index_col=0)
if w in [0, 1, 2, 3, 4, 5]:
studentList = assessment1A.index
elif w in [6, 7, 8, 9, 10, 11]:
studentList = assessment2A.index
temp = temp.loc[temp.index.isin(studentList)]
# temp = graphLearning.mapNewLabel(temp, reLabelIndex)
# temp = temp.drop(['Practice_0-Practice_0'],axis=1)
# if w == 1:
# temp = temp.drop([8])
transitionDataMatrixWeeks.append(temp)
transitionDataMatrixWeeks_directFollow_standardised = []
for w in range(0, 12):
transitionDataMatrixWeeks_directFollow_standardised.append(
dataProcessing.normaliseData(transitionDataMatrixWeeks[w].T))
transitionDataMatrixWeeks_directFollow_normalised = []
for w in range(0, 12):
transitionDataMatrixWeeks_directFollow_normalised.append(
dataProcessing.normaliseData(transitionDataMatrixWeeks[w],
'normalised'))
#transpose transition data matrix
for w in range(0, 12):
transitionDataMatrixWeeks[w] = transitionDataMatrixWeeks[w].T
transitionDataMatrixWeeks_directFollow_normalised[
w] = transitionDataMatrixWeeks_directFollow_normalised[w].T
transitionDataMatrixWeeks_directFollow_standardised[
w] = transitionDataMatrixWeeks_directFollow_standardised[w].T
# a = activityDataMatrixWeeks[11]
cmap = sns.cm.rocket_r
sns.heatmap(a.corr(), annot=True, center=0.8, yticklabels=True, xticklabels=True, cmap='coolwarm')
plt.title('Correlation matrix')
plt.show()
#transpose transition data matrix
for w in range(0,12):
transitionDataMatrixWeeks[w] = transitionDataMatrixWeeks[w].T
transitionDataMatrixWeeks_directFollow_normalised = []
for w in range(0,12):
transitionDataMatrixWeeks_directFollow_normalised.append(dataProcessing.normaliseData(transitionDataMatrixWeeks[w]))
# correlation processing
corrList = []
corrDistanceList = []
for w in range(0,12):
corrTemp = transitionDataMatrixWeeks[w].corr()
corrList.append(corrTemp)
corrDistance = (0.5*(1 - corrTemp)).apply(np.sqrt)
corrDistanceList.append(corrDistance)
# cmap = sns.cm.rocket_r
# sns.heatmap(corrList[3], annot=False, center=0.8, yticklabels=False, xticklabels=False, cmap='coolwarm')
# plt.title('Correlation heatmap week ' + '3')
temp = temp.loc[temp.index.isin(studentList)]
# temp = graphLearning.mapNewLabel(temp, reLabelIndex)
# temp = temp.drop(['Practice_0-Practice_0'],axis=1)
# if w == 1:
# temp = temp.drop([8])
transitionDataMatrixWeeks.append(temp)
for w in range(0,10):
transitionDataMatrixWeeks[w].index = transitionDataMatrixWeeks[w].index + ['-2020']
for w in range(0,10):
transitionDataMatrixWeeks[w] = graphLearning.mapNewLabel(transitionDataMatrixWeeks[w], reLabelIndex)
transitionDataMatrixWeeks_directFollow_standardised = []
for w in range(0,10):
transitionDataMatrixWeeks_directFollow_standardised.append(dataProcessing.normaliseData(transitionDataMatrixWeeks[w].T))
transitionDataMatrixWeeks_directFollow_normalised = []
for w in range(0,10):
transitionDataMatrixWeeks_directFollow_normalised.append(dataProcessing.normaliseData(transitionDataMatrixWeeks[w].T, 'normalised'))
# correlation processing
corrList = []
corrDistanceList = []
for w in range(0,10):
corrTemp = transitionDataMatrixWeeks_directFollow_standardised[w].corr()
corrList.append(corrTemp)
corrDistance = (0.5*(1 - corrTemp)).apply(np.sqrt)
corrDistanceList.append(corrDistance)
graph_all_weeks = []
sns.heatmap(corrDistanceList[3],
annot=False,
center=0,
yticklabels=False,
xticklabels=False,
cmap='coolwarm')
plt.title('Distance correlation heatmap week ' + '3')
plt.show()
#normalised data--------------------------------------------------------------
transitionDataMatrixWeeks_directFollow_normalised = []
for w in range(0, 12):
temp = transitionDataMatrixWeeks_directFollow[w][:-1]
transitionDataMatrixWeeks_directFollow_normalised.append(
dataProcessing.normaliseData(temp))
#correlation processing
corrList_dataNormalised = []
corrDistanceList_dataNormalised = []
for w in range(0, 12):
corrTemp = transitionDataMatrixWeeks_directFollow_normalised[w].corr()
corrList_dataNormalised.append(corrTemp)
corrDistance = (0.5 * (1 - corrTemp)).apply(np.sqrt)
corrDistanceList_dataNormalised.append(corrDistance)
w = 11
matrix = corrList_dataNormalised[w]
# denoised_matrix = libRMT.denoisedCorr(matrix, transitionDataMatrixWeeks_directFollow_normalised[w].shape[0], transitionDataMatrixWeeks_directFollow_normalised[w].shape[1])
