Integration of Automated Essay Scoring Models by Item Responce Theory

Abstract

Automated Essay Scoring (AES) is the task of automatic grading essays instead of using human raters. Many AES models offering different benefits have been proposed in the past few decades. This study proposes a new AES model averaging framework using item response theory. The proposed framework can improve scoring accuracy because it averages prediction scores from various AES models while considering characteristics of each model for evaluation of examinee ability.