Introducción

Abstract

Nowadays, physical and software systems, designed and built through engineering processes and software, are everywhere: our home and our office are full of electronic devices, our factories are almost fully automatized, we have cars and trucks full of complex electronic systems, almost every electronic system contains hundreds or thousands of lines of code, our computers run operating systems made up of hundreds of small programs, etc. Hence, it is required that these systems work as expected and as safer as possible. For these tasks, automated diagnosis is mandatory because for most devices, it is almost impossible to obtain necessary experience to build knowledge-based systems before they become obsolete, or there are enough number of variants for the same mechanism that it is not possible to adjust existing solutions. In those cases where a lot of data are available, it would be possible to learn models using data-driven techniques