ZETEMA: A web service for automatic short-answer questions grading

ZETEMA is a prototype of a web service that allows a user to train an automated system for scoring short-answer questions, and use it for scoring new student answers. In addition, the service provides already-trained models based on standard benchmarks such as SciEntsBank and Beetle, used in SemEval 2013 Student Response Analysis challenge. The service is based on the system that obtained the best results in that competition (soft cardinality) and includes some improvements in time, simplifications in the model preserving performance. ZETEMA was developed in Python and can be used via a web interface using XML format for sending, training and testing texts, and also for retrieving result.