Predicting Users' Gender and Age based on Mobile Tasks

Abstract

Demographic attributes are a key factor in marketing products and services, which enable a business owner to find the ideal customer. Users' app usage behaviors could reveal rich clues regarding their personal attributes since they always determine what apps to use depending on their personal needs and interests. Prior studies [1, 2] have tried to predict users' gender and age through their app usage behavior. However, most of the existing methods for users' demographic prediction are straightforward, simply using popular used apps or app usage frequency as features, without considering the internal semantic relationship of apps usage. Recently, mobile tasks [3] have been identified from mobile app usage logs, representing a more accurate unit for capturing users' goals and behavioral insights, where a "mobile task" can be thought of as a group of related used apps to accomplish a single discrete task. For example, to plan dinner with friends, multiple apps (e.g., WhatsApp, Yelp, Uber and Google Maps) might be accessed for completing the task. In this talk, I will introduce how we leverage the fine-grained task units for generating user representation aims at predicting users' gender and age. We analyzed the effectiveness of using tasks to infer users' demographics, especially when compared to only treating apps independently. We explored different approaches for constructing users' representation and models with both mobile apps and tasks. Finally, we validated that the two-level hierarchical structure of "apps to tasks" and "tasks to users" is an important factor that should be taken into consideration for improving mobile user modelling. This work shed light on whether and how the extracted mobile tasks could be effectively applied. We believe that the task-based representations could be further explored for improving many other applications.