Terminator3
Advanced Member level 3
- Joined
- Feb 25, 2012
- Messages
- 802
- Helped
- 71
- Reputation
- 142
- Reaction score
- 63
- Trophy points
- 1,308
- Activity points
- 9,027
I just wondering. Yesterday i made a test: using simple function y=k*x + random(30)-random(30).
Compared moving average with embedding in 8-dimensional space (and more). That method with moving N-dimensional vector to nearest N-dimensional neighbor. Results are interesting, but pretty useless, as moving average gives better result. Does it mean, that i need multiple measurements of same function + noise to get better results than moving average? If anyone know some useful material on this topic please provide authors or paper name i can search for.
Compared moving average with embedding in 8-dimensional space (and more). That method with moving N-dimensional vector to nearest N-dimensional neighbor. Results are interesting, but pretty useless, as moving average gives better result. Does it mean, that i need multiple measurements of same function + noise to get better results than moving average? If anyone know some useful material on this topic please provide authors or paper name i can search for.