Save the date! Next SNOLA webinar will be on October 8th, 16:00

Sign up for the webinar here.

While topic modeling has been accessible to the ‘regular’ person since about the late ’80s and early ’90s, it effectively became of mainstream interest roughly around 2010, and it has been on a relatively steady rise ever since. Topic modeling is, arguably, one of the core techniques or methods currently available to perform semantically relevant analyses of texts, and its outputs can be extremely important for consequent analyses (e.g., document classification, information retrieval, content recommendation, text summarization, etc.)

As an unsupervised learning technique, topic modeling reduces human error and bias, at least in its early stages. It is also ‘human-friendly’, meaning that it often is much more interpretable than related NLP techniques. Therefore, its wide adoption (and preference) by scientists from across a huge variety of academic and scientific fields is not surprising. However, its ‘human-friendliness’ does not preclude the need for rigor and consistency.

In this webinar, we will cover the basis of topic modeling and its main approaches, its applications in learning analytics, a few practical examples of topic modeling in R, and several broad perspectives about its place in a larger research methodology (rather than as an end in itself).

Speaker: Dorin Stanciu

Biography: Dorin is an Associate Professor at the Technical University of Cluj-Napoca (Romania), the most fertile ‘source’ of automation and computer science graduates in Romania. He holds a PhD in Education Sciences and a PhD in Psychology by Babes-Bolyai University. Dorin is Editor in Chief of the journal Computers in Human Behavior Reports and Editorial Board Member for CHB. Besides teaching Teacher’s Training programs, he also teaches Ergonomics (Human Factors), with a heavy focus on the psychological aspects of human-computer, human-robot interaction, and UI/UX.