搜索竞价广告中关键字最优化策略  

Keyword Optimization in Search Auctions

作　　者：杨彦武[1] 郭迅华[2] 曾大军[1] 张杰[1] 秦蕊[1] 

机构地区：[1]中国科学院自动化研究所,北京100190 [2]清华大学经济管理学院,北京100084

出　　处：《信息系统学报》2013年第2期9-23,共15页China Journal of Information Systems

基　　金：国家自然科学基金(71272236,71071152,71110107027,71025001);教育部人文社会科学重点研究基地项目基金(12JJD630001);北京市自然科学基金(4132072)

摘　　要：关键字是搜索竞价广告中联结广告主、搜索用户和搜索引擎之间的重要纽带。关键字的生成、选择与分组等过程一直都是搜索竞价广告主所面临的核心决策问题，并贯穿于搜索竞价推广活动的整个生命周期。本文提出了一个针对搜索竞价推广活动整个生命周期的多层次、闭环关键字策略优化框架模型，分别从关键字的生成、扩展、定位、选择与组合以及动态调整几个方面研究相应的最优化策略。进而，利用实际关键字竟价推广活动的日志数据，设计了相应的搜索竞价关键字策略实验场景，对本文提出的关键字优化模型及其算法进行实际验证与评估。实验结果表明，本文所提出的关键字策略优化框架模型及其算法能够收敛且在很大程度上优于一般广告主所采取的关键字策略。In sponsored search auctions, keyword is a kernel bridge among advertisers, search users and search engines. Advertisers have to deal with a series of primary keyword-related decisions （e. g. , keyword generation, selection and portfolio） throughout the entire lifeeycle of search advertising campaigns. This paper proposes a hierarchical and close- loop keyword optimization framework to support various keyword-related decisions during the entire lifecycle of search advertising campaigns. Based on this framework, we develop various optimal strategies for keyword generation, expansion, allocation, selection and portfolio with consideration of search advertising structures, and dynamical adjustment of keywords aeeording to the realtime advertising performance. Furthermore, with field reports and logs from historical and ongoing search advertising campaigns, we conduct some experiments to validate our keyword optimization framework and identified strategies. Experimental results show that our keyword optimization framework and strategies outperform the baseline keyword strategy applied in common practices.

关 键 词：电子商务 搜索竟价广告 策略优化 关键字策略 广告策略 

分 类 号：C934[经济管理—管理学] C931.6