AN AUTOMATIC EDUCATIONAL QUALITY EVALUATION FUZZY SYSTEM C. García 1 , L.G. Pérez 2 , L. Martínez 2 , B. Montes 3 and P.J. Sánchez 2 1 Dept. of Computer Science and A.I., University of Granada, 18071 - Granada, Spain e-mail: gcarlos@fedro.ugr.es 2 Dept. of Computer Science, University of Jaén, 23071 - Jaén, Spain, e-mail:martin@ujaen.es, pedroj@ujaen.es 3 Dept. of Psicology, University of Jaén, 23071 - Jaén, Spain e-mail:bmontes@ujaen.es Abstract The surveys about educational quality in the University are more and more important in nowadays. However, current surveys are not flexible at all, due to the fact that all the questions are assessed in the same domain despite that they are not related to or they have different nature. The aim of this paper is to present an evaluation approach for educational quality that can manage heterogeneous information (numerical, interval valued and linguistic) based on the fuzzy linguistic approach and afterwards will be outlined its Unified Modelling Language (UML) scheme that facilitates its implementation in a JAVA based system. To do so, this approach is based on linguistic decision and evaluation approaches presented in the literature [4,5,8]. The UML scheme will be the basis to develop an object oriented evaluation system implemented in Java. This system will offer a great flexibility to the users (students who answer the pool) because they could provide their responses in different domains according to the nature of the questions used to evaluate the educational quality. In addtion this software system improve the procees of obtaining the final results about the qualification obtained by each lecturer or professor due to it is a automatic system and the pools can be filled, processed and stored in a automatic way. Keywords Fuzzy system, heterogeneous information , quality evaluation. 1. INTRODUCTION The evaluation of the educational quality in the Universities referred to the skills of their lecturers and professors is more and more important nowadays. This evaluation is usually carried out by means of surveys in which the students qualify different aspects related to the educational and research skills of the lecturers and professors. These surveys force the students to express their opinions or preferences in a given numerical scale belong to a specific domain in spite of the different aspects to qualify may have different nature and the knowledge about them is vague and imprecise. The use of fuzzy techniques have provided good results in different evaluation problems [3,10]. In this contribution we present the UML scheme of the JAVA based system developed for the Office of Quality for the Universities of Andalucía (South of Spain) that facilitates, improves and automatizes its current evaluation processes. This system is based on linguistic decision and evaluation models presented in [5,6,8] that manage heterogeneous information (numerical, interval valued and linguistic), so it will offer a greater flexibility and better rapport with reality to face this type of evaluation problems than the current ones. This contribution is structured as follows: in the section 2 we review briefly the decision process that has been used to design the UML scheme for implement the evaluation JAVA system dealing with heterogeneous information, in the section 3 we shall present the UML scheme of the evaluation JAVA system. Eventually some concluding remarks and future work is pointed out.