TY - JOUR
T1 - Discovering intimate partner violence from web search history
AU - Zaman, Anis
AU - Kautz, Henry
AU - Silenzio, Vincent
AU - Hoque, Md Ehsan
AU - Nichols-Hadeed, Corey
AU - Cerulli, Catherine
N1 - Funding Information:
In this paper, we show that individual level search histories can be used for identifying abusive relationship. Our work is an empirical demonstration that daily online search history has the potential to capture IPV among couples. We employed both the linguistic and contextual aspects of search logs to build model for identifying individuals in abusive relationships. We look forward to expanding our analysis across different forms of online activities and eagerly explore signals of other forms of related phenomena. This work has been funded by National Science Foundation Award #IIS1319378 , New York State Center for Excellence, Intel Science and Technology Gift, and DARPA IHMC sub-contract #W911NF-15-1-0542 .
Publisher Copyright:
© 2020
PY - 2021/3
Y1 - 2021/3
N2 - Intimate partner violence is a public health problem with increasing prevalence and harmful influence to both individuals and society. Automated screening for intimate partner violence is still an unsolved problem in academic research and practical applications. Current detection methods use self-reporting scales and in-person interviews, which are laborious, expensive, and often lack precision and sensitivity, making it essential to develop new approaches. This paper proposes a scalable and lightweight ubiquitous screening technique, validated via ground truth data collected through self assessment survey, for detecting signs of intimate partner violence by analyzing individual-level Google search histories. Initial analysis shows that there are temporal, textual, contextual differences in search behavior between individuals who have/haven't experienced intimate partner violence. Using these differentiating signals, we were able to build a model that can detect violence in intimate relationships with an F1 score of 0.80. Though preliminary, we hope our findings pave the way for the AI community to address this important public health problem.
AB - Intimate partner violence is a public health problem with increasing prevalence and harmful influence to both individuals and society. Automated screening for intimate partner violence is still an unsolved problem in academic research and practical applications. Current detection methods use self-reporting scales and in-person interviews, which are laborious, expensive, and often lack precision and sensitivity, making it essential to develop new approaches. This paper proposes a scalable and lightweight ubiquitous screening technique, validated via ground truth data collected through self assessment survey, for detecting signs of intimate partner violence by analyzing individual-level Google search histories. Initial analysis shows that there are temporal, textual, contextual differences in search behavior between individuals who have/haven't experienced intimate partner violence. Using these differentiating signals, we were able to build a model that can detect violence in intimate relationships with an F1 score of 0.80. Though preliminary, we hope our findings pave the way for the AI community to address this important public health problem.
KW - Intimate partner violence
KW - Online search history
KW - Public health
KW - Ubiquitous sensing from search engine logs
UR - http://www.scopus.com/inward/record.url?scp=85098696951&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85098696951&partnerID=8YFLogxK
U2 - 10.1016/j.smhl.2020.100161
DO - 10.1016/j.smhl.2020.100161
M3 - Article
AN - SCOPUS:85098696951
SN - 2352-6483
VL - 19
JO - Smart Health
JF - Smart Health
M1 - 100161
ER -