SmokingOpp: Detecting the Smoking ‘Opportunity’ Context Using Mobile Sensors

on YouTube ↗06m

" summary

Defines smoking opportunity context as presence of cigarette availability plus smoking allowance and shows those items Granger-cause future smoking. Uses GPS and accelerometer data to build event-based assessment windows, mines personal and general smoking spots, and adds indoor/outdoor, time-of-day, and time-since-last-cigarette-POS features. Trains models on 15 million GPS points, 13,000 hours of activity and 3,412 self-reports from 90 newly abstinent smokers, validated with leave-one-subject-out CV; logistic regression achieves recall 79.8% and F1 74.3%.

AI-written summary · about

" chapters (11)

1. 0:00Start
2. 0:21Introduction
3. 0:46Predictors of Impulsive Adverse Behaviors
4. 1:33Defining Opportunity' Contexts
5. 2:28Sensors: GPS and Accelerometer
6. 2:40Data Segmentation
7. 3:14Feature Computation
8. 3:51Model Training: Dataset
9. 4:21Model Training: Label Assignment
10. 4:43SmokingOpp Model
11. 5:03Limitations and Future Works