Taxonomy of Fuzzy Multi-Attribute Decision Making Systems in Terms of Model , Inventor and Data Type

Decision support systems are one of the choices decision-makers make in an attempt to cope with the problems related to the time length required in decision-making process. Such systems are known to improve the efficiency and accuracy in the decision-making processes. In developing a decision support system, a certain calculation method is required as part of its processing. One of the most commonly used methods is FMADM. This research discusses the clustering of decision support system using FMADM in an attempt to provide a taxonomy of decision support system based on FMADM. Keywords-artificial intelligence; decision support system; fuzzy; taxonomy


INTRODUCTION
A decision support system (DSS) is a computerized system that will provide results in the form of ranking based on the assessment aspects determined by decision makers.DSSs are derived from expert systems and are part of the artificial intelligence (AI) field and of the applications that aim to help solving common knowledge-based cases [1].DSSs are systems that try to gather and exploit human knowledge and experience in artificial intelligence systems so that they may assist in, or even perform, decision making [2].Some examples of research on expert systems are stroke detection [3], animal disease identification [4,5] and motor engine damage detection [6].One of the algorithms used in DSSs is the Multiple Criteria Decision Making (MCDM) algorithm.However, MCDM is divided into several types.This paper, following a similar approach to the one in [7], provides a short literature review on MCDM taxonomy focusing on Multi Attribute Decision Making (MADM) aiming to provide a taxonomy of Fuzzy Multi-Attribute Decision Making Systems in Terms of Model, Inventor, and Data Type Methods.

II. RESULT AND DISCUSSION
MCDM is a decision-making method that can be used to establish the best choice from a number of alternatives based on certain criteria, e.g.size, standard etc [8].However, MCDM has a minor disadvantage: if the data provided by the decision maker or the attribute of the data is incomplete, then the resulting decision will contain uncertainty.The problem of uncertainty can be caused by several things, namely: 1.
Information that cannot be calculated, 2. Incomplete information, 3. Unclear information and 4. Partial abandonment [9].To solve these problems, some research on the use of Fuzzy MCDM began to be conducted in order to find methods that proved to have excellent performance.FMCDM can be divided into 2 models: fuzzy multi objective decision making (FMODM) and fuzzy multi attribute decision making (FMADM).FMADM model then can be further divided into 2 models namely the Yager and the Baas & Kwakernaak model.Based on the type of data, FMADM can be divided into 3 types, namely fuzzy data, crisp data, fuzzy and crisp data [10].While based on the method of application, FMADM can be divided into 3 types, namely SAW method, WP method and TOPSIS.FMADM taxonomies are shown in Figures 1-4 and are presented below.

A. FMADM Inventor-Based Taxonomy 1) Yager Model
The Yager model FMADM is the standard form of FMADM.According to [11],Yager model has 5 completion stages, which are: 1. Set a pairwise comparison matrix between attributes M based on Saaty's hierarchy procedure.
2. Determine the consistent weight of w j for each attributes for each attribute based on the eigenvector method of Saaty.
3. Calculate the value of 4. Determine the intersection of all 5. Select with the largest membership degree in , and set the optimal alternatives.
One of the researches related to DSS using Yager method is [12] which emphasize on theapplication of DSS to solve cases about the determination of families as poor.A similar research, [13], was conducted to solve the best customer selection case.Both researches resulted in a desktop-based decision support system that was able to assist the decision-making process in their respective cases.

2)
Baas In contrast odel is not a ften used by s he Baas &Kw nking of som ts [14].

Fig
Fig