Methods for building models of long-term prediction of financial time series
AbstractMethods for prediction of the financial time series with external conditions are developed. An algorithm of step-by-step construction of linear regression equations with different combinations of repressors is offered. The linear optimization problem is applied to meet the external conditions. The algorithm is used for long-term prediction of the time series according to the requirements of Ukrainian banks on juristic persons credits in 2007.
Mathematical methods, models, problems and technologies for complex systems research