Opinion Mining in Higher Education: A Corpus-Based Approach - Supplementary material – Grljević O.*, Bošnjak Z.**, Kovačević A.*** University of Novi Sad, Faculty of Economics in Subotica, Subotica, Serbia, oliverag@ef.uns.ac.rs University of Novi Sad, Faculty of Economics in Subotica, Subotica, Serbia, bzita@ef.uns.ac.rs University of Novi Sad, Faculty of Technical Science, Novi Sad, Serbia, kocha78@uns.ac.rs Corresponding author: Aleksandar Kovačević, Faculty of Technical Science, University of Novi Sad, Trg Dositeja Obradovića 6, 21000 Novi Sad, kocha78@uns.ac.rs 1. HiEd-Sent corpus analysis The HiEd-Sent corpus 1 was collected from the review website ‘Oceni Profesora’ (www.oceniprofesora.com), a Serbian language equivalent of the ‘Rate a Professor’ website. The website contains profiles of the teaching staff of all of the Serbian universities. By July 2015, the site had 3.5 million visits and 6,000 students had rated their professors. 2 Besides the unstructured text fields that were included in the corpus, each review contains data in structured format, i.e. four values that represent the rating (on the scale from 1 to 5) of the following features: usefulness and interestingness of the lectures, the quality of the communication with the students, and the availability and quality of the course materials. These features represent the basis of our annotation scheme. 1 The corpus and annotation guidelines in the Serbian language are freely available at http://www.ef.uns.ac.rs/hiedsent/. 2 Official Facebook page of website Oceni profesora https://www.facebook.com/rangiranjeprofesora.