Automated generation of test units for diagnosis and development of cognitive skills

Abstract

The paper presents the architecture of a web-based system for automated generation of test units for diagnostics and development of cognitive skills. The system supports a database (DB) containing information objects of different type (text, images), including metadata used in the generation process. The database is updated through queries to external information sources (to supply the objects themselves) and with next addition of attributes. For example, an image can be accompanied by a textual description and metadata (as part of an e-learning process), a Bulgarian lexema - with automatically generated phonetic transcription and syllabicate, and so on. The supported and updated specific image metadata system (including color, size, purpose, etc.) and for lexams (composite sounds, syllables, etc.) provides an opportunity to generate test units for diagnosis and cognitive development. The selection of objects from DB suitable for insertion in a test unit of a certain type (for example, to distinguish objects with different characteristics) is performed by a query to the database for extraction of objects according to the value of the accompanying metadata.

Keywords

• cognitive skills
• cognitive language assessment tasks
• cognitive task model
• automated test unit generation