Psychological Stress Detection from Online Shopping
The increasingly faster life pace in modern society makes people always feel stressful and it is of great significance to discover a users suffering stress in time.According to psychological study,shopping is chosen as an effective way for stress relief,especially for females.Compared with non-stress cases,a user may perform different shopping patterns when under stress.An interesting issue then arises: can we detect ones psychological stress from online shopping data? By investigating stress-related outlier features from both content and behavior of online purchase orders,we learn a users stress status by classification.A real user study of 20 experienced female online customers aged 23-30 verifies the effectiveness of shopping based stress detection,achieving an F1-measure of more than 80% with J48 classifier.None of the features negatively affect the detection result.Feature combinations bring dramatic improvements than single feature.In total,shopping content features are proved to be more significant than behavior features.
Liang Zhao Hao Wang Yuanyuan Xue Qi Li Ling Feng
Tsinghua National Laboratory for Information Science and Technology,Department of Computer Science a School of Computer Science and Technology,North China University of Technology,Beijing 100144,China
国际会议
International Asia-Pacific Web Conference(第18届国际亚太互联网大会)
苏州
英文
431-443
2016-09-23（万方平台首次上网日期，不代表论文的发表时间）