Predicting Source Gaze Fixation Duration: A Machine Learning Approach

Tanik Saikh, Srinivas Bangalore, Michael Carl, Sivaji Bandyopadhyay

    Research output: Contribution to conferencePaperResearchpeer-review

    75 Downloads (Pure)

    Abstract

    In this paper an attempt has been made to predict the gaze fixation duration at source text words using supervised learning method, namely Support Vector Machine. The machine learning models used in the present work make use of lexical, syntactic and semantic information for predicting the gaze fixation duration. Different features are extracted from the data and models are built by combining the features. Our best set up achieves close to 50% classification accuracy.

    Original languageEnglish
    Publication date3 Mar 2015
    Number of pages6
    Publication statusPublished - 3 Mar 2015
    Event2015 International Conference on Cognitive Computing and Information Processing - JSS Academy of Technical Education, Noida, India
    Duration: 3 Mar 20153 Mar 2015
    http://www.ccip.jssaten.ac.in/

    Conference

    Conference2015 International Conference on Cognitive Computing and Information Processing
    LocationJSS Academy of Technical Education
    CountryIndia
    CityNoida
    Period03/03/201503/03/2015
    SponsorJSS Academy of Technical Education, Noida
    Internet address

    Keywords

    • Support Vector Machine
    • Eye Tracking
    • Gaze Fixation Duration
    • Machine Learning

    Cite this