Sentiment-Oriented Summarisation of Peer Reviews

Authors: 

Sunghwan Mac Kim and Rafael A. Calvo

Paper type: 
Poster

Abstract: 

It is common that students peer-review other students’ writing, and these reviews are useful information to instructors, both on the particulars of the essay being reviewed, the feedback provided and the overall progress of the class. This paper describes a novel approach to summarising feedback in academic essay writing. We present a summarisation method for identifying and extracting representative opinion sentences from each feedback. Sentiment score-based techniques are employed and SentiWordNet is used as a linguistic lexical resource for sentiment summarisation. We evaluate our approach with the reviews written by a group of 50 engineering students.

Keywords: 

sentiment summarisation, peer review, student feedback