| Abstract [eng] |
Forecasting the dynamics of financial funds, often are difficult to choose one of the many existing prediction methods, because each of them has advantages and disadvantages. Performed widely used forecasting techniques Soros, Bayeso, AR, Wiener analysis. It is widely used in prediction methods of mathematical statistics. Often mathematical statistics methods examine the functions of the large number of variables. Large numbers of variables are the main problems in many cases for discreet optimization methods. To limit the number of errors in calculating it is need for more testing. Often number of experiments should increase exponentially. It is possible that the statistics of events do not behave according to the data of a certain situation. For these and other reasons there was seek for other forecasting methods. In this work, fuzzy logic used to predict the dynamics of financial funds. The developed system predicts event’s groups effects of funds curve, and are not based on statistics. This is expert prediction method. Fuzzy logic makes possible for experts to analyze and define the problematic area of vague rules. Fuzzy rules are made from statements which are easy to understand. Experts can detect the input data. Such input data error depends on the expert experience and intuition to assess the impact of events groups funds curve. In this work where analyzed approximation methods as interpolation and smoothing the curve. Approximation algorithm of Sklansky and Gonzalez was modified and realized. There are algorithms to predict changes of the graph . There was proposed prediction method which is based on fuzzy logic. |
