Abstract: This paper presents developed genetic-based algorithm for time series forecasting problem and describes approaches to learning procedures design. Different techniques of population representation, recombination, formation of niches, calculation of fitness, conflict resolution methods are proposed. Results of computational experiments with real time series are analyzed.

Keywords: forecasting, genetic-based machine learning, rule-based forecasting, genetic algorithm, time series forecasting, evolutionary algorithms.

ACM Classification Keywords: I.2.8 Problem Solving, Control Methods, and Search

Link:

DEVELOPMENT AND ANALYSIS OF GENETIC ALGORITHM FOR TIME SERIES FORECASTING PROBLEM

Leonid Hulianytskyi, Anna Pavlenko

http://www.foibg.com/ijima/vol04/ijima04-01-p02.pdf