- How precisely we know mass of ISS?
- Could electrodynamic tethers power long-term unmanned satellites around the Sun, Jupiter, or Saturn?
- How are gravitational losses on ascent related to horizontal velocity?
- What are the parameters of the new Iodine electrical rocket engine developed by RSC Energia?
- Does any site track the mass of the ISS?
- ISS' change in mass over time; might there be a fairly high cadence plot?
- ISS mass gain or loss due to meteorites and the solar wind?
- is there a way to lock everything so settings won't get changed between rendering sessions?
- How to activate an action after another action happens
- How can I export a character animation using constraints and IK in FBX Format?
- How do you access and set a specific bone animation fcurve in python ? e.g. set pinky02_L z rotation = 8 degrees
- Cisco 220 Switch fail over
- Reverse Engineering OLD DOS Programm
- what is COERCE_FLOAT in ida Hex-Rays' C++ pseudocode?
- Should I say “laundry poles on the veranda” or “laundry poles in the veranda”?
- “bigger” vs. “more big”
- meaning of the verb “model” in context
- “Bring them back to society” vs “bring them back in to society”
- I will ring you in the evening
- meeting on a weekly-basis
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?