Weakly Supervised Models of Aspect-Sentiment for Online Course Discussion Forums

Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, pages 74–83, Beijing, China, July 26-31, 2015. c⃝2015 Association for Computational Linguistics Weakly Supervised Models of Aspect-Sentiment for Online Course Discussion Forums Arti Ramesh,1 Shachi H. Kumar,2 James Foulds,2 Lise Getoor2 1University of Maryland, College Park 2University of California, Santa Cruz artir@cs.umd.edu, {shulluma, jfoulds, getoor}@ucsc.edu Abstract Massive open online courses (MOOCs) are redefining the education system and transcending boundaries posed by tradi- tional courses. With the increase in pop- ularity of online courses, there is a cor- responding increase in the need to under- stand and interpret the communications of the course participants. Identifying top- ics or aspects of conversation and inferring sentiment in online course forum posts can enable instructor interventions to meet the needs of the students, rapidly address course-related issues, and increase student retention. Labeled aspect-sentiment data for MOOCs are expensive to obtain and may not be transferable between courses, suggesting the need for approaches that do not require labeled data. We develop a weakly supervised joint model for aspect- sentiment in online courses, modeling the dependencies between various aspects and sentiment using a recently developed scal- able class of statistical relational models called hinge-loss Markov random fields. We validate our models on posts sam- pled from twelve online courses, each con- taining an average of 10,000 posts, and demonstrate that jointly modeling aspect with sentiment improves the prediction ac- curacy for both aspect and sentiment. 1 Introduction Massive Open Online Courses (MOOCs) have emerged as a powerful medium for imparting edu- cation to a wide geographical population. Discus- sion forums are the primary means of communica- tion between MOOC participants (students, TAs, and instructors). Due to the open nature of these courses, they attract people from all over the world leading to large numbers of participants and hence, large numbers of posts in the discussion forums. In the courses we worked with, we found that over the course of the class there were typically over 10,000 posts. Within this slew of posts, there are valuable problem-reporting posts that identify issues such as broken links, audio-visual glitches, and in- accuracies in the course materials. Automati- cally identifying these reported problems is impor- tant for several reasons: i) it is time-consuming for instructors to manually screen through all of the posts due to the highly skewed instructor-to- student ratio in MOOCs, ii) promptly address- ing issues could help improve student retention, and iii) future iterations of the course could ben- efit from identifying technical and logistical is- sues currently faced by students. In this paper, we investigate the problem of determining the fine-grained topics of posts (which we refer to as “MOOC aspects”) and the sentiment toward them, which can potentially be used to improve the course. ied,While aspect-sentiment has been widely stud- the MOOC discussion forum scenario presents a unique set of challenges. Labeled data are expensive to obtain, and posts containing fine- grained aspects …

Content truncated to 500 words