The analysis of students’ feedback written in natural language has been poorly considered in academic institutions, looking more frequently at students’ ratings as a base to evaluate courses and instructors. Statistical text analyses offer the possibility of exploring text collections from a quantitative viewpoint. Particularly interesting is Opinion Mining (OM), a family of techniques at the crossroads of Statistics, Linguistics and Computer Science. OM allows evaluating the sentiment of individual opinions, highlighting their semantic orientation. In an educational context, this approach allows processing students’ comments and creating powerful analytics. This paper aims at introducing readers to OM, presenting a strategy to compute the sentiment polarity of students’ comments. After explaining the rationale of the proposal and its mathematical formalisation, a toy example is presented to show how it works in practice. A discussion about theoretical and empirical implications offers some hints about its potentiality in a learning environment.

在学术机构中，对以自然语言编写的学生反馈的分析很少得到考虑，更经常地将学生的评分作为评估课程和讲师的基础。统计文本分析提供了从定量角度探索文本集合的可能性。特别有趣的是意见挖掘（OM），这是统计，语言学和计算机科学交叉领域的一种技术。OM可以评估单个意见的情感，突出其语义取向。在教育方面，这种方法可以处理学生的评论并创建功能强大的分析。本文旨在向读者介绍OM，提出一种计算学生评论情绪极性的策略。在解释了该建议的原理及其数学形式之后，给出了一个玩具示例来展示其在实践中的工作方式。有关理论和经验含义的讨论为它在学习环境中的潜力提供了一些提示。