Online discussion forums (or discussion boards) are one of the most common tools in web-based teaching- learning environments. Students' activity in discussion threads can be a relevant source of information that facilitates the monitoring tasks during the course by providing teachers with relevant indicators of their students' needs and lacks. In the present paper, the use of time series and an agglomerative hierarchical clustering algorithm is proposed with the aim of determining what different behavior patterns are adopted by students in online discussion forums. To this end, the actions carried out by students along the threads (e.g., writing and reading) are used to represent their activity in times series form. The use of an agglomerative hierarchical clustering algorithm is proposed in order to group similar students according to their activity profile. Some strategies on how to cut the obtained dendrograms are discussed and preliminary experimental results are presented.