Keyword targeting optimization in sponsored search advertising: Combining selection and matching

In sponsored search advertising (SSA), advertisers need to select keywords and determine matching types for selected simultaneously, i.e., keyword targeting. An optimal targeting strategy guarantees reaching the right population effectively. This paper aims address problem, which is a challenging task because of incomplete information historical performance indices high uncertainty in SSA environments. First, we construct data distribution estimation model apply Markov Chain Monte Carlo method make inference about unobserved (i.e., impression click-through rate) over three broad, phrase exact). Second, formulate stochastic (BB-KSM) combining operations selection maximize expected profit under chance constraint budget, develop branch-and-bound algorithm incorporating simulation process our model. Finally, based on realworld dataset collected from field reports logs past campaigns, computational experiments are conducted evaluate strategy. Experimental results show that, (a) BB-KSM outperforms seven baselines terms profit; (b) shows its superiority as budget increases, especially situations with more combinations; (c) proposed approach can effectively problem turn significantly promotes decisions. research makes important contributions literature offer critical insights into management advertisers.