Time series processing and forecasting using soft computing tools

  • Authors:

  • Nadezhda Yarushkina;Irina Perfilieva;Tatiana Afanasieva;Andrew Igonin;Anton Romanov;Valeria Shishkina

  • Affiliations:

  • Ulyanovsk State Technical University, Ulyanovsk, Russia;Institute for Research and Applications of Fuzzy Modeling, University of Ostrava, Ostrava 1, Czech Republic;Ulyanovsk State Technical University, Ulyanovsk, Russia;Ulyanovsk State Technical University, Ulyanovsk, Russia;Ulyanovsk State Technical University, Ulyanovsk, Russia;Ulyanovsk State Technical University, Ulyanovsk, Russia

  • Venue:
  • RSFDGrC'11 Proceedings of the 13th international conference on Rough sets, fuzzy sets, data mining and granular computing
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

The aim of this contribution is to show that the combination of F-transform, fuzzy relations, neural networks and genetic algorithms can be successfully used in analysis and forecast of short time series encountered in financial analysis of a small enterprize. We propose to represent a time series trend by the direct F-transform components and to model it by one of three different models that are based on a linear autoregressive equation, neural network or fuzzy relation autoregressive equation. An optimal model of the trend will be chosen by a genetic algorithm. In comparison with other time series techniques the proposed one is simple and effective in computation and forecast. In the application part, we present a description of a new software system that has been elaborated on the basis of the proposed theory. It includes analysis of time series and their tendencies in linguistic terms.