Empleo de las listas de adyacencia para la paralelización del algoritmo matricial de los árboles de decisión

摘要: The difficulty in the use of multiprocessors systems consists which maximum and nearer productivity system can only be reached when all or majority processors are loaded with useful work. equitable load depends on two necessary components: structures parallelization used device parallelization. In this article a method algorithm sets out represented by an information graph. input data for analysis modification evidence interrelation operations data. stored form adjacency listings, allows to save memory considerably accelerate process processing these. output borders interval that includes volume calculation resources allow make decisions following given as well planning execution operation. exposes one results verification algorithms example construction matrix decision trees.