MMSE profiles vs sex

QPC test 1

• 18 Features: basic stats from MMSE x 3 (avg, std, kr, sk, min, max)
• 2 classes (male, female)
• Normalization [-1,1]

>> w=qpctrain(x,sex,'plot','qpc','display','short')
procedure     = qpc
dataname      = data
vectors       = 43
features      = 18
OptMethod     = gd
learningRate  = 0.100000
prototypes    = 0
proto lr      = 0.000000
eps           = 0.001000
start init.   = 10
end init      = 3
multistart    = yes
maxIterations = 1000
qpc function  = @(wx,p)qpcfunction(x,ci,wx,[],func)
function      = @(x)f_gauss(x,parameters.beta)
beta          = 0.100000
Performing 10 initiations - Gradient based optimization (MultiStart)

Initialization  7 was the best
QPC      W
0.1407      -0.0610  -0.3720  -0.0549  0.0928  -0.3889  -0.1171  -0.2340  0.2090  -0.4526  -0.2194  0.0225  0.1967  -0.2217  -0.2602  -0.0588  0.0048  -0.4011  -0.0981

Initialization  1 was the best
QPC      W
0.1722      0.0116  -0.3085  -0.2951  0.1922  0.0427  -0.0071  -0.0002  0.2227  0.3176  -0.0118  -0.2655  -0.5617  -0.2525  -0.2163  -0.1833  0.0171  -0.1770  -0.2576
-----------------------------------------------

>> bgraph(x*w(1,:)',sex,'labels',{'F','M'})

>> scaterplot(x*w',sex,'labels',{'F','M'})

>> bar(w(1,:))

QPC test 2

w2=qpctrain(x,y)
  
Columns 1 through 13

-0.0249   -0.3698    0.1600    0.1945   -0.4369   -0.0992   -0.0704   -0.0648   -0.1939   -0.2658   -0.4138   -0.0841   -0.3087
 0.0943   -0.1567   -0.3461    0.0129   -0.0332   -0.0622   -0.0284   -0.2094    0.0890   -0.2581   -0.1387   -0.0162    0.1340

Columns 14 through 18

 -0.1321    0.2038   -0.3028   -0.1533    0.1825
 -0.3923   -0.4241    0.0668   -0.2750   -0.5207

scaterplot(x*w2',y,'labels',{'F','M'})

bar(w2(1,:));

QPC Proto

• Dane: jak wyżej

wp=qpctrain_proto(x,sex,'display','short','prototypes',10);
  
procedure     = proto
dataname      = data
vectors       = 43
features      = 18
OptMethod     = gd
learningRate  = 0.100000
prototypes    = 2
proto lr      = 0.100000
eps           = 0.001000
start init.   = 10
end init      = 3
multistart    = no
maxIterations = 1000
qpc function  = @(wx,p)qpcfunction_dt(x,ci,wx,p,func)
function      = @(x)f_gauss(x,parameters.beta)
beta          = 0.100000
---------------Sumary--------------------------
Initialization  1 was the best
N     QPC      W
0  0.3684      0.1588  -0.2728  -0.2091  0.0229  -0.2330  -0.4201  -0.2694  0.2895  -0.3135  0.2957  -0.2428  0.1705  0.1866  -0.0747  0.0811  0.1172  -0.2520  -0.2665
-----------------------------------------------
procedure     = proto
dataname      = data
vectors       = 43
features      = 18
OptMethod     = gd
learningRate  = 0.100000
prototypes    = 2
proto lr      = 0.100000
eps           = 0.001000
start init.   = 10
end init      = 3
multistart    = no
maxIterations = 1000
qpc function  = @(wx,p)qpcfunction_dt(x,ci,wx,p,func)
function      = @(x)f_gauss(x,parameters.beta)
beta          = 0.100000
---------------Sumary--------------------------
Initialization  8 was the best
N     QPC      W
0  0.3485      -0.1843  -0.3654  0.1899  0.1058  -0.3556  0.1592  -0.0685  0.1587  -0.2727  -0.4180  0.3167  0.0428  -0.2293  -0.0986  0.1989  -0.0137  -0.3471  0.1729
-----------------------------------------------

 scaterplot(x*wp',sex,'labels',{'F','M'})

 bar(wp(1,:))