Početna » FINIZ 2020 - People in the focus of process automation » Papers from the Thematic Areas of the Conference
ECHO STATE NETWORKS USAGE FOR STOCK PRICE PREDICTIONS
Abstract:
The subject of research is the consideration of potential use of Echo State Networks (ESN) for prediction of stock prices. The reasons are: 1) stock prices show non stationary behaviour and ESN is well suited to time series prediction of chaotic systems; 2) implementation in optical domain can bring very fast inference and financial market transactions requires prompt brought decisions; 3) training time of ESN is short and doesn’t require special hardware like training deep Feedforward Neural Network (NNs). Although there is only a few studies about using ESN for prediction financial data (pricing of securities subject of trade, and their volumes), it can be concluded than ESNs have great potential to be used in prediction of the stock prices. Having in mind the intention of the Serbian government to stimulate issuing of bonds, ESN can be applied on Belgrade Stock Exchange.
CITATION:
IEEE format
L. Barjaktarović, M. Barjaktarović, S. Konjikušić, “Echo State Networks Usage for Stock Price Predictions,” in FINIZ 2020 - People in the focus of process automation, Belgrade, Singidunum University, Serbia, 2020, pp. 97-102. doi: 10.15308/finiz-2020-97-102
APA format
Barjaktarović, L., Barjaktarović, M., Konjikušić, S. (2020). Echo State Networks Usage for Stock Price Predictions. Paper presented at FINIZ 2020 - People in the focus of process automation. doi:10.15308/finiz-2020-97-102
BibTeX format
Download
RefWorks Tagged format
Download
