A Hidden Markov Model-Based Analysis of the Interaction Patterns between Cognitive Styles and Lexical Productivity in Second Language English Learners

Abstract

Vocabulary productiveness refers to learners' ability to use vocabulary correctly and actively in oral or written expressions, and its learning pattern is often influenced by learners' cognitive styles. In this paper, we integrate the feature indicators related to learning effect in learners' behavioral data, predict learners' learning effect through machine learning algorithms, and construct a model of learning effect of learners' cognitive style based on multimodal data. Meanwhile, based on the learner behavioral feature data, the decision tree of learner cognitive styles is established, and the Hidden Markov Algorithm is used to establish a Hidden Markov Model for the preference of each cognitive style, thus proposing the recognition method of learner cognitive styles. With the support of this method, the average score of learners' vocabulary output test in the experimental group increased by 23.08 points compared to the preexperimental period, and showed a statistically significant difference with the average score of the students in the control class (P=0.000). The proposed method of identifying learners' cognitive styles provides a new design idea and empirical basis for teaching vocabulary productiveness and interaction model for second language English learners.