Any "method" will take some form of context into account, but that context is usually static. Meaning that the relationship to a prior day high or a Fibonacci level (as an example) is a 1 or a 0. The method will do what it is told to do or display what it is told to display under the condition that a pre-contextualized condition has been met.
But what if the condition has still been met, but the market structure has changed? And now the market does the opposite every time it sees those pre-defined levels? The method would then fail because the structure of the market has changed. Systems or Methods fade over time because unless programmed to adapt, their context is nearly always baked in.
Machine Learning can adapt and handle this stuff just like discretionary trading. Output values can be used to train deep learning neural networks, but this is Black Mirror stuff that I dont know much about.
Context is a difficult topic to discuss because it is so individual. In my opinion, market structure and context is organic and needs to be handled in a fluid manner.
If you are a profitable trader then (even if you dont specifically know it) it appears you are probably reading the context of the market correctly.
If you are not profitable but you are reading the markets actions correctly then I would suggest your context is sound and your execution is a problem.
If you are neither profitable nor reading the markets actions/context correctly, there is more work to do.