Presentation 2012-03-29
Estimation of answer correctness for recall tests with eye movements using hidden Markov models
Naoya TAKAHASHI, Minoru NAKAYAMA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this study, A mdel-based approach was proposed to estimate users' contextual understanding from features of their eye movements while viewing question statements in response to definition statements. For the estimation, 2 alternative forced choice tasks were performed, and features of subjects' eye movements while viewing question statements and features of correct and incorrect responses were extracted. Features of eye movements were defined as a sequence of symbols, saccade and fixation. Eye movements were modeled using hidden Markov models for the sequential data. This model have 2 states, "certain" and "uncertain". The former expressed high confidence in their answers to the questions, the latter expressed low confidence. While viewing question statements, state transitions happen repeatedly and the states output the symbols, saccade and fixation. Viewers' responses were classified into correct and incorrect from features of eye movements, with discriminative hidden Markov models estimated from training data. And certainty were estimated from the stationary distribution of each models. As a result, the prediction accuracy still leaves room for improvement, but this study shows that time series analysis can apply data of eye movements.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) eye movements / answer correctness / hidden Markov model / response estimation / discriminant analysis
Paper # PRMU2011-253,HIP2011-81
Date of Issue

Conference Information
Committee PRMU
Conference Date 2012/3/22(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Pattern Recognition and Media Understanding (PRMU)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Estimation of answer correctness for recall tests with eye movements using hidden Markov models
Sub Title (in English)
Keyword(1) eye movements
Keyword(2) answer correctness
Keyword(3) hidden Markov model
Keyword(4) response estimation
Keyword(5) discriminant analysis
1st Author's Name Naoya TAKAHASHI
1st Author's Affiliation Human System Science, Tokyo Institute of Technology()
2nd Author's Name Minoru NAKAYAMA
2nd Author's Affiliation Human System Science, Tokyo Institute of Technology
Date 2012-03-29
Paper # PRMU2011-253,HIP2011-81
Volume (vol) vol.111
Number (no) 499
Page pp.pp.-
#Pages 6
Date of Issue