I'm using SAX (Symbolic Aggregate approXimation) on a time series data. There are some SAX's parameters which can be adjusted - word size, vocabulary size, etc.
So, I wanted to have a metric in order to check how "accurate" is each of the SAX approximations, to see which one to use for further analysis (data mining, for example). What metrics would you recommend for that?
$R^2$ (coefficient of determination) came to mind. However I'm not sure it's enough.