Detecting Deception and Suspicion in Dyadic Game Interactions

摘要: In this paper we focus on detection of deception and suspicion from electrodermal activity (EDA) measured left right wrists during a dyadic game interaction. We aim to answer three research questions: (i) Is it possible reliably distinguish truth based EDA measurements interaction? (ii) the state trust card game? (iii) What is relative importance wrists? To our questions conducted study in which 20 participants were playing Cheat pairs with one sensor placed each their wrists. Our experimental results show that measures provide more information for than person-dependent reliable person-independent detection. particular, classifying signal Support Vector Machine (SVM) yields accuracies 52% 57% prediction respectively, 63% 76% respectively. Also, found that: optimal interval informative about 1 s while around 3.5 detection; relevant deception/suspicion can be captured after 3.0 seconds stimulus occurrence regardless type (deception/truthfulness/suspicion/trust); features extracted both are important classification suspicion. best knowledge, first work uses data automatically detect interaction setting.