Pattern recognition in the structure of price curves and prediction of their future development using counterpropagation neural networks

There is proposed the methodological approach to forecasting of financial indices changes, which based on the recognition of images at price curves structures. This task is solved in the article with usage of counterpropagation neural networks. The analysis of results of carried out experiments shown the high ability to identify of typical patterns in financial indices time series and precise of forecasting of theirs following changes.