def analyses_classes(X, noOfTrainData, noOfExamples):
train_T_indices = [x for x in range(0, noOfTrainData, noOfExamples)]
for i in range(0, len(train_T_indices)):
print('i = {}'.format(i))
if i == 0:
this_class=X[0:train_T_indices[1]]
else:
this_class=X[train_T_indices[i - 1]:train_T_indices[i]]
code.min_hamming_distance(this_class)
distances = all_hamming_distances(this_class)
print('Mean Hamming distance for class {} is {}'.format(i,np.mean(distances)))
print('Std of Hamming distance for class {} is {}'.format(i, np.std(sum(this_class))))
print('Mean no. of activiations per class {}'.format(np.mean(sum(this_class))))
print('Example vector weight per class {}'.format(sum(this_class[0])))
return
n = n + 1
# at the moment, we are not doing validation:
XTrain = X
TTrain = T
allInputData = X
#############################################################
# get some input stats
#############################################################
if verbose == True:
# temp=code.min_hamming_distance(T)
# print('Input code: This is a [{0}, {1}, {2}] code'.format(noOfInputs, sizeOfInput, temp))
# print("min Hamming distance: %i" % temp)
# weights = code.weight(T, verbose=True)
temp = code.min_hamming_distance(t)
print('Output code: This is a [{0}, {1}, {2}] code'.format(
noOfOutputs, sizeOfOutput, temp))
print("Outputmin Hamming distance: %i" % temp)
weights = code.weight(t, verbose=True)
temp2 = code.min_hamming_distance(XTrain)
print('Input code: This is a [{0}, {1}, {2}] code'.format(
noOfTrainData, lenOfInput, temp2))
print("Input min Hamming distance: %i" % temp2)
# N.B. for the analysis scripts
np.save("allInputDataCurrent", X)
np.save("allOutputDataCurrent", T)
# N.B. we are just using this to get intel on the input
dataset = np.append(X, T, axis=1)