Extracting Robust Process Features with Time-Embedding N-Grams and Machine Learning Methods (Poster 10)

• In Event: Division D Graduate Student Committee In-Progress Research Gala Poster Session

Abstract

The computer-based assessment enables the collection of a broader range of records in log files throughout human-machine interactions. These granular records, often referred to as process data, were found informative to disclose respondents’ cognitive process when completing a digital task such as transforming geographical knowledge to spatial recognition in numerical problem. This study focuses on identifying robust process features in group disparities by latent numerical levels and gaining deeper insights into the potential reasons for problem-solving success and failure. This study also provided new evidence of the significance of process data in supporting the investigation of latent numeracy skills.

Authors

• Yiming Chen, University of Minnesota

• Qiwei He, Georgetown University