CASPIAN JOURNAL
MANAGEMENT AND HIGH TECHNOLOGIES
Experience with using fuzzy sets in forecasting of currency exchange rate
 |
Read | Pilyugina Anna V., Boiko Andrey A. Experience with using fuzzy sets in forecasting of currency exchange rate // Caspian journal : management and high technologies. — 2014. — №3. — pp. 143-157. |
Pilyugina Anna V. - Ph.D. (Economics), Assistant Professor, Moscow State Technical University named after Bauman, 5 (building 1), 2-ya Baumanskaya str., Moscow, 105005, Russian Federation, pilyuginaanna@mail.ru
Boiko Andrey A. - postgraduate student, Moscow State Technical University named after Bauman, 5 (building 1), 2-ya Baumanskaya str., Moscow, 105005, Russian Federation, boiko_andrew@mail.ru
In article is shown practical implications of currency exchange rate forecasting for all currency market participants. Classification of forecasting methods is presented. Authors paid particular attention to formalized forecasting methods. Such method using fuzzy sets is fully considered as one of structural models. Approaches to forecasting accuracy are considered. Forecast is performed by the example of fifteen monthly average exchange rate of the US dollar to the Russian Federation rubble (from October, 2011 till December, 2012). Bell-shaped membership function, Mamdani’s min fuzzy implication and centre of gravity for singletons defuzzification methods are used. Described forecasting procedure resulted in Mean Absolute Percentage Error (MAPE) of 2,43 percent. To reduce MAPE, and thus, improve forecasting accuracy, additional researches of other fuzzification, implication and defuzzification methods capabilities have to be performed.
Key words: валютный курс, прогнозирование, временной ряд, нечеткая логика, нечеткие множества, лингвистическая переменная, функция принадлежности, нечеткая импликация Мамдани, точность прогнозирования, средняя ошибка аппроксимации, currency exchange rate, forecastin