All about MQL5 Wizard : create robots without programming. - page 9

[Sergey Golubev]

MQL5 Wizard Techniques you should know (Part 60): Inference Learning (Wasserstein-VAE) with Moving Average and Stochastic Oscillator Patterns

In examining the patterns generated from pairing the MA and stochastic oscillator, we have looked to machine learning as a means of systemizing our approach. There are mainly three methods of training networks in machine learning and these are supervised-learning, reinforcement-learning, and inference. By taking the view that each of these learning methods can be used at different stages of model/ network development, we have made the case that a model can be enriched by incorporating all of them.

[Sergey Golubev]

MQL5 Wizard Techniques you should know (Part 61): Using Patterns of ADX and CCI with Supervised Learning

We continue our look at how indicator pairings that track different aspects of the markets can be paired with machine learning to build a trading system. For these next articles, we are looking at the pairing of the Average Directional Index (ADX) oscillator with the Commodity Channel Index (CCI). The ADX is a predominantly a trend confirmation indicator, while the CCI is a momentum indicator. We touched on these two properties when we were looking at the patterns for individual indicators in past articles like this one. To recap, though, trend confirmation measures how strong a given price trend is; with a strength pointing to suitability for entry. Momentum indicators on the other hand measure the rate of price change. The more rapidly price is changing in a given direction, the less likely one is to suffer adverse excursions.

[Sergey Golubev]

MQL5 Wizard Techniques you should know (Part 62): Using Patterns of ADX and CCI with Reinforcement-Learning TRPO

We continue our look at how technical indicators that track different parts of price action can be paired in machine learning. In the last piece, we saw how supervised learning in a Multi-Layer-Perceptron (MLP) lays the groundwork of forecasting price action. We refer to the inputs of the MLP as features and its forecast outputs as states. From the way we defined our features in the last article which was slightly different from our approach in pieces 57–60, we aimed at having a more continuos input vector as opposed to the discrete option we had used. The move towards continuous data and regression and away from discrete data and classification can perhaps best be argued if we look at our AI trends.

[Sergey Golubev]

MQL5 Wizard Techniques you should know (Part 63): Using Patterns of DeMarker and Envelope Channels

For this article, we are pairing a momentum oscillator with a support/resistance channel. This may seem like an odd pairing, considering that most indicator pairings typically involve a trend following indicator, however this route could be explored because: of a need to avoid lags from trend identification; or a focus on mean reversion plays; or a need for a simpler trade system; or the need to adapt to choppy or range-bound markets; or the need to exploit momentum divergences, etc.

We therefore pair the DeMarker a momentum oscillator with the Envelopes Channel a support/resistance tool. In doing so, we are going to look, as always, at the top 10 patterns from pairing these two while testing with the GBP USD pair.

[Sergey Golubev]

MQL5 Wizard Techniques you should know (Part 64): Using Patterns of DeMarker and Envelope Channels with the White-Noise Kernel

We follow up our last article that paired momentum based DeMarker indicator with the support/resistance Envelopes bands by examining how their signals could be harnessed in machine learning. We have had similar approaches on indicator pairings in recent articles, and readers seeking some introduction can look into those. Essentially, we implement the MQL5 indicators in the Python language, while using price data imported with the MetaTrader 5 Python module. This module allows you to log onto your broker’s server and retrieve price data and symbol information.

[Sergey Golubev]

MQL5 Wizard Techniques you should know (Part 65): Using Patterns of FrAMA and the Force Index

We continue our series, where we had last looked at the DeMarker and Envelopes channels, by considering the pairing of the Fractal Adaptive Moving Average (FrAMA) and the Force Index Oscillator. FrAMA being a moving average is a trend signalling indicator while the Force Index provides a check on volume to see if the trend has sustenance. We will consider the typical 10 patterns that can be generated from combining these two indicators, as we have in past articles. We are training or optimizing with EUR USD on the 4-hour time frame for the year 2023. Forward walks or testing are done with this symbol for the year 2024.

[Sergey Golubev]

MQL5 Wizard Techniques you should know (Part 66): Using Patterns of FrAMA and the Force Index with the Dot Product Kernel

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.

[Sergey Golubev]

MQL5 Wizard Techniques you should know (Part 67): Using Patterns of TRIX and the Williams Percent Range

We continue our series on studying signal patterns generated by pairing technical indicators. Last time, we looked at the fractal adaptive moving average when paired with the force index oscillator. For this article, we are looking at the Triple Exponential Moving Average Oscillator (TRIX) when paired with another oscillator, the Williams Percent Range (WPR). TRIX being a moving average oscillator is a trend signalling indicator, while the Williams Percent Range acts as a compliment on support and resistance levels.

[Sergey Golubev]

MQL5 Wizard Techniques you should know (Part 68): Using Patterns of TRIX and the Williams Percent Range with a Cosine Kernel Network

Of the ten signal-patterns, we examined in the last article, only 3 were able to forward walk. These patterns were generated from combining indicator signals of the TRIX, a trend indicator and the Williams Percent Range (WPR), a support/ resistance oscillator. The training/ optimizing of the Expert Advisor was restricted to just one year, 2023, with the forward walk being performed over the subsequent year, 2024. We were testing with CHF JPY on the 4-hour time frame.

In extending our patterns that forward walk with machine learning, we typically use Python because it codes and trains networks very efficiently. This is true even without a GPU. In past articles, we have been prefacing with Python implementations of the functions of patterns that were able to forward walk. For this article, we will touch on the indicator implementations in Python, but mostly dwell on the network setup that takes the indicator signals as inputs. It is a convolutional 1-Dim network that uses the cosine kernel in its designs.

[Sergey Golubev]

MQL5 Wizard Techniques you should know (Part 69): Using Patterns of SAR and the RVI

The Parabolic Stop-And-Reverse (SAR) indicator and the Relative Vigour Index (RVI) oscillator are our next pairing to consider within these series. This time, this duo brings together trend identification with momentum tracking. As we have in past articles, we consider 10 possible signal patterns that can be generated from combining these two indicators, and we have them tested in a wizard assembled Expert Advisor. With that, let's look at the definitions of these 2 indicators.