MODELS AND METHODS OF TECHNICAL ANALYSIS OF FINANCIAL MARKETS

Abstract

This paper is devoted to the study of existing models and methods of analysis and prediction of financial markets. Basic information about fundamental and technical analysis of financial markets and their main assumptions is provided. The main tasks and problems that arise in the process of analysis and forecasting of financial markets are highlighted. The relevance of the topic is ensured by the fact of significant increase in the number of financial instruments in stock and other financial markets, as well as the rapid computerization of the trading process in these markets. Analysis of existing models and methods used to solve problems such as: analysis of the current market situation, search for patterns and anomalies in the financial time series and forecast the future price of the asset is provided. Authors mainly focus on statistical models and forecasting methods, pattern recognition methods, machine learning models and methods, sentimental analysis models and hybrid models. Study on the results of such models and methods as long short-term memory, gated recurrent units, support vector machine, perceptually important points is provided. In particular, given results of research of models that are used both independently and as components of a hybrid model for technical analysis of various financial markets. Namely, an overview of the achievements in the application of these models for short- and long-term forecasting in the United States stock market and Korean stock market. It has been found that hybrid artificial neural networks, which are able to take into account the public mood of market players, are the most promising for short-term forecasting of the company's stock price in the stock market. Based on the study, feasibility of using statistical models in combination with methods of pattern recognition or machine learning.