DSPRelated.com
Forums

Predictors

Started by pawl March 28, 2008
Hello,	
  could somebody write me a list of predictors used for data compression 
in transmission and maybe some books where I can read about them? I 
could find only a little about Linear Predictor (forward and backward).

Thanks

pawl wrote:
> Hello, � > � could somebody write me a list of predictors used for data compression > in transmission and maybe some books where I can read about them? I > could find only a little about Linear Predictor (forward and backward).
There are many algorithms to compute linear prediction coefficients (if that is what you mean). Here is a list of the better known methods: - autocorrelation method (aka Yule-Walker method) - covariance method & modified covariance method - Burg method (aka Maximum Entropy method) There are many more, they usually differ in the domain (lag, time or frequency) and the error that is minimized. There is an intorductory section in Proakis&Manolakis "DSP" on linear prediction. Regards, Andor
Hello,

> There are many algorithms to compute linear prediction coefficients > (if that is what you mean). Here is a list of the better known > methods: > > - autocorrelation method (aka Yule-Walker method) > - covariance method & modified covariance method > - Burg method (aka Maximum Entropy method) >
Thanks for that. But are there any other predictors? Or is there only the linear one? I'm asking, because I do not know anything about it. I am studying Power Electronics and Control. In two weeks I will have an exam and there will be a professor from Data Transmission Institute (I am not sure if it is correct english name). He told me that he will ask me about predictors in data transmission (for compression purposes mostly). The only one thing which I found and which was called predictor is the linear predictor. Of course just like you told it's weights can be calculated in many different ways. But I think he meant that there are other predictors, not only linear. Do you know any names of them? Thanks.