Journals →  Tsvetnye Metally →  2009 →  #5 →  Back

ArticleName Creation of share indexes for forecasting securities of «Norilsk nickel» mining and metallurgical corporation
ArticleAuthor Butyrin A. A.
ArticleAuthorData Butyrin A. A., e-mail: nii@norvuz.ru.
Abstract Methods of reception of correct neural networks (NN) entrances adapted for forecasting of securities of Open Society «Norilsk Nickel» mining and metallurgical corporation are considered. The described a way of formation of entrance images for training the NN, named «a method of windows». The opportunity of use of known share indexes is examined and quality of use of the new index entered by the author for forecasting a portfolio of securities is analyzed. For increase of accuracy of the correlation analysis, it has been decided to lead it in separate subgroups. As criterion of splitting into groups the factor of liquidity has been chosen. At training in NN, it is not necessary to be limited to any one index as it not in forces to reflect a condition of the market simultaneously in all sections. Expedience to choose set of the indexes that are most suitable for forecasting the chosen securities and to make the analysis of divergences indexes.
keywords valuable securities, forecast, neuron network, initial image, method of windows, stock index, representativeness, liquidity, correlation analysis.
References 1. Бутырин А. А. // Материалы Всерос. науч.-практ. Интернет-конф. «Информационные системы и технологии в социально-экономических и правовых процессах» / под ред. Е. Н. Шиянова, В. В. Неверовой, Ж. В. Игнатенко. — Ставрополь : СКСИ, 2007. С. 12–17.
2. Маргевич А., Волков А. // Рынок ценных бумаг. 2007. № 21. С. 75–79.
3. Бельзецкий А. // Банкирский вестник. 2006. С. 19–26.
Language of full-text russian
Full content Buy
Back