Course abstract
This short course introduces fundamental knowledge in data science methods that could be used to tame complex process data in sequential format associated with timestamps and highlights advanced applications of sequence mining in analyzing process data to better support group-level (in)variance explorations of behavioral patterns in large-scale assessments. Specifically, the presenters will focus on four subtopics, including (1) how to extract and select gram-based features from clickstream sequence, (2) how to compute sequence distance to identify pairwise sequence similarity, (3) how to integrate timing information into sequence-based analysis, and (4) how to use latent sequence models (e.g., hidden Markov model) to identify latent process states and transitions.
During the three-hour short course, participants will be provided with an overview of process data collected from computer-based large-scale assessments, learn about various approaches to analyzing process data with sequence mining methods, and obtain hands-on experience with sequential process data analysis through examples and exercises. Intended audience are researchers, students, and practitioners with basic knowledge of process data and familiarity with fundamental programming skills and interested in learning or applying data-driven methods to process data analysis.