KernelPCA

import math as m
import sys
import numpy as n
from numpy import *
import matplotlib.pyplot as p
import graph as g

d=[[0 for x in xrange(50)] for x in xrange(2)]
for x in range(50):
    d[0][x]=x

i=1
while(i<51):
    yy2=m.log(i)
    d[1][i-1]=yy2
    i=i+1
d2=[[0 for x in xrange(50)] for x in xrange(2)]
for x in range(50):
    d2[0][x]=x


i=1
while(i<51):
    yy2=m.log(i)-1
    d2[1][i-1]=yy2
    i=i+1
p.plot(d[0][:],d[1][:],'x',d2[0][:],d2[1][:],'o')
p.title('input data')
p.show()
data=hstack((d,d2))


parameter_list = [[data,0.01,1.0], [data,0.05,2.0]]

def preprocessor_kernelpca_modular (data, threshold, width):
    from shogun.Features import RealFeatures
    from shogun.Preprocessor import KernelPCA
    from shogun.Kernel import GaussianKernel
    features = RealFeatures(data)

    kernel=GaussianKernel(features,features,width)

    preprocessor=KernelPCA(kernel)

    preprocessor.init(features)
    preprocessor.set_target_dim(2)
    preprocessor.apply_to_feature_matrix(features)
    X=preprocessor.get_transformation_matrix()
    X2=preprocessor.apply_to_feature_matrix(features)
#   print 'X=',X2[2][50:100]
    print 'apply to feature matrix=%d, ',n.shape(X2)
    p.plot(X2,'o')
    p.title('apply_to_feature_matrix')
    p.show()
    #sys.exit(0)
    print 'type of features=%',(type(X))
    print 'le X=\n',len(X[0])
    print 'size of the original data is= ,and the returned matrix is',n.shape(data),n.shape(X2)
    lx0=len(X2)
    lx1=len(X2[0])
    modified_d1=[[0 for x in xrange(len(d[0]))] for x in xrange(lx0)]
    modified_d2=[[0 for x in xrange(int(len(d2[0])))] for x in xrange(lx0)]


    #print 'X2[:][len(d[0]):]',X2
    for i in range(lx0):
        for j in range(len(d[0])):
            modified_d1[i][j]=X2[i][j]
    for i in range(lx0):
        for j in range(lx1-len(d[0])):
            modified_d2[i][j]=X2[i][j+len(d[0])]
    p.plot(modified_d1,'o',modified_d2,'x')
    p.title('final data')
    p.show()

    return features



if __name__=='__main__':
    print('KernelPCA')
    preprocessor_kernelpca_modular(*parameter_list[0])