Theory-based Habit Modeling for Enhancing Behavior Prediction

Chao Zhang,Joaquin Vanschoren,Arlette van Wissen,Daniel Lakens,Boris de Ruyter,Wijnand A. IJsselsteijn

DOI: https://doi.org/10.48550/arXiv.2101.01637

2021-01-06

Abstract:Psychological theories of habit posit that when a strong habit is formed through behavioral repetition, it can trigger behavior automatically in the same environment. Given the reciprocal relationship between habit and behavior, changing lifestyle behaviors (e.g., toothbrushing) is largely a task of breaking old habits and creating new and healthy ones. Thus, representing users' habit strengths can be very useful for behavior change support systems (BCSS), for example, to predict behavior or to decide when an intervention reaches its intended effect. However, habit strength is not directly observable and existing self-report measures are taxing for users. In this paper, built on recent computational models of habit formation, we propose a method to enable intelligent systems to compute habit strength based on observable behavior. The hypothesized advantage of using computed habit strength for behavior prediction was tested using data from two intervention studies, where we trained participants to brush their teeth twice a day for three weeks and monitored their behaviors using accelerometers. Through hierarchical cross-validation, we found that for the task of predicting future brushing behavior, computed habit strength clearly outperformed self-reported habit strength (in both studies) and was also superior to models based on past behavior frequency (in the larger second study). Our findings provide initial support for our theory-based approach of modeling user habits and encourages the use of habit computation to deliver personalized and adaptive interventions.

Artificial Intelligence,Human-Computer Interaction,Machine Learning