pattern recognition, correlation recognition, rule violation

h_song

Hello everybody!

I am a newby here so same, forgive me if my questions are too easy!
but i really couldn't find my answers in the forum. nor in the internet in general.

I will be working on very large database, so I'm currently testing the software with simple usecases.

1. My database shows many points of the Sinus curve in function of the time.
How can I make rapidminer recognize that is is sinus function here?
I've tried nearly every modeling and learning tools, but none of their output show this. And applied to a simple time database table, the model won't give sinus as output either..

2. My second database has also the maximum function. Its values are 0 when sin is negative, and 1 when positive.
I want the software to know this correlation.
But the correlation matrix only gives values of coefficients, etc.

3. My third database shows the Sinus values once again, but this time with 3anomalies which are random(time) instead of sinu(time). which is a rule violation.
I tried the similarity measure with a normal sinus database among other methods...

...nothing..!

so you see....i would more than appreciate some tip!
thanks^^