System ekspertowy do oceny układu systemu dystrybucji wody sporządzony za pomocą wnioskowania indukcyjnego

At present often in computer programs the methods of computational intelligence are used, in this expert systems. In building of expert systems the process of knowledge acquisition is one of the principle problem. The improvement of knowledge acquisition is received by use of the machine learning methods. The most popular strategy of knowledge acquisition in the machine learning methods is the inductive inference, in this induction of decision trees. Inductive inference is the process of reaching a general conclusion from specific examples. This paper presents results of the induction of the decision tree intended to evaluation of water distribution system. Rules kept in the decision tree make possible to estimate the new project variants of the water supply network. Applying the induction of decision trees entails the preparation of a set of examples learners. Collection of examples should be representative and sufficiently describe the specific features of the problem. Using information on the eight water supply systems and the parameters of their work, the computations were performed water distribution system. The calculation results were the basis for the calculation defined in this study the variables that characterize the solution of the water distribution system. Five classes are defined to describe the water supply problems due to improper water distribution system and one associated with the correct solution. Each class will be selected on the basis of defined variables. Therefore, the decision tree should lead to the assignment of variables describing the system of water distribution to the appropriate class, characterized by the solution of the system. Using inductive inference obtained decision rules that can be used in expert system that can work with the program for the simulation of hydraulic water distribution systems.