FileBase	PeterP-2013-10-10_dataset1.kwik
ElecNo	12
UseFeatures	
DropLastNFeatures	0
UseDistributional	1
MaskStarts	100
MinClusters	100
MaxClusters	110
MaxPossibleClusters	500
nStarts	1
StartCluFile	
SplitEvery	100
SplitFirst	20
PenaltyK	0.000000
PenaltyKLogN	1.000000
Subset	1
FullStepEvery	10
MaxIter	10000
RandomSeed	654
Debug	0
SplitInfo	1
Verbose	1
DistDump	0
DistThresh	6.907755
ChangedThresh	0.050000
Log	1
Screen	1
PriorPoint	1
SaveSorted	0
SaveCovarianceMeans	0
UseMaskedInitialConditions	1
AssignToFirstClosestMask	1
RamLimitGB	0.000000
help	0
help	0
Setting RAM limit to total physical memory, 16.00 GB.
WARNING: Not all physical memory will be available, but on Macs it is not possible
         to get the available physical memory.
-------------------------------------------------------------------------
Using Distributional EM with Maskstarts
NOTE: Maskstarts overides above values of MinClusters and MaxClusters                 
MinClusters = 100 
MaxClusters = 100 
 This run is expected to use between 0.01 and 0.01 GB of RAM.
Memory request: 0.00 GB would take us to total of 0.00 GB.
----------------------------------------------------------
Loaded 6 data points of dimension 25.
MEMO: A lower score indicates a better clustering 
 Masked EM: Computing Noise Means and Variances 
 -----------------------------------------
Starting from 100 clusters...
------ Clustering full data set of 6 points ------
Starting initial clusters from distinct float masks 
 Not enough masks (3) to generate starting clusters (100), so starting with (4) clusters instead.
Assigned 4 initial classes from 3 unique masks.
Unmasked E-step: Deleting class 1 (4 points): covariance matrix is singular 
Unmasked E-step: Deleting class 2 (1 points): covariance matrix is singular 
Unmasked E-step: Deleting class 3 (1 points): covariance matrix is singular 
Iteration 0F (0.002202 sec): 1 clusters Score: Raw 0.000000 + Penalty 275.123485 = 275.123485 nChanged 2
Iteration 1F (0.002411 sec): 1 clusters Score: Raw 11.675461 + Penalty 275.123485 = 286.798946 nChanged 6
Iteration 2F (0.002606 sec): 1 clusters Score: Raw 0.000000 + Penalty 275.123485 = 275.123485 nChanged 0
Writing temp clu file 
Memory request: 0.00 GB would take us to total of 0.01 GB.
Compute initial score before splitting: Score: Raw 0.000000 + Penalty 275.123485 = 275.123485Time taken for this iteration:0.002350 seconds.
 100->1 Clusters: Score 275.123485, best is 100000000000000005366162204393472.000000
THE BEST YET!

Main iterations: 3 (time per iteration =1.445333 ms)
Total iterations: 3 (time per iteration =1.445333 ms)

Def. Iteration metric 2:
Iteration_metric2 += (scalar)(nDims*nDims)*(scalar)(nPoints)
Iterations metric 2: 11250.000000 (time per metric unit =385.422222ns)

Def. Iteration metric 3:
Iteration_metric3 += (scalar)(nDims*nDims)*(scalar)(nDims*nPoints)
Iterations metric 3: 281250.000000 (time per metric unit=15416.888889ps)

That took 0.004336 seconds.