Loading...
Publication
Inteligência artificial aplicada na previsão de mercados financeiros
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 6.13 MB | Adobe PDF |
Authors
Abstract(s)
Neste documento, são investigados vários métodos usados na inteligência artificial, com o
objetivo de obter previsões precisas da evolução dos mercados financeiros. O uso de
ferramentas lineares como os modelos AR, MA, ARMA e GARCH têm muitas limitações,
pois torna-se muito difícil adaptá-los às não linearidades dos fenómenos que ocorrem nos
mercados. Pelas razões anteriormente referidas, os algoritmos como as redes neuronais
dinâmicas (TDNN, NARX e ESN), mostram uma maior capacidade de adaptação a estas
não linearidades, pois não fazem qualquer pressuposto sobre as distribuições de
probabilidade que caracterizam estes mercados. O facto destas redes neuronais serem
dinâmicas, faz com que estas exibam um desempenho superior em relação às redes
neuronais estáticas, ou outros algoritmos que não possuem qualquer tipo de memória.
Apesar das vantagens reveladas pelas redes neuronais, estas são um sistema do tipo black
box, o que torna muito difícil extrair informação dos pesos da rede. Isto significa que estes
algoritmos devem ser usados com precaução, pois podem tornar-se instáveis.
In this document, several methods used in the field of artificial intelligence are investigated, with the objective of obtaining more precise forecasts of the financial markets. The use of linear tools such the models AR, MA, ARMA and GARCH has many limitations, because it becomes very difficult to adapt them to the non linear phenomena that occur in the markets. For the reasons mentioned above, algorithms like dynamic neural networks (TDNN, NARX and ESN), show a better adaptation to these non linearities, because they don't make any assumptions of the probability distributions of these markets. The fact that these neural networks are dynamic, leads to a superior performance in relation to the static neural nets, or other algorithms that don't possess any type of memory. In spite of these advantages, neural networks are a black box type of system, making very difficult for the developer to extract information from the weights of a neural net. This means that these algorithms should be used with caution because they suddenly might become unstable.
In this document, several methods used in the field of artificial intelligence are investigated, with the objective of obtaining more precise forecasts of the financial markets. The use of linear tools such the models AR, MA, ARMA and GARCH has many limitations, because it becomes very difficult to adapt them to the non linear phenomena that occur in the markets. For the reasons mentioned above, algorithms like dynamic neural networks (TDNN, NARX and ESN), show a better adaptation to these non linearities, because they don't make any assumptions of the probability distributions of these markets. The fact that these neural networks are dynamic, leads to a superior performance in relation to the static neural nets, or other algorithms that don't possess any type of memory. In spite of these advantages, neural networks are a black box type of system, making very difficult for the developer to extract information from the weights of a neural net. This means that these algorithms should be used with caution because they suddenly might become unstable.
Description
Keywords
AR MA ARMA GARCH TDNN NARX ESN Inteligência artificial Redes neuronais Black box Pesos Artificial intelligence Neural networks Weights