Using Copman Theory in Machine Learning?

There's a lot of systems where you can find very good, low dimension acupent in badis and that's kind of what we should be going after with these kind of approaches. For example, energy forecasting or atmospheric pollution forecasting. These are some of the things we explord in our paper. I'm not actually sure about this, i heard that you can't have a linear system that has multipled racpers yxat cande velina system will to attractivs. So if your noneinamica system has two attractives, then copman theory is going to be difficult to be employed. Kind of, you'd also mention that language was one place maybe you