Unit:
Κατεύθυνση Γλωσσολογία: θεωρία και εφαρμογέςLibrary of the School of Philosophy
Author:
Nektariou Angeliki
Supervisors info:
Γεώργιος Μικρός, Καθηγητής, Τμήμα Ιταλικής Γλώσσας και Φιλολογίας, Φιλοσοφική Σχολή ΕΚΠΑ
Βασιλική Μητσικοπούλου, Αναπληρώτρια Καθηγήτρια, Τμήμα Αγγλικής Γλώσσας και Φιλολογίας, Φιλοσοφική Σχολή ΕΚΠΑ
Νικόλαος Λαβίδας, Επίκουρος Καθηγητής, Τμήμα Αγγλικής Γλώσσας και Φιλολογίας, Φιλοσοφική Σχολή ΕΚΠΑ
Original Title:
A Statistical Approach to Automatic Essay Scoring
Translated title:
A statistical approach to automatic essay scoring
Summary:
Taking into consideration escalating need for testing writing ability and the potential of Automatic Essay Scoring (AES) to support writing instruction and evaluation, the aim of the present study is to explore the relationship between stylometric indices, widely used in AES systems, and the degree of sophistication of learner essays, captured by the score provided by expert human raters. The data analyzed were obtained from a recently organized public AES competition and comprise persuasive essays written in the context of public school in the United States. Stylometric information taken into consideration greatly focuses on measures of cohesion, as well as lexical diversity and syntactic sophistication. Results indicate a clear relationship between quantifiable features of learners’ written responses and the impression which they have made on expert raters. This observation reinforces the importance of pursuing further experimentation into AES, which would yield significant educational and social benefits.
Main subject category:
Language – Literature
Keywords:
automatic essay evaluation, natural language processing, stylometry
