Discussing the article: "Gaussian Processes in Machine Learning (Part 2): Implementing and Testing a Classification Model in MQL5"

[MetaQuotes]

Check out the new article: Gaussian Processes in Machine Learning (Part 2): Implementing and Testing a Classification Model in MQL5.

In this section, we will look at the implementation of the key interfaces of the library of Gaussian processes in MQL5: IKernel, ILikelihood, and IInference. We will also demonstrate its operation on synthetic data and implement indicators for classification and regression, demonstrating its operation in online mode - with retraining of the model on each new bar.

In the previous article, we learned about the theoretical foundations of the Bayesian machine learning model — Gaussian Processes — and began creating a GP library in MQL5, describing two key classes: GaussianProcess and GPOptimizationObjective.

Here we will complete the library by taking a detailed look at the implementation of the key interfaces: IKernel, ILikelihood, and IInference. After this, we will test the library on synthetic data and write indicators for classification and regression, demonstrating its operation in online mode — with retraining the model on each new bar.

Author: Evgeniy Chernish

[cemal]

Hello,

Is it possible for you to give an example and show how to fit the variance of price series with your GP regressor with radial basis function like the attched picture instead of predicting the t+1 etc..?

[Evgeniy Chernish]

cemal #:
Hello,

Could you please provide an example to show how to adjust the variance of a price series using your GP regressor with a radial basis function, as shown in the attached image, rather than forecasting t+1 and so on?

Good afternoon!

Is this what you wanted?

It now displays the mean and variance of the training sample rather than a one-step-ahead forecast.

[cemal]

Thank you very much  very kind of you.

Can the  Gaussian Process Regression  models be trained online mode , allowing them to adapt in real time mean and variance of the data?

[Evgeniy Chernish]

cemal #:

Thank you very much, that’s very kind of you.

Is it possible to train regression models based on Gaussian processes in an online mode, so that they can adapt to the mean and variance of the data in real time?

Yes, of course; that is exactly what the GPRegressor.mq5 script does.