Vimal wrote: > Hi, > > Can someone explain or give some urls on 'difference' between > expectation-maximization and entropy maximization. > > To me both seems to maximize the E(log(p(x))) where p(x) is the pdf, > although both originate from different theories. > > Thanks, > > Cheers, > Vimal
I think what you say is not quite right... ... let p(x) be the true pdf and q(x) a candidate for what might be the true pdf. Then "expectation-maximization" seeks to maximise E(log(q(x))) (where E derives from p(x), usually when p(x) is unknown but information derives from a data sample, but p(x) is treated as fixed, while q(x) is variable). ... and, "entropy maximization" seeks to maximise E(log(p(x)) (where E derives from p(x), and both instances of p(x) are the same and variable). David Jones . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
