#numVectors = len(lsaData.pz)
numVectors = 2
lsaVector.append(lsaData.pwz[:, 3])
lsaVector.append(lsaData.pwz[:, 4])
#Stip down to the optimal number }
#for i in range (0, numVectors):
# lsaVector.append(lsaData.pwz[:, index[i]])
#lsaVector.append(lsaData.pwz[:, i])
projections = []
for v in lsaVector:
projections.append([])
for d in dVector:
projections[-1].append(analysis.lsaProjection(d, v))
pr = []
for s in range(len(projections[0])):
tmp = []
for r in range(len(projections)):
tmp.append(projections[r][s])
pr.append(tmp)
pointsPlot = [[] for i in range(2)]
numtosplit = 20
dim = 2
for s in range(len(times)):
spot = int(s / numtosplit)
pointsPlot[spot].append((pr[s][0:dim], str(times[s])))
#if len(tmpDoc) >= minBehaviour:
dVector.append(tmpDoc)
times.append(oldSplit)
oldSplit = newSplit
#Load plsa data
lsaData = dataio.loadData(lsaLocation)
#Take counts data and lsa data to make graph.
index = [ i for (i,j) in sorted(enumerate(lsaData.pz), key=operator.itemgetter(1))]
#numVectors = len(lsaData.pz)
numVectors = 2
lsaVector.append(lsaData.pwz[:, 0])
lsaVector.append(lsaData.pwz[:, 3])
projections = []
for v in lsaVector:
projections.append([])
for d in dVector:
projections[-1].append(analysis.lsaProjection(d, v))
x = range(splitLength, numSplits + splitLength)
x *= skipLength
visualizer.plotLines(x, projections)
print tmpDoc
lsaVector = []
index = [
i for (
i,
j) in sorted(enumerate(lsaData.pz), key=operator.itemgetter(1))
]
#Stip down to the optimal number
for i in range(numVectors):
lsaVector.append(lsaData.pwz[:, index[i]])
ranges = [[(0, 1)], [(0, 1)]]
for t in range(len(lsaVector)):
#Now project this document to each latent class and store the results
val = analysis.lsaProjection(tmpDoc, lsaVector[t])
yProj[t].append(val)
ct += slide
xProj.append(count)
count += 1
xProj = numpy.array(xProj)
visualizer.makeLatentTimeGraph(xProj, yProj, ranges)
print tmpDoc
lsaVector = []
index = [ i for (i,j) in sorted(enumerate(lsaData.pz), key=operator.itemgetter(1))]
#Stip down to the optimal number
for i in range (numVectors):
lsaVector.append(lsaData.pwz[:, index[i]])
ranges = [[(0, 1)], [(0, 1)]]
for t in range(len(lsaVector)):
#Now project this document to each latent class and store the results
val = analysis.lsaProjection(tmpDoc, lsaVector[t])
yProj[t].append(val)
ct += slide
xProj.append(count)
count += 1
xProj = numpy.array(xProj)
visualizer.makeLatentTimeGraph(xProj, yProj, ranges)