A Novel Approach for Time Series Forecasting with Multiobjective Clonal Selection Optimization and Modeling

摘要

In this paper a novel approach for time series forecasting with multiobjective clonal selection optimization and modeling has been considered. At first, the main principals of the forecasting models (FM) on the base of the strictly binary trees (SBT) and the modified clonal selection algorithm (MCSA) have been discussed. Herewith, it is suggested, that the principles of the FMs on the base of the SBT can be applied to creation the multi-factor FMs, if we are aware of the presence of the several interrelated time series (TS). It will allow increasing the forecasting accuracy of the main factor (the forecasting TS) on the base of the additional information on the auxiliary factors (the auxiliary TS). Then, it is offered to develop the multiobjective MCSA (MMCSA) on the base of the notion of the "Pareto dominance", and use the affinity indicator (AI) based on the average forecasting error rate (AFER), and the tendencies discrepancy indicator (TDI) in the role of the objective functions in this algorithm. It will allow to improve the results of the solution of a problem of the short-term forecasting and to receive the adequate results of the middle-term forecasting. This MMCSA can be applied for solving problems of individual and group forecasting. Also, the application of the principles of the attractors' forming on the base of the long TSs to creation of the training data sequence (TDS) with the adequate length for the FM on the base of the SBT has been discussed. aBesides, the possibilities of the FMs on the base of the SBT and the MMCSA in the problem of the TS restoration with aim of the fractal dimension definition have been discussed. It is offered to carry out restoration of the TS elements' values as for the timepoints in the past as for the timepoints in the future simultaneously, using two FMs of the middle-term forecasting. The experimental results which confirm the efficiency of the offered novel approach for time series forecasting with multiobjective clonal selection optimization and modeling have been given.