High Frequency Time Series Forecasting.

      Аннотация: Time series are a special form of data where past values in the series may influence future values, depending on the presence of underlying deterministic forces. These forces may be characterised by trends, cycles and nonstationary behaviour in the time series and predictive models attempt to recognise the recurring patterns and more particularly potential linear or nonlinear relationships between past and actual values, or with other exogenous variables which may be linked to the variable studied. Time series forecasting is the use of a model to forecast future time series values based on known past events: to predict data points before they are measured. Forecasting is an important and recurrent issue in business world since good forecasting models can lead to a major position in the market. Indeed a firm can anticipate the temporal evolution of a given data in order to implement solutions before its competitors. Forecasting problems find their applications in many fields: for example sales in marketing, production volume in operations and logistics, economic variable like GDP in macroeconomic studies or financial variables like stock prices in finance.

Ключевые слова: ARFIMA, ARIMA, artificial neural networks, Exponential Smoothing, non-linearity, star, stock price forecasting, high frequency time series, ARMA-GARCH, intraday data, one-step-ahead prediction, forecasting performances