An Ensemble-Based Model to Improve the Accuracy of Automatic Short Answer Grading

Since 1966 much research has been done on the automatic grading of student answers task, and it was divided into short answer grading and essay scoring. In this paper, on the short answer grading challenge, we are working with text similarity approaches that are being classified into string, semantic (corpus and knowledge-based), and embedding text similarity approaches are the three types of text similarity techniques. On the Texas data-structure data set, different experiments were examined individually before being merged to give a maximum correlation result of 0.65.