Discussing the article: "MQL5 Wizard Techniques you should know (Part 66): Using Patterns of FrAMA and the Force Index with the Dot Product Kernel"
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Check out the new article: MQL5 Wizard Techniques you should know (Part 66): Using Patterns of FrAMA and the Force Index with the Dot Product Kernel.
The FrAMA Indicator and the Force Index Oscillator are trend and volume tools that could be paired when developing an Expert Advisor. We continue from our last article that introduced this pair by considering machine learning applicability to the pair. We are using a convolution neural network that uses the dot-product kernel in making forecasts with these indicators’ inputs. This is done in a custom signal class file that works with the MQL5 wizard to assemble an Expert Advisor.
From our last article, where we introduced the pair of these indicators as a source of entry signal patterns for an Expert Advisor, the forward walk results were not as promising. We provided a few reasons why this was and also caveated that the training and optimization we perform is for only 1 year and therefore for any pattern, it is imperative to test as extensively as possible on vasts amount of history. We follow up that piece as always by examining those patterns that were able to forward walk. This is with machine learning.
When applying machine learning algorithms in MQL5, OpenCL is always an option, however this often requires one to have GPU hardware. This is nice to have, but Python’s code library has become quite extensive, a lot of efficiencies can be reaped with just a CPU. That is what we are exploring in these article series and so for this piece, as we have done on some in the past, we code our neural networks in Python because the coding, and training in Python is very efficient.
Of our ten patterns that we optimized or trained in the last article, only 2 were able to walk. Pattern-6 and Pattern-9. We therefore test these further with a neural network, as we did in past particles, the difference being that we are using a convolution neural network aka CNN. This CNN will implement the dot product kernel. First, however, as always with Python implementations, we define the indicator functions first that we need to provide signals to our network.
Author: Stephen Njuki