Prediction of good abandonment behavior in mobile search

Abstract

Good abandonment behavior means that in a query, the user can obtain the required information directly through the search results without clicking any linked page or reformulating the search query, which is common in mobile search at present. In this paper, a good abandonment prediction model in mobile search was constructed from 5 groups of features: session features, query features, SERP features, mobile touch interaction features and visual attention features. The visual attention features are introduced into good abandonment prediction for the first time and proved to be able to improve the prediction accuracy.