Evaluation and selection of projects based on anfis: a call center case study

Fuzzy logic occurred as a follow-up of conventional set theory. Main point of the fuzzy logic is, for the question ?Is a member of a set?? is not answered definitely as yes or no like in conventional set theory, answer is a constant membership function which gives result between 0 and 1. Value of a membership function for a member is called as membership degree. As membership degree changes between 0 and 1, that provides opportunity to use verbal knowledge together with quantitative data for the solution of a problem. Using verbal expressions is the most important difference of fuzzy logic compared to other methodologies.Fuzzy systems are generally defined as the systems which use fuzzy set principles in order to procure output variables using selected present input variables. Fuzzy system?s most important advantage is, adding experiments and verbal information to fuzzy model in order to solve the problem. Fuzzy model is a system that is formed by fuzzy rules called as IF-THEN rules. Fuzzy logic is used for the solution of industrial engineering problems and decision making processes as it is applied at many other engineering areas. There two common fuzzy logic management methods. They are; Mamdani and Takagi Sugeno methods. Mamdani is fuzzy logic method that is used for any kind of problem which needs experimental knowledge. On the other hand Takagi Sugeno method is used for the problems which have not got many variables or these variables are not divided into many subsets. This report targets to find out the most useful project using Takagi Sugeno method.Adaptive network based fuzzy inference system (ANFIS) is used to model decision making process. Hybrid learning which is formed with the combination of subtractive clustering, back propagation and least square methods, is used to train ANFIS. ANFIS is a network structure which has learning capability that represents Takagi Sugeno type fuzzy systems. That network is composed of the nodes which are placed as layers in order to execute a definite function.