Innovative methods for improving portfolio management based on artificial intelligence instruments

摘要

Financial markets represent one of the most complex environments for business and there are a lot of types of external factors which impact their dynamics. The recent financial turbulence materialized by the global financial crisis 2008-2009 and the European sovereign debt crisis (2010-2012) made serious pressure on financial markets that proved their fragility and sensitivity in a different manner. The use of different instruments used on artificial intelligence could be applied in decision making process in financial markets because they offer a unique capability of learning. The conventional theories regarding the anticipation of financial markets evolution are represented by the efficient market hypothesis (Fama, 1970) and the paradigm regarding the methods to anticipate the future performance of financial assets. The actual interest is to identify optimal strategies for portfolio management by using artificial intelligence. The basic steps of incorporating different types of artificial intelligences on the study of the future dynamic of the performance of different financial assets are the following: the analysis of the strategies used by different portfolio managers and their performances; the identification of new instruments capable to improve the strategy references; the selection/ development and testing of the new instrument; the analysis of the differential performance. Actual artificial intelligence instruments are difficult to create/develop and to use because in this paper will be presented a new concept in which the basis will be the application data transformation in order to build different sets of training artificial neural networks in order to optimize/modify in an easy way their behavior. This module for simulating the artificial neural network is improved by using genetic algorithms to select the best network regarding the predictions of the performance of financial instruments, but also the optimal timing in the process of portfolio management.