FCM Clustering on Interaction Pattern Analysis of Chinese Language Learner Behavior.

Area of Science:

• Educational Technology
• Computer Science
• Artificial Intelligence

Background:

• Current digital learning systems lack personalization, hindering effective education.
• Personalized learning requires analyzing user behavior patterns to adapt educational content.
• Existing clustering methods like Fuzzy C-means (FCM) are sensitive to initial cluster centers.

Purpose of the Study:

• To develop a personalized language learning system using user behavior analysis.
• To propose an Improved Harmony Search-FCM (HIS-FCM) algorithm for clustering Chinese language learners.
• To enhance the accuracy and efficiency of learner level division and course optimization.

Main Methods:

• Obtaining user language learning interaction data to determine initial and current language levels.
• Utilizing an adaptive algorithm to update the initial learning model based on current language level.
• Applying a Harmony Search-FCM (HS-FCM) algorithm, specifically an Improved HS (HIS)-FCM, for clustering learning behavior.
• Selecting participation, focus, regularity, interaction, and academic performance as behavior analysis indicators.

Main Results:

• The HIS-FCM algorithm demonstrates higher clustering accuracy compared to HS-FCM and decision tree methods.
• The proposed algorithm exhibits faster convergence speed and lower fitness values.
• Learner levels are effectively divided into five categories: excellent, good, medium, qualified, and poor.

Conclusions:

• The HIS-FCM algorithm offers a robust approach for analyzing online Chinese language learning behavior.
• This method provides new opportunities for accurate learner level division and personalized course optimization.
• The integration of service concepts and user behavior analysis is crucial for advancing digital learning systems.