Automatic evaluation of ERD in e-learning environments

Abstract

Entity Relationship Diagram (ERD) are largely in computer science degrees, related to database discipline. They are an important part of many evaluations. Currently, computer aided assessments are widely used for multiple choice questions, however, they do not have the ability to evaluate a student's knowledge more comprehensively, going beyond right or wrong, which is necessary for the job with the diagram. This research project presents an innovative approach for automatic evaluation of ERD. This approach proposes a solution to the challenge of encouraging students to perfect their solution: seeking, in addition to a response that returns the correct result, a note that approaches the ideal solution. This approach has the following advantages: (1) the student receives instant feedback during the practical diagramming activity, which allows the student to redo his solution for an optimal solution; (2) complete integration of ERD teaching concepts with examples of online diagrams; (3) monitoring the student's activities, i.e. how many examples were executed in each exercise, how many attempts were made. This research is a first step in building a fully assisted environment, for example, with automatic evaluation for ERD teaching, where the professor is freed from the arduous work of correcting diagrams and can perform more relevant pedagogical tasks. The method, based on machine learning techniques, uses structured query language (SQL) metrics extracted from the ERD and experts grade to create the prediction model. The solution can be adapted to other types of diagrams, such as the Unified Modeling Language (UML).