摘要：Straipsnyje nagrinėjamos finansinio kapitalo sektoriaus veiklos analizės ir prognozės problemos ir siūlomi efektyvūs jų sprendimo būdai. Parodoma, kad tradicinė finansinio kapitalo rinkų kitimo analizė, besiremianti tiesine paradigma, jau nėra efektyvi kriziniais ir pereinamaisiais (nepusiausviriais) laikotarpiais, kai vyrauja chaotiškoms sistemoms būdingi neperiodiniai svyravimai. Nagrinėjama, kaip šių problemų optimaliam sprendimui siūloma pritaikyti dirbtinio intelekto sistemų metodus, kurie šiuolaikinių informacinių technologijų dėka išplečia tradicinės analizės galimybes. Pasiūlyta bendra tyrimų schema, susidedanti į nuoseklių ir metodiškai pagrįstų tyrimo etapų pradedant tyrimo objekto dinamikos analize ir baigiant jo elgesį prognozuojančių modelių sudarymu ir tyrimu.
其他摘要：A review of financial capital markets, presented in the article, helped to reveal the extended use of modem information technologies and artificial intelligence methods in today’s financial markets. Artificial intelligence systems (AIS) embrace artificial neural networks, chaos theory, fractal theory, fuzzy logic, and genetic algorithm’s methods. Those theories and methods are well suited for the modeling of non linear dynamics and are capable to overtake other techniques in short term forecasting, trend prognosis, recognition of structural shifts, nonlinear correlations, and chaotic behavior. AIS are capable to coup with the modem financial markets problems, which more resemble adaptive, chaotic and evolutionary then static or equilibrium nature.The authors have stressed on the description of drawbacks of the traditional capital investment theories, formulation of theoretical and practical premises for the nonlinear approaches using modem information technologies like distributed databases, world-wide communication channels, parallel processing, and OLAP systems. Extended review of a related literature, helped to create the overall research scheme, based on AIS methods, which has mutually bounded and consistent research stages. The overall research scheme embraces: 1) research and description, using chaos and fractal methods, of the nonlinearities and their dynamics in financial time series; 2) creation of non linear complex models for SE indices approximation and prognosis using artificial neural network’s methods; 3) application of the different AIS methods like a consistent analytical tool for the bank clients crediting risk decision support model creation. The scheme made possible systematic review of various analytical possibilities for the different AIS methods. It also helped to create compound models for the analysis and prognosis of the dynamics in the financial markets.
