A0596
Title: A path signature perspective of process data feature extraction
Authors: Xueying Tang - University of Arizona (United States) [presenting]
Abstract: Computer-based interactive items have become prevalent in recent educational assessments. In such items, the entire human-computer interactive process is recorded in a log file as timestamped action sequences. Such response process data are noisy, diverse, and in a nonstandard format. Several methods have been developed to extract information from response processes. However, these methods often focus on the action sequences and ignore the timestamps in response processes. A new feature extraction method is introduced that incorporates the information in both action sequences and timestamp sequences. Based on the concept of path signature, the proposed method extracts features characterizing the response processes at different levels of detail. The proposed method is applied to both simulated data and real response process data from PIAAC to compare the information contained in action and timestamp sequences and demonstrate the potential benefit of incorporating time information for assessing respondents' latent ability.
