- Gruul side against Jund decks
- 2 players. Can I play the draw 2 after a draw 4 if it was the last card to the previous player?
- What is contained in a backup.ab file?
- Allow login with multiple passwords
- Sizing Resistors for Voltage Divider Circuit Feeding Solid-State Relay
- Preferable pattern for differential length matching
- NIMH volt recovery
- Energies of the fields in a plane wave, attenuated and unattenuated
- Why doesn't it matter if a resistor is before or behind an LED wrt voltage drop?
- tracking error state space, linear control example
- Indicator LED directly to AC mains. Which method would you use?
- DC Motor - Ke and Kt are different…?
- Large drift (temperature could be a culprit) in this circuit. Anyone can help?
- Sending audio data in chunks using visible light communication
- Should client ideas about the UI turn into User Stories?
- How to use separated methods?
- Are “technical user stories” allowed in Scrum?
- Cyclomatic Complexity Question
- Rear derailleur speed number interchangability
- mismatched chain and chainring
GMM preprocessing steps
I am trying to classify MNIST data (28x28x60000) using a GMM. I am looking for hints on preprocessing steps, what I've tried so far:
PCA and run GMM on first 10 principal components, looking for 10 GMM components: very poor performance
ICA and run GMM, slightly better performance than PCA but still not great
i've tried some manifold dimensionality reudction techniques like isomap and MDS but these don't work on such a large dataset, at least not on my pc.
In both cases I have tried initialising GMM weights with means equal to Kmeans cluster centers on the PCs.
bonus question: if using GMM with more than 10 components, is there a way to map these components back to the original 10 classes?