Sensors Model Student Self Concept in the Classroom

Cooper, David G.; Arroyo, Ivon; Woolf, Beverly Park; Müldner, Kasia; Burleson, Winslow; Christopherson, Robert

doi:10.1007/978-3-642-02247-0_6

Cited by 35 publications

(35 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the Fall of 2008 and the Spring of 2009 the geometry tutor Wayang Outpost was deployed with a set of sensors into real classroom environments [7,15,16]. The set of sensors included: a mouse that captured degree of pressure placed on its various points, a bracelet that measured skin conductance of the wrist, a chair that sensed the level of pressure on the chair back and seat, and a camera supplemented with software for facial emotion recognition.…”

Section: Setupmentioning

confidence: 99%

“…Each student's physiological data and interactions with the tutor were logged. Subsequently, the interaction and sensor data were time-aligned and converted into tutor and sensor features, as described in [7]. At intervals of five minutes in the Fall, and three minutes in the Spring, students were presented with an emotional query about one of four affective states (confident, interested, frustrated, or excited ) selected from a uniform random distribution.…”

Section: Setupmentioning

confidence: 99%

“…For each student and for each emotion there are between two and five affective-state labels. For more detail on the full specification of these features see [7]. …”

Section: Tutor and Sensor Featuresmentioning

confidence: 99%

“…We also performed the same cross-validation on a linear classifier with a constant model, which we used as our baseline. This step differs from [7] in two ways: 1) The mean was taken across each test student instead of across tests. 2) We calculated sensitivity and specificity in addition to accuracy.…”

Section: Cross Validationmentioning

confidence: 99%

“…In each experiment, students were queried about four emotional states (confident, interested, frustrated, and excited ), providing the standard for validating our models. The study data was used to construct a number of linear classifiers for each emotional state, as we reported in [7]. The best classifiers for a given emotion obtained accuracies between 78% and 87% according to a leave-one-student-out cross-validation.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Ranking Feature Sets for Emotion Models Used in Classroom Based Intelligent Tutoring Systems

Cooper

Müldner

Arroyo

et al. 2010

User Modeling, Adaptation, and Personalization

Self Cite

View full text Add to dashboard Cite

Abstract. Recent progress has been made by using sensors with Intelligent Tutoring Systems in classrooms in order to predict the affective state of students users. If tutors are able to interpret sensor data with new students based on past experience, rather than having to be individually trained, then this will enable tutor developers to evaluate various methods of adapting to each student's affective state using consistent predictions. In the past, our classifiers have predicted student emotions with an accuracy between 78% and 87%. However, it is still unclear which sensors are best, and the educational technology community needs to know this to develop better than baseline classifiers, e.g. ones that use only frequency of emotional occurrence to predict affective state. This paper suggests a method to clarify classifier ranking for the purpose of affective models. The method begins with a careful collection of a training and testing set, each from a separate population, and concludes with a non-parametric ranking of the trained classifiers on the testing set. We illustrate this method with classifiers trained on data collected in the Fall of 2008 and tested on data collected in the Spring of 2009. Our results show that the classifiers for some affective states are significantly better than the baseline model; a validation analysis showed that some but not all classifier rankings generalize to new settings. Overall, our analysis suggests that though there is some benefit gained from simple linear classifiers, more advanced methods or better features may be needed for better classification performance.

show abstract

Section: Setupmentioning

confidence: 99%

Section: Setupmentioning

confidence: 99%

“…For each student and for each emotion there are between two and five affective-state labels. For more detail on the full specification of these features see [7]. …”

Section: Tutor and Sensor Featuresmentioning

confidence: 99%

Section: Cross Validationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations