Machine learning-driven model construction for automated classification of cognitive styles.

Area of Science:

• Cognitive Science
• Human-Computer Interaction
• Machine Learning

Background:

• Personalized learning and HCI systems require accurate cognitive style identification.
• Traditional self-report methods for cognitive style assessment are prone to subjectivity bias.

Purpose of the Study:

• To develop an objective machine learning model for classifying cognitive styles using physiological data.
• To differentiate between verbal and representational cognitive styles non-invasively.

Main Methods:

• Collected eye-movement data from 85 participants during a standardized cognitive task using eye-tracking technology.
• Extracted multidimensional eye-movement features.
• Evaluated six machine learning algorithms (DT, KNN, NB, SVM, LR, EL) for classification performance.

Main Results:

• All evaluated machine learning algorithms effectively utilized eye movement features for cognitive style classification.
• The Support Vector Machine (SVM) algorithm achieved optimal performance with 82.1% classification accuracy (F1=0.715) after parameter optimization.
• The proposed method provides a novel, non-invasive approach to cognitive style assessment.

Conclusions:

• Machine learning models analyzing eye-movement data offer a viable alternative to subjective self-report measures for cognitive style identification.
• This approach has significant implications for real-time adaptive learning systems, personalized educational technology, and HCI design.
• Findings provide valuable insights for educational psychology and human-computer interaction research.