How to do multivariate time series classification using C# and either Accord.NET, Encog or ML.NET?

I have a time series based on financial security prices with additional features. I wish to feed this series into some ML construct in order to perform multi-class classification.

Most of the solutions that I found in my search offer predictions. I am not interested in predicting future prices. I merely wish to train the ML construct to offer the most likely class for the given time-series input frame.

I am looking for C# solutions or links to tutorials that use either the Accord.NET, Encog or ML.NET libraries. I would be most appreciative for answers that lead my eyes to view C# code that demonstrates a solution to my question.

In lieu of the above, I would also appreciate a description of the types of ML constructs that would satisfy my requirement.

I have no interest in Python solutions. Please, do not chastise me or praise Python. I need the code to be in C# so that it easily integrates into existing code.

Thank you.