Keyword selection on Google Ads

Abstract

Google Advertising is a publicity agency that provides marketers with advertisements. By choosing keywords relevant to their ad material, advertisers fit the user's search terms and push advertising. Keywords will decide the type of users being pushed by an advertiser, the efficacy of the ad promotion, and therefore the ad product's sales. The main objective is to automatically choosing keywords that are satisfactory to advertisers from an outsized number of keywords given by Google Advertising. But there’s not an excessive amount of time for the framework to make a decision whether keywords are chosen and to pick the proper keywords within the shortest time. Therefore, a model structure which can obtain some helpful keywords for advertisers is built also to accomplish this multipurpose task, an enhanced method of multi-objective particle swarm optimization is introduced. Many technical challenges need to be solved to accomplish this multi-objective mission, such as the issue of mixed language, the problem of data imbalance, the issue of obtaining features from the collection, and so on. The mixture of evolutionary computation, deep learning, machine learning and text processing approaches is used here to solve the issue of keyword selection.