Preliminary Investigation of Predicting Next-Use Mobile Apps Using App Semantic Representations

Cheng Chen, Takuya Maekawa, Daichi Amagata, Takahiro Hara

Abstract: This study presents our preliminary investigation of predicting the next-use mobile Apps based on the App usage history of a target user that can facilitate the user to select an App from an entire list of Apps. The proposed method is designed to train a next-use App prediction model using the usage history of other users (source users) to cope with the cold-start problem of a system in which the training data from a target user is considered to be insufficient when the user begins to use the prediction system. We predict the usage of cold-start users and Apps by leveraging the semantic similarities between the Apps that are installed on the smartphones of the source users and the Apps that are installed on the smartphones of the target user.