Разработка оболочки системы прогнозирования временных рядов с использованием эволюционных алгоритмов
Analyzing the reasons of financial crises in the book «The Black Swan» N.N. Taleb concludes that modern economic models badly describe reality for they are not able to forecast such crises in advance. We tried to present processes on stock exchange as two random processes one of which happens rather often (regular regime) and the other one - rather rare. Our answer is that if regular processes are correctly recognized with the probability a bit higher than 1/2, this allows to get positive average gain. We believe that this very phenomenon lies in the basis of unwillingness of people to expect crises permanently and to try recognizing them.
ФИНАНСОВЫЕ КРИЗИСЫ, биржа, пуассоновский процесс, financial crises, Stock exchange, Poisson processes
This paper the experimental research of dynamic characteristics self-similar scales of measurement financial time lines and quality check statistical, econometrics and intellectual methods of their analysis and forecasting is described. Research was carried out on 25 various financial time lines, including on lines of the prices of actions of the Russian and foreign companies, the prices for gold, oil, indexes of the MMVB, S&P, exchange rates, etc. The Analysis of these lines has confirmed presence of the common laws in change of structure of lines depending on scale.
In this article we describe a system allowing companies to organize an efficient inventory management with 40 suppliers of different products. The system consists of four modules, each of which can be improved: demand planning, inventory management, procurement planning and KPI reporting. Described system was implemented in a real company, specializing on perishable products totaling over 600 SKUs. The system helped the company to increase its turnover by 7% while keeping the same level of services.