Fork me on GitHub

Hi, I have some data coming from a trading simulation, so I have time values and some Y ranging between 10000 and 15000. I was curious and tried some linear regression around it, but, to no avail. Knowing almost nothing about this field I wonder if there are models that could provide some prediction or what else could be done with such data? Any hints, clues or readings are appreciated.


Where I mapped time values to the range of my Y values


consider reading some machine learning books or taking a ML course, like this one


time-series prediction is one of ML topics, there are many methods for doing that


@be9 I already started doing so, I watched half of the openclassroom stuff from stanford and read through the first two chapters of machine learning for clojure. The thing is, I want to solve that specific problem (if possible at all) and for this, I have to know what exactly to look for. Time-series prediction sounds reasonable, thanks for the hint simple_smile


e.g. Holt-Winters method. there are also ways to use neural networks (even complex-valued!) and support vector machines for forecasting. not to mention deep learning which is said to be able to solve anything 😄


@be9 thank you very much simple_smile