Towards receptivity-aware computing for mobile just-in-time health intervention

Abstract

Advances in ubiquitous computing and sensor network technologies in a recent decade enable a variety of mobile-based health care. In particular, just-in-time (JIT) adaptive health intervention which delivers the right type of support at the right time has recently been in the spotlight. One challenge for designing JIT intervention is how to make users are $ \textit{receptive} $ to JIT support. This thesis proposes the receptivity-aware computing which understand, detect, and improve receptivity to JIT support with mobile technologies. For this, I first propose a user's underlying decision-making process on an incoming support and investigate a variety of factors relevant to decisions. Subsequently, I devised machine learning models that predict users' decisions using a variety of mobile data. Furthermore, to encourage users to adhere to JIT support, I design a stochastic micro-incentive, in which users' micro-behavior changes are compensated for micro-incentives that are dynamically adjusted depending on users' adherence on different incentives and contexts. Such receptivity-aware computing is expected to not only increase the effectiveness of mobile JIT intervention, but also contribute to the development of JIT intervention that efficiently utilize various resources such as cognition, cost, and time.