Decision models evaluation using fuzzy pattern recognition

Abstract

Artificial intelligent models such as Artificial Neural Network, Genetic Algorithms, etc., play important roles in decision making. However, diversity in natures of intelligent models makes it difficult to decide which model is the most suitable one in solving certain decision problems. This paper proposes criteria of deferent suitability levels to evaluate the suitability of intelligent methods to decision environments, and presents a procedural method to calculate the suitability levels using Fuzzy Pattern Recognition. It analyzes various attributes related to decision problems and decision support systems' performance, then proposes a term SMP (Suitability of Model to decision Problem) for evaluating the suitability of decision models to certain decision problems. The influences of various attributes and preferences of decision makers are further explored, and a concept of SSMP (Synthetic SMP) is proposed to address human aspects in decision making. The paper presents a procedural method to calculate SMP and SSMP. Experimental calculation result shows this method is applicable in decision models selection.