Automated Essay Grading: An Empirical Analysis of Ensemble Learning Techniques

Abstract

Listen

Translate article

Automated essay grading refers to the application of natural language processing tools for assigning scores to student essays. It is an important research domain as teachers are often required to grade a large amount of student essays in educational settings. Fair grading of essays is a challenging and tedious task. Teachers often consider this as unproductive work. Thus, there is a need for an automated approach for teachers so that they are no longer required to manually grade the student essays. Various automated essay scoring systems using machine learning and information retrieval concepts have been developed in the past studies. In the recent years, ensemble classification techniques have gained popularity. Ensemble techniques use multiple classifiers for making a prediction and have proved to be outperforming classical machine learning. In this paper, we present an empirical study of ensemble learning techniques for classification of student essays. We studied performance of five machine learning and four ensemble learning techniques for conducting experiments. We further utilized feature selection technique to improve the prediction efficiency. The performance results on automated student assessment prize dataset available on Kaggle showed that ensemble techniques outperform the efficiency of traditional machine learning techniques.