Expert system to evaluate development tests in the Universidad Continental students

Abstract

Objectives: To build an expert system to evaluate development tests in informatics and to prove if its application enables the development tests evaluation with a quantitative score which allows students to evaluate themselves at any time and receive instant information about his/her knowledge level. Methods: The research type was applied, field research, descriptive and transversal. Exainfo basically has two parts: the knowledge base and the inference engine to obtain a test development score. For the applicability evaluation, we built the subject evaluation matrix, after the features model to obtain the rating and empty words, in that way we qualify students built responses and indicate the missing concepts in learning; in this process we were careful with the concurrence principle by facial capture validation, to prevent student impersonation. The sample was formed by 40 Universidad Continental first semester students from the Informatics subject. Results: The recovery rate average of 72,38 % indicates that most of the basic concepts are recovered by students. The retrieved basic concepts are too many so the accuracy rate is reduced. Conclusions: The Exainfo expert system makes possible to qualify development tests and diagnose the student learning state through the free text analysis of student responses.