Deep Modeling Complex Couplings within Financial Markets

Authors: Wei Cao, Liang Hu, Longbing Cao

AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results on data of stock and currency markets from three countries show that our approach outperforms other baselines, from both technical and business perspectives.
Researcher Affiliation Academia Wei Cao Advanced Analytics Institute University of Technology, Sydney Wei.Cao@student.uts.edu.au Liang Hu University of Technology, Sydney and Shanghai Jiaotong University lianghu@sjtu.edu.cn Longbing Cao Advanced Analytics Institute University of Technology, Sydney Long Bing.Cao@uts.edu.au
Pseudocode No The paper describes mathematical equations and inference steps but does not provide a clearly labeled pseudocode or algorithm block.
Open Source Code No The paper does not provide any specific statement or link indicating the availability of open-source code for the described methodology.
Open Datasets Yes The data set used includes weekly closing prices from Jan 2007 to Dec 2013 2, and the prices are decoded into returns by RIt = P It P It 1 P It 1 100%, here RIt and PIt are, respectively, the return and closing price at time t. As indexes in different markets may appear on different trading days, we delete those days on which some market data is missing and only choose the days with data from all financial markets. 2http://research.stlouisfed.org/
Dataset Splits Yes The testing data includes the financial crisis period (2007-2009) and a non-crisis period (2010-2013) (Here we split the data by years, and we use the last five years data as the training set before the testing year so as to learn the model parameters).
Hardware Specification No The paper does not provide specific details about the hardware used for running the experiments.
Software Dependencies No The paper does not specify software dependencies with version numbers.
Experiment Setup Yes This is our deep learning approach, where the order n of the both CGRBM and CCRBM are set equal to 2 which yields good results in this experiment.