Over-emphasizing peaks in a GMM
8 visualizzazioni (ultimi 30 giorni)
Mostra commenti meno recenti
I have a GMM made from gmdistribution.fit which I want to modify the probability distribution so as to emphasize the high probabilities even more. For example, if I could square all the individual probabilities and renormalize, it would raise the peaks even higher and concentrate samples into a few peak areas. How is it best to accomplsh this task?
1 Commento
Umar
il 14 Lug 2024 alle 7:34
Hi Eugene,
Square all the individual probabilities obtained from the GMM and renormalize them to ensure they sum up to 1. By squaring the probabilities, you are effectively amplifying the differences between high and low probabilities, thereby emphasizing the peak areas. After squaring the probabilities, you need to renormalize them to maintain the total probability mass at 1. This step ensures that the adjusted probabilities accurately represent the relative likelihoods of different components in the GMM.
Risposte (0)
Vedere anche
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!