def fitWindowing(self, samples, windowSize, wordLength, symbols, normMean,
lowerBounding):
self.transformation = MFT(windowSize, normMean, lowerBounding)
sa = {}
index = 0
for i in range(samples["Samples"]):
new_list = getDisjointSequences(samples[i], windowSize, normMean)
for j in range(len(new_list)):
sa[index] = new_list[j]
index += 1
sa["Samples"] = index
self.fitTransform(sa, wordLength, symbols, normMean)
def fitTransformDouble(self, samples, wordLength, symbols, normMean):
if self.initialized == False:
self.initialize(wordLength, symbols, normMean)
if self.transformation == None:
self.transformation = MFT(len(samples[0].data), normMean,
self.lowerBounding, self.MUSE_Bool)
transformedSamples = self.fillOrderline(samples, wordLength)
if self.HistogramType == "EQUI_DEPTH":
self.divideEquiDepthHistogram()
elif self.HistogramType == "EQUI_FREQUENCY":
self.divideEquiWidthHistogram()
elif self.HistogramType == "INFORMATION_GAIN":
self.divideHistogramInformationGain()
return transformedSamples
def fitWindowing(self, samples, Labels, windowSize, wordLength, symbols,
normMean, lowerBounding):
self.transformation = MFT(windowSize, normMean, lowerBounding)
labels = []
sa = []
for i in range(samples.shape[0]):
new_list = getDisjointSequences(samples.iloc[i, :].tolist(),
windowSize, normMean)
for j in range(len(new_list)):
sa.append(new_list[j])
labels.append(Labels[i])
SA = pd.DataFrame(np.zeros((len(sa), len(sa[0]))))
for r in range(len(sa)):
for c in range(len(sa[0])):
SA.iloc[r, c] = sa[r][c]
if self.SUP:
self.fitTransformSupervised(SA, labels, wordLength, symbols,
normMean)
else:
self.fitTransform(SA, labels, wordLength, symbols, normMean)