ONNX Trader - MetaTrader 5 Expert | MT5 EA Download - MetaTrader 5 Resources

Archived articles for models. You can train your own model as the models can be very different depending on the training hyperparameters.

The machine learning model is trained using Python and converted to ONNX format. You do not need to change the parameters and code in encludnik.

Setup is simple:

 Enter Double Maximum Risk = 0.001 ;    //Progressive lot coefficient
Enter double manual lot = 0 ; //Fixed lot size
Enter Oolong Order Magic= Chapter 666 ; //Order Magic
Enter the integer maximum order number = 1 ; //Order number
Enter integer stop loss = 2000 ; //stop loss
Enter integer take profit = 2000 ; // take profit
input string comment = "ONNX EA" ;

But you can add your own filters to try and make the model trade better.

At least you can optimize your stop loss and take profit on any time frame

The model is connected via header files, for example:

 #include

"ONNX Trader Clusters" robot, ONNX model and include files. Clustering is done by k-means method.

The training period is from 2010 to 2020. Future training will take place from 2020 to 2024.

This “Onnx Trader ” bot supports the second training method in the article “Using clustering to match trades ”

To do this, concatenate the ininclude file , the file will pull out the corresponding models "catmodel10.onnx" and "catmodel_m10.onnx"

Models are trained on EURUSD H1 but can be used on any time frame!

This script calculates autocorrelation and partial autocorrelation functions and displays them on a chart

Tillson T3 performs EMA calculation without auxiliary indicator buffer.

This library is an updated version of the ErrorDescription.mqh library published by MetaQuotes and contains some functions.

Use keyboard data

Attachment download

📎eurusd_onnx_include0.mqh (0.59 KB)

📎 catmodel0.onnx (496.69 KB)

📎 catmodel_m0.onnx (522.47 KB)

📎 eurusd_onnx_include_propensity_without_matching.mqh (0.64 KB)

📎 catmodel_propensity_without_matching.onnx (285.31 KB)

📎 catmodel_m_propensity_without_matching.onnx (326.56 KB)

📎 eurusd_onnx_include_propensity_matching_naive.mqh (0.64 KB)

📎 catmodel_propensity_matching_naive.onnx (249.23 KB)

📎 catmodel_m_propensity_matching_naive.onnx (522.47 KB)

📎 eurusd_onnx_include_propensity_matching_original.mqh (0.64 KB)

📎 catmodel_propensity_matching_original.onnx (522.47 KB)

📎 catmodel_m_propensity_matching_original.onnx (249.22 KB)

📎 eurusd_onnx_include_propensity_matching_original_ipw.mqh (0.61 KB)

📎 catmodel_propensity_matching_original_ipw.onnx (522.46 KB)

📎 catmodel_m_propensity_matching_original_ipw.onnx (290.47 KB)

📎eurusd_onnx_include_clusters.mqh (0.97 KB)

📎 catmodelclusters.onnx (1018.09 KB)

📎catmodel_mclusters.onnx (22.37 KB)

📎 catmodel10.onnx (159 KB)

📎 catmodel_m10.onnx (522.47 KB)

📎eurusd_onnx_include10.mqh (0.59 KB)

📎 onnx_trader.mq5 (8.66 KB)

📎 onnx_trader_clusters.mq5 (9.06 KB)

Source: MQL5 #48482