본문 바로가기
  • Home

Reliable Growth Estimation Using Theil-Sen: A Comparative Study in AI-Driven ALEKS Program

  • Journal of The Korea Society of Computer and Information
  • Abbr : JKSCI
  • 2025, 30(1), pp.241-249
  • Publisher : The Korean Society Of Computer And Information
  • Research Area : Engineering > Computer Science
  • Received : November 1, 2024
  • Accepted : December 27, 2024
  • Published : January 31, 2025

Hyunsuk Han 1 WANG XINYI 1

1전북대학교

Accredited

ABSTRACT

The recent acceleration in the expansion of Artificial Intelligence (AI) powered learning platforms has enabled the collection of large volumes of learning data, thereby overcoming the limitations of traditional data collection methods. This study focuses on analyzing progress data from students using ALEKS, an AI-based online platform, and compares the ordinary least squares method with the non-parametric Theil-Sen method for estimating progress slopes. The findings indicate that while both models produce similar slope estimates, the Theil-Sen model is more reliable, especially when dealing with extreme values, which are often present among low achievers. These results highlight the necessity of considering alternative analytical methods when parametric estimation methods are inappropriate, to obtain a reliable estimates. Furthermore, this study underscores the need for ongoing research to explore more precise methods for analyzing AI-driven data, particularly in light of the limitations of parametric assumptions inherent in traditional estimation models.

Citation status

* References for papers published after 2023 are currently being built.

This paper was written with support from the National Research Foundation of Korea.