摘要： 在关键词拍卖中,每个广告主都有成千上万的关键词可选,为了在预算限制条件下给选择的关键词设置合理的投标价格,提出一种新的基于广告主的拍卖策略,包括选词策略和出价策略。在选词策略中,提出基于词频-反转文件频率算法的关键词关联度计算方法,通过该方法选出的关键词不仅能提高网站的关联度,增加转化率,还能避免因使用过度普遍的关键词而增加竞争成本。在出价策略中,运用改进的粒子群优化算法,在若干约束条件限制下对每个关键词的出价做适当调整,以增加广告主所获利润。实验结果表明,采用拍卖策略选出的关键词组可增加网站的转化率,降低竞争成本,所获得的利润比传统人为投标所获得的利润高,并且在初期和中期呈现持续上升趋势,后期趋于稳定。

关键词: 关键词拍卖, 选词策略, 关联度, 出价策略, 词频-反转文件频率算法, 粒子群优化

Abstract: There are millions of available keywords for each advertiser in keywords auction.How to set a reasonable bid price for the selected keywords under a limited condition,such as budget,becomes the most difficult work for the advertiser.As it is hard for many advertisers to select keywords and set the price,a novel model of auction strategy based on advertisers is proposed for these problems.This auction strategy includes the keywords selection strategy and the bidding strategy.The keywords selection strategy presents a calculation method for keywords correlation which is based on the Term Frequency-Inverse Document Frequency(TFIDF)algorithm.The keywords,selected through this method,not only improve the correlation with the promoted website,increase the conversion rate,but also avoid increasing the competition cost due to the overuse of common keywords.The bidding strategy uses an improved Particle Swarm Optimization(PSO)algorithm to properly adjust the bids of each keyword under some constraint conditions so as to increase the profits of advertisers.Experimental results show that keywords,selected through the auction strategy,increase the conversion rate of website and reduce the competition cost.Moreover,its profit is higher than that of traditional bidding method.The algorithm presents a continuous rising trend in the early-middle period and becomes stable in the late period.

Key words: keyword auction, word selection strategy, correlation degree, bidding strategy, Term Frequency-Inverse Document Frequency(TFIDF)algorithm, Particle Swarm Optimization(PSO)

中图分类号: 

• TP391