Examining the Social Context of Alcohol Drinking in Young Adults with Smartphone Sensing
summary
Dataset covers 241 participants over three months with about 1250 self-reports of weekend drinking and eight passive phone modalities. Seven social-context inference tasks (two- and three-class) use leave-one-user-out training and SMOTE; overall accuracies run 75–86% with Random Forests dominating. Accelerometer features rank highest across tasks while location, Bluetooth and app-usage also contribute. A three-class sex-composition inference reaches about Random Forest ~75.9% for sex composition, suggesting feasible but imperfect group-level social context sensing.
AI-written summary · about
chapters (4)
- 0:00Start
- 1:51What is Social Context of Drinking Alcohol? We borrow the conception used in alcohol research, which refers to either one or both of the following Type of Relationship family members, spouse or partner, Friends or colleagues, others Number of People
- 2:15Research Questions RQ1: What social contexts around drinking events can be observed by analyzing self- reports and smartphone sensing data? RQ2: Can young adults' social context of drinking be inferred using sensing data?
- 5:32Conclusion We emphasized the importance of using smartphones to understand the social context of drinking
related talks
06mPredicting Subjective Measures of Social Anxiety from Sparsely Collected Mobile Sensor Data
21m“Trust Us”: Mobile Phone Use Patterns Can Predict Individual Trust Propensity
07mPredicting Brain Functional Connectivity Using Mobile Sensing
06mMachine Learning for Phone-Based Relationship Estimation: The Need to Consider Population ...
09mSocial Media Image Analysis for Public Health
05mExploring Smartphone Keyboard Interactions for Experience Sampling Method driven Probe Generation
12mUnderstanding Driving Distractions: A Multimodal Analysis on Distraction Characterization
34mMining smartphone sensor data with python