# R to $R^{20}$ linear transformation for classification?

I try to do some type of classification.I have 2 timeseries signals to distinguish from eachother.

Here is the way i was told to do it: I take into account one of these signals and call it MySignal hereforth. I place a 20-sample consecuitive window on this signal and then these 20 values will be considered as elements of a vector of 20-Dimension(MyVector). In this way i do a mapping from R to $R^{20}$ so if MySignal has 200 points, we will have 10 MyVector vectors.Therefore in $R^{20}$ space we will have 10 points which creats a 20_window signal.

I repeat the same action for the second timeserie signal and plot both 20_window signals in $R^{20}$ space. then I feed a decision tree algorithm with these two 20_window signals. if these two timeseries signals are not distinguishable from each other in R space by a straight line, can they be distinguishable in $R^{20}$ by a straight line parallel to any of 20 axes?

Let me clarify what i'm doing. After creating these 20_sample signals, i feed them to a decision tree (WEKA) which creates a very small decision tree. I try to understand how this decision tree is created (meaning how these test functions in each node are created by this tree).

btw, is there any keyword that i can use to look for this type of Transformation from R to $R^{20}$ without entering to the concept of Decision trees or Classification I mean by mathematics?

Thanks

-
Do your texts employ eccentric random boldfacing of terms? It is pretty distracting... –  rschwieb Dec 10 '12 at 18:01
so you suggest not to use boldface in text? –  Michelle Dec 13 '12 at 11:56
Well some general rules of thumb would be to not use boldface for every instance of a particular word, and to not use boldface for mundane words like "mathematics". –  rschwieb Dec 13 '12 at 15:29