I’m still Hello Worlding my way towards submitting my first model but I must know! Can I make my model produce visual representations of data so it can learn from pictures of graphs too? (To be clear, I have no idea what I’m talking about but I want my model to stare at itself using a camera)
That seems to be a very inefficient way to develop your model. Any visual representation of the data could just be intuited from the data itself.
For example, you could have a camera examine an image of a cluster of data points and have a neural net try to figure out the line of best fit (would take very long). Or you could just run OLS in a couple seconds.