Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Heterogeneous Interaction Network Analysis (HINA): A New Learning Analytics Approach for Modelling, Analyzing, and Visualizing Complex Interactions in Learning Processes

Shihui Feng, Baiyue He, Dragan Gasevic, Alec Kirkley

TL;DR

The paper addresses the inadequacy of existing learning analytics to capture heterogeneous interactions in modern learning environments. It introduces HINA, a framework that encodes learning processes as Heterogeneous Interaction Networks (HINs) and applies a tri-level analytics pipeline—node-level, dyadic-level, and meso-level—along with interactive visualization. Core contributions include quantity and diversity measures for individual engagement, binomial-null model-based edge significance testing, and a novel MDL-based nonparametric clustering for meso-level group discovery, demonstrated in a case study of AI-mediated small-group collaboration. The results show distinct engagement profiles and cluster-specific interaction patterns, supporting theory-building and informing the design of adaptive educational AI systems with broader applicability across contexts.

Abstract

Existing learning analytics approaches, which often model learning processes as sequences of learner actions or homogeneous relationships, are limited in capturing the distributed, multi-faceted nature of interactions in contemporary learning environments. To address this, we propose Heterogeneous Interaction Network Analysis (HINA), a novel multi-level learning analytics framework for modeling complex learning processes across diverse entities (e.g., learners, behaviours, AI agents, and task designs). HINA integrates a set of original methods, including summative measures and a new non-parametric clustering technique, with established practices for statistical testing and interactive visualization to provide a flexible and powerful analytical toolkit. In this paper, we first detail the theoretical and mathematical foundations of HINA for individual, dyadic, and meso-level analysis. We then demonstrate HINA's utility through a case study on AI-mediated small-group collaborative learning, revealing students' interaction profiles with peers versus AI; distinct engagement patterns that emerge from these interactions; and specific types of learning behaviors (e.g., asking questions, planning) directed to AI versus peers. By transforming process data into Heterogeneous Interaction Networks (HINs), HINA introduces a new paradigm for modeling learning processes and provides the dedicated, multi-level analytical methods required to extract meaning from them. It thereby moves beyond a single process data type to quantify and visualize how different elements in a learning environment interact and co-influence each other, opening new avenues for understanding complex educational dynamics.

Heterogeneous Interaction Network Analysis (HINA): A New Learning Analytics Approach for Modelling, Analyzing, and Visualizing Complex Interactions in Learning Processes

TL;DR

The paper addresses the inadequacy of existing learning analytics to capture heterogeneous interactions in modern learning environments. It introduces HINA, a framework that encodes learning processes as Heterogeneous Interaction Networks (HINs) and applies a tri-level analytics pipeline—node-level, dyadic-level, and meso-level—along with interactive visualization. Core contributions include quantity and diversity measures for individual engagement, binomial-null model-based edge significance testing, and a novel MDL-based nonparametric clustering for meso-level group discovery, demonstrated in a case study of AI-mediated small-group collaboration. The results show distinct engagement profiles and cluster-specific interaction patterns, supporting theory-building and informing the design of adaptive educational AI systems with broader applicability across contexts.

Abstract

Existing learning analytics approaches, which often model learning processes as sequences of learner actions or homogeneous relationships, are limited in capturing the distributed, multi-faceted nature of interactions in contemporary learning environments. To address this, we propose Heterogeneous Interaction Network Analysis (HINA), a novel multi-level learning analytics framework for modeling complex learning processes across diverse entities (e.g., learners, behaviours, AI agents, and task designs). HINA integrates a set of original methods, including summative measures and a new non-parametric clustering technique, with established practices for statistical testing and interactive visualization to provide a flexible and powerful analytical toolkit. In this paper, we first detail the theoretical and mathematical foundations of HINA for individual, dyadic, and meso-level analysis. We then demonstrate HINA's utility through a case study on AI-mediated small-group collaborative learning, revealing students' interaction profiles with peers versus AI; distinct engagement patterns that emerge from these interactions; and specific types of learning behaviors (e.g., asking questions, planning) directed to AI versus peers. By transforming process data into Heterogeneous Interaction Networks (HINs), HINA introduces a new paradigm for modeling learning processes and provides the dedicated, multi-level analytical methods required to extract meaning from them. It thereby moves beyond a single process data type to quantify and visualize how different elements in a learning environment interact and co-influence each other, opening new avenues for understanding complex educational dynamics.
Paper Structure (25 sections, 7 equations, 4 figures, 2 tables)

This paper contains 25 sections, 7 equations, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Heterogeneous Interaction Network Analysis (HINA): Concepts and Definitions
  4. Theoretical Underpinnings
  5. Review of Learning Analytics Methods
  6. Methods
  7. HINA Framework
  8. Methodological Details
  9. Individual-level analysis
  10. Dyadic-level analysis
  11. Meso-level analysis
  12. Case Study: Applying HINA to Model Human-AI Interactions in Small-Group Collaborative Learning
  13. Study Context and Participants
  14. Data Collection and Preprocessing
  15. Data Analysis Using HINA
  16. ...and 10 more sections

Figures (4)

  • Figure 1: The HINA multi-level analytical Framework
  • Figure 2: Individual interaction profiles of students within the student-partner HIN, showing the (unnormalized) quantity $s_i=WQ_i$ (Eq. \ref{['eq:quantity']}) and diversity $D_i$ (Eq. \ref{['eq:diversity']}) of their engagement with AI and peer (other student) targets. Two divergent profiles are visible: students with a diverse engagement strategy (non-zero diversity) and students with a focused strategy (zero diversity).
  • Figure 3: Engagement clusters identified through mesoscale analysis of the student-(content code, partner) HIN. Student nodes are colored by cluster membership (orange and blue for Clusters 1 and 2 respectively), while (content code, partner) nodes are colored in green. Double asterisks indicate the composite entities, i.e. "Appreciation ${}^{\ast\ast}$AI".
  • Figure 4: Aggregated interactions among content codes and partners within each cluster identified through the mesoscale clustering analysis of the student-(content code, partner) HIN. Interactions of the form student-AI are highlighted in red, and content codes are colored by their corresponding dimension---Cognitive, Metacognitive, Coordinative, or Socio-Emotional. Only statistically significant associations at $\alpha=0.05$ are presented in the figure, after the dyadic-level network pruning. Both Fig. \ref{['fig:engagement_clusters']} and Fig. \ref{['fig:engagement_each_cluster']} were generated using the HINA web-tool at hina-network.com.