Method of Neuronet Classification of Situations in Financial Markets and Its Implementation in IDSS

Abstract

The task of developing tools for analyzing financial markets based on modern methods of artificial intelligence, imitating the work of market traders, who need to analyze a large amount of heterogeneous information about the market for decision-making, is becoming increasingly relevant at the moment. The paper deals with the application of a neural network classifier as the core of an intellectual decision support system for a trader in financial markets. The main characteristics of systems such as an intelligent decision support system are identified. The technique of neuronet classification of market situations, using various approaches to the analysis of financial markets, such as technical, fundamental analysis and econometric modeling, is presented. The basis of the methodology is the integrated information analysis based on the neural network topology multilayer perceptron. The results of system design using the algorithm of neural network classification of market situations, the architecture of the system, the study of efficiency and efficiency of the intellectual decision support system for traders in the financial markets based on the described methodology of classifications of market situations are presented. According to the results of the experimental study, the efficiency of the system in 83% is indicated on the sample, which is normalized by the parameters of the experiment.