“Trust Us”: Mobile Phone Use Patterns Can Predict Individual Trust Propensity
summary
A 10-week field study collects call, SMS and GPS logs from 50 Android users and extracts 24 phone-derived features across social, spatial and temporal categories to predict a standard trust-propensity score (median 73). A regression model using phone data reaches correlation 0.53 versus 0.27 for demographics and mean absolute error ±7.7; classification improves roughly from 68% to 81% accuracy. An app is implemented; privacy limits raw content access.
AI-written summary · about
chapters (9)
- 0:00Introduction to presentation
- 2:34Problem with trust measurement
- 6:51Mobile phone metadata approach
- 9:26Predicting trust propensity
- 13:38Classification results analysis
- 14:43Privacy and ethical considerations
- 15:47Study limitations and future work
- 16:53Conclusion and acknowledgments
- 17:19Q&A session
related talks
22mA Classification Model for Sensing Human Trust in Machines Using EEG and GSR
06mFacial Recognition: Understanding Privacy Concerns and Attitudes Across Increasingly
06mMachine Learning for Phone-Based Relationship Estimation: The Need to Consider Population ...
26m"Reality Mining for Compan
23mKeynote: Digital Psychometrics and it's Future Effects on Technology
06mExamining the Social Context of Alcohol Drinking in Young Adults with Smartphone Sensing
05mComparing Security and Privacy Attitudes Among U.S. Users of Different Smartphone and
25mThe State of Mobile App Privacy: Compliance Trends & Takeaways