Revisit Prediction by Deep Survival Analysis (PAKDD 2020, To appear)

Abstract

In this manuscript, we introduce SurvRev, a next-generation revisit prediction model that can be tested directly in the business. The SurvRev model has many advantages. First, SurvRev can use partial observations which were considered as missing data and removed in the previous regression framework. By using deep survival analysis, we are able to estimate the next customer arrival from unknown distribution. Second, SurvRev is an event rate prediction model. It generates the predicted event rate of the next k days rather than predicting revisit interval and revisit intention directly. We showed the superiority of the SurvRev model by comparing with diverse baselines including the feature engineering model and the state-of-the-art deep survival models.

Publication
In the 24th Pacific-Asia Conference on Knowledge Discovery and Data Mining