It has been over three months since my last post, due to working away from home for some of the summer, a summer holiday and moving home. However, during this time I have continued with my online reading and some new thinking about my conditional restricted boltzmann machine based trading system has developed, namely the use of a nonlinear autoregressive exogenous model in the bottom layer gaussian units of the CRBM. Some links to reading on the same are shown below.
- https://www.mathworks.com/help/nnet/ug/design-time-series-narx-feedback-neural-networks.html
- http://deeplearning.cs.cmu.edu/pdfs/Narx.pdf
- http://www.wseas.us/e-library/transactions/research/2008/27-464.pdf
- https://rucore.libraries.rutgers.edu/rutgers-lib/24889/
- https://arxiv.org/abs/1607.02093
- https://hal.archives-ouvertes.fr/hal-00501643/document
- http://www.theijes.com/papers/v3-i11/Version-1/C0311019026.pdf
- http://www.sciencedirect.com/science/article/pii/S0925231208003081
The exogenous time series I am thinking of using, at least for the major forex pairs and perhaps precious metals, oil and US treasuries, is a currency strength indicator based on the US dollar. In order to create the currency strength indicator I will have to delve into some data wrangling with the historical forex data I have, and this will be the subject of my next post.