上的限制,系统常不能准确定位用户的真实搜索需求.为了最大化覆盖用户的搜索需求,搜索系统显示的结果不仅需要最大化同关键词的相关性,而且需要最大化结果之间的差异性.考虑了最大和搜索结果多样性问题,给出了贪婪算法,并针对实际中的差异性度量常常不满足三角不等式的情况下,分析证明了该贪婪算法具有的近似性能比.结果表明贪婪算法具有很好的理论近似性能.
[Abstract]:The optimal diversity of Internet search results is studied. Given user search terms are less, and the ambiguity of the keywords themselves. At the same time, due to the number of the results presented in the search system, the system is often unable to accurately locate the real search requirements of users. In order to maximize the search requirements of the users, the results displayed by the search system need not only maximize the correlation with the keywords, but also maximize the differences between the results. In this paper, the problem of maximum and diversity of search results is considered, and the greedy algorithm is given. The approximate performance ratio of the greedy algorithm is proved by analyzing the fact that the difference measurement often does not satisfy the triangular inequality. The results show that the greedy algorithm has good theoretical approximation performance.
【作者单位】： 电子科技大学经济与管理学院;
【基金】：国家自然科学基金(70901012) 中央高校基本科研业务费专项资金(ZYGX2013J134)~~
【分类号】：TP391.3
[Abstract]:The optimal diversity of Internet search results is studied. Given user search terms are less, and the ambiguity of the keywords themselves. At the same time, due to the number of the results presented in the search system, the system is often unable to accurately locate the real search requirements of users. In order to maximize the search requirements of the users, the results displayed by the search system need not only maximize the correlation with the keywords, but also maximize the differences between the results. In this paper, the problem of maximum and diversity of search results is considered, and the greedy algorithm is given. The approximate performance ratio of the greedy algorithm is proved by analyzing the fact that the difference measurement often does not satisfy the triangular inequality. The results show that the greedy algorithm has good theoretical approximation performance.
【作者单位】： 电子科技大学经济与管理学院;
【基金】：国家自然科学基金(70901012) 中央高校基本科研业务费专项资金(ZYGX2013J134)~~
【分类号】：TP391.3
【相似文献】
相关期刊论文 前10条
1 陈洪;蔡佳;张玉成;;多分类贪婪算法的一致性[J];湖北大学学报(自然科学版);2005年04期
2 杨洁;;基于贪婪算法的卫星区域观测摆角方案选择方法[J];广西科学院学报;2006年02期
3 申时凯;吴绍兵;申浩如;王付艳;管彦庆;;计算最短公共超串的贪婪算法[J];计算机工程与设计;2007年08期
4 王杰;刚轶金;李凤光;吴伟巍;;改进贪婪算法在博客突发事件检测中的研究[J];计算机工程与应用;2008年34期
5 周柳阳;高珩;梁翥;;贪婪算法的实际应用[J];硅谷;2009年02期
6 李e，
本文编号：2421281
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2421281.html
