Regresar

Sentiment Analysis in the Feedback of Peer Evaluation Activities

Abstract:

Sentiment analysis is a technique used more frequently in the educational field. For the present work, the analysis and classification of the feedback comments issued by the students in the peer evaluation activities has been taken as the main application approach. Determining the polarity of these comments can help the teacher to identify characteristics and patterns in the criteria issued by the students to enrich the teaching-learning process. The present work aims to determine the polarity of feelings of the feedback comments of the peer evaluation activities planned as challenges within the courses offered by the Open Campus initiative. To do this, experimentation is carried out in three training scenarios and tests of the classification model using the corpus of tweets written in Spanish TASS and a corpus of comments extracted from the learning platform, manually classified by experts. Among the main results, it is observed that many students give feedback that is useful, be it positive or negative. However, there is a significant percentage of comments that are perceived as unjustified or incomprehensible, and this is observed in the number of comments classified as neutral and without polarity.