One More Bite? Inferring Food Consumption Level of College Students Using Smartphone Sensing and ...
summary
Study collects data from 84 college students with 3,278 self-reports and one-hour windows of passive sensing (accelerometer, app and screen usage) plus time of eating. Food categories (fats, meats, sugary drinks, restaurant dishes), social context and mood correlate with overeating vs undereating. A random forest with leave-k-users-out achieves 87.81% accuracy for three classes using sensing plus self-report and 83.49% accuracy using only sensing and eating time, suggesting viable context-aware food diaries.
AI-written summary · about
related talks
06mNeckSense: A Multi-Sensor Necklace for Detecting Eating Activities in Free-Living Conditions
06mPredicting Subjective Measures of Social Anxiety from Sparsely Collected Mobile Sensor Data
09mSocial Media Image Analysis for Public Health
27mHARSHA URLAM & SWETA SINGH at Cypher 2017
10mMy AI Thinks I'm Eating My Feelings (and Other Nutritional Insights)
44mQuantified Self: Analyzing the Big Data of our Daily Life
07mPredicting Brain Functional Connectivity Using Mobile Sensing
21m“Trust Us”: Mobile Phone Use Patterns Can Predict Individual Trust Propensity