An enhanced Elo-based student model for polychotomously scored items in adaptive educational system

ABSTRACT

The adaptive learning environment provides learning support that suits individual characteristics of students, and the student model of the adaptive learning environment is the key element to promote individualized learning. This paper provides a systematic overview of the existing student models, consequently showing that the Elo rating system has greater potential as compared to the other models regarding application in the online learning environment. Based on the Elo model, this study proposes the EELO, an enhanced Elo rating system, in consideration of the application scenarios of polychotomously scored items and multi-dimensional granularity evaluations not covered by the basic Elo rating system. The EELO model estimating students’ cognitive abilities and predicting their future performances on unknown questions is evaluated based on one public set (Assigment2) and one proprietary dataset (HSK), and achieved an AUC of 0.92 for Assigment2 and 0.84 for HSK, which shows that the EELO model has the best performance regarding the above-mentioned objectives as compared with the latest extensions of the IRT and BKT models. Subsequently, the EELO model was tested and applied successfully in a real large-scale online learning environment to demonstrate the potential of the EELO model in adaptive learning applications.

KEYWORDS:

• adaptive learning system
• student model
• Elo rating system
• polychotomously scored item
• online learning environment

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the National Natural Science Foundation of China: [Grant Number 62007024]; China Educational Equipment Industry Association: [Grant Number CEFR18009R4]; Shanghai Sailing Program: [Grant Number 17YF1428400].

Notes on contributors

Bingxue Zhang

Bingxue Zhang received her Ph.D. in Computer Science in 2014 from Ecole Centrale de Lyon, France. She works as a lecturer at the University of Shanghai for Science and Technology China. Her research activities are focused on human contextual learning, mobile learning, smart learning, etc.

Yang Shi

Yang Shi received a B.S. degree in Computer Science and Technology from Shanghai Second Polytechnic University, Shanghai, in 2019. And he is currently pursuing a master degree at the University of Shanghai for Science and Technology, China. His research activities are focused on smart learning, etc.

Yuxing Li

Yuxiang Li received a B.S. degree in Computer Science from Shanghai DianJi University, Shanghai, in 2016. And he is currently a graduate student at the University of Shanghai for Science and Technology China. His research activities are focused on natural language processing, smart learning, etc.

Chengliang Chai

Chengliang Chai received a B.S. degree in Computer Science and Technology from Shandong Yingcai University, Shanghai, in 2019. And he is currently pursuing a master degree at the University of Shanghai for Science and Technology, China. His research activities are focused on smart learning, etc.

Longfeng Hou

Longfeng Hou received his Ph.D. in 2015 from INSA Lyon, France. He works as a lecturer at University of Shanghai for Science and Technology China. His research activities are focused on data mining, smart learning, etc.