Analysis of Trainee Performance for Automating Training and Scenario Recommendations
Thorough trainee evaluation requires subject matter experts to assess current performance and recommend future training that is optimized for each individual. Presently, with limited subject matter expert availability and large volumes of data generated from today’s training environments, a comprehensive and consistent trainee assessment is impractical, if not impossible. Typically, this impacts the quality of analysis required to determine a trainee’s actual proficiency, which in turn degrades the ability to select the most appropriate future training, which further degrades the ability to determine if a scenario can truly ensure that training objectives are mastered.
To overcome this problem, we have designed a system that ingests training data, calculates predetermined Measures of Performance and then provides scenario recommendations that ensure trainees master warfighting skills as intended. By tracking trainees throughout their career, we can identify when a skill is mastered and how often trainees need to practice that skill to stay proficient.
This presentation will provide an overview of the selected approach, system design, and evaluation. In close cooperation with operational users, a concept of operations has been defined and a system architecture was developed that allows processing of large volumes of training data and is adaptable to various file types and file formats. Leveraging commercial off-the-shelf products, a demonstration system has been built. Using example datasets from two different training environments producing multiple data formats, we were able to successfully demonstrate data recognition and ingestion, Measure of Performance evaluation, trainee tracking, and optimized scenario recommendations.