@AmpBenzScientist
I don't understand: are you designing a new machine learning algorithm?
@AmpBenzScientist
Cool, I'd be interested in an overview of how you would go about designing a new methodology
@rastinza It depends on the task at hand. If one finds a task where no ML algorithm adequately fits, they need to do some analysis on the target.
They will eventually find some way they are all connected so some constant like form. This would be the base of the ML algorithm. The training data needs to be constructed to where it ramps up in differences so it can be trained to eventually go for classification or sorting on the target data.
This is how I understand it. The training data is going to be where accuracy comes from. I'm quite new to ML design and this is the first time I'm going to implement a ML algorithm. I could be wrong
@rastinza Well two of them but yes. There aren't any that I have found that can do what I need. It strays a little out of ML territory and potentially into other fields.