def runClassification(K, dynamic_datasets_path):
print('Start KNN Classification..')
dataSets = readDatasets.read_dataset(True, False, True,
dynamic_datasets_path)
trainSet = dataSets[0]
testSet = dataSets[1]
makeListsOfNeighborsForAllTests = True
plotPatterns = False # We just want the KNN, not the html-maps.
maxWarpingWindowPercentage = 0.33 # For testSet_a2, we need a bigger window to get the right patternIDs.
### Run KNN for test_a2
neighborsTestsLists = findKnearestNeighbors(
K, maxWarpingWindowPercentage, plotPatterns,
makeListsOfNeighborsForAllTests, trainSet, testSet)
testData = getVotes(neighborsTestsLists)
write_predictions_to_csv(testData)
### Run cross-validation.
crossValidation(trainSet, K, maxWarpingWindowPercentage, num_folds=10)
def data_visualization(K, dynamic_datasets_path):
print("Going to visualize bus-trajectories..")
dataSets = readDatasets.read_dataset(True, False, False,
dynamic_datasets_path)
trainSet = dataSets[0]
#print trainSet.shape[0] # DEBUG!
#print trainSet['Trajectory'] # DEBUG!
journeyPatternIDs, trainTrajs, trainListSize = TrainData.getListsOfTrainData(
trainSet)
storeMapsDir = os.path.join('..', '..', 'Resources', 'maps', 'task1')
if not os.path.isdir(storeMapsDir):
os.makedirs(storeMapsDir)
selectedPatternIDs = []
numOfSelectedPatterns = 0
start_time = time.time()
maxSecondsToWait = 120
while True:
if numOfSelectedPatterns == K:
print('Finished plotting ' + K.__str__() +
' distinct random patterns.')
break
elif (time.time() - start_time) > maxSecondsToWait:
print('The program could not find ' + K.__str__() +
' distinct random patterns in the specified time: ' +
maxSecondsToWait.__str__())
break
randomTrain = random.randint(0, trainListSize - 1)
curPatternID = journeyPatternIDs[randomTrain]
if curPatternID not in selectedPatternIDs:
selectedPatternIDs.append(curPatternID)
# plot the new pattern
print('Going to plot a new random train..')
longitudes, latitudes = GetCoordinates.getCoordinates(
trainTrajs[randomTrain])
fileName = "train" + randomTrain.__str__(
) + "_Pattern_" + curPatternID + ".html"
GmPlot.gmPlot(latitudes, longitudes,
os.path.join(storeMapsDir, fileName))
numOfSelectedPatterns += 1
def runA1andA2(K, dynamic_datasets_path):
dataSets = readDatasets.read_dataset(True, True, False,
dynamic_datasets_path)
trainSet = dataSets[0]
testSet = dataSets[1]
maxWarpingWindowPercentage = 0.11
makeListOfAllNeighbors = False
plotPatterns = True
task2A1.findKnearestNeighbors(K, maxWarpingWindowPercentage, plotPatterns,
makeListOfAllNeighbors, trainSet, testSet)
# This doesn't take any dataSets as parameters.. it gets them on its own
task2A2.runLCSS(K, dynamic_datasets_path)
def data_visualization(K):
dataSets = readDatasets.read_dataset(True, False, False)
trainSet = dataSets[0]
#print trainSet.shape[0] # DEBUG!
#print trainSet['Trajectory'] # DEBUG!
journeyPatternIDs, trainTrajs, trainListSize = TrainData.getListsOfTrainData(
trainSet)
storeMapsDir = "../../Resources/maps/task1"
if not os.path.isdir(storeMapsDir):
os.makedirs(storeMapsDir)
selectedPatternIDs = []
numOfSelectedPatterns = 0
start_time = time.time()
maxSecondsToWait = 120
while True:
if numOfSelectedPatterns == K:
print 'Finished plotting ' + K.__str__(
) + ' distinct random patterns.'
break
elif (time.time() - start_time) > maxSecondsToWait:
print 'The program could not find ' + K.__str__(
) + ' distinct random patterns in the specified time: ' + maxSecondsToWait.__str__(
)
break
randomTrain = random.randint(0, trainListSize - 1)
curPatternID = journeyPatternIDs[randomTrain]
if curPatternID not in selectedPatternIDs:
selectedPatternIDs.append(curPatternID)
# plot the new pattern
print 'Going to plot a new random train..'
longtitutes, latitudes = GetCoordinates.getCoordinates(
trainTrajs[randomTrain])
GmPlot.gmPlot(
latitudes, longtitutes, storeMapsDir + "/train" +
randomTrain.__str__() + "_Pattern_" + curPatternID + ".html")
numOfSelectedPatterns += 1
def runLCSStest():
# get two lists
# f = open("lists.txt")
# contents = f.read().split("\n")
# list1 = [int(i) for i in contents[0].split(",")]
# list2 = [int(i) for i in contents[1].split(",")]
dataSets = readDatasets.read_dataset(True, False, True)
list1 = dataSets[0]
list2 = dataSets[1]
lists = lcs(list1, list2, True)
# here we have all most common subs
# so now we pic the top 5 and plot them....
for l in lists:
print(l)
def runLCSS(K, dynamic_datasets_path):
print('\nLCSS start..')
dataSets = readDatasets.read_dataset(True, False, True,
dynamic_datasets_path)
trainSet = dataSets[0]
testSetA2 = dataSets[1]
journeyPatternIDs, trainTrajs, trainListSize = TrainData.getListsOfTrainData(
trainSet)
storeMapsDir = os.path.join('..', '..', 'Resources', 'maps', 'task2A2')
if not os.path.isdir(storeMapsDir):
os.makedirs(storeMapsDir)
kMaxs = KMaxs(K)
start_time = time.time()
lastTime = start_time # For in-the-middle elapsed-time.
testNum = 0
for trajectoryTest in testSetA2['Trajectory']:
# print trajectoryTest
testNum += 1
# if testNum <= 2:
# continue
# if testNum == 2:
# break
nearestNeighbors = []
sorted_subSeqsSizes = []
print('\nChecking for ' + K.__str__() +
' longest common sub-sequences of test ' + testNum.__str__())
# print trajectoryTest # DEBUG!
iterations = 0
sorted_subSequences = []
for i in range(
0,
trainListSize): # IDs and Trajectories are of the same size.
# print i
trajectoryTrain = trainTrajs[i]
LongestCS = lcs(trajectoryTrain, trajectoryTest)
if not LongestCS:
continue
# else:
# print i.__str__() + ') LCS length: ' + len(LongestCS).__str__()
# if iterations == 10:
# break
# else:
# iterations += 1
kMaxs.checkMinLengthAndInsert([i, LongestCS, len(LongestCS)])
curTime = time.time()
curElapsedTime = curTime - lastTime
lastTime = curTime
print('\nTest: ' + testNum.__str__() + ') finished in ' +
time.strftime("%H:%M:%S", time.gmtime(curElapsedTime)))
# Plot test
full_longitudes, full_latitudes = GetCoordinates.getCoordinates(
trajectoryTest)
fileName = "lcss" + testNum.__str__(
) + "-test-Time(sec)_" + curElapsedTime.__str__() + ".html"
GmPlot.gmPlot(full_latitudes,
full_longitudes,
os.path.join(storeMapsDir, fileName),
zoom=13)
# So now we pic the top 5 and we plot them....
sorted_subSequences = sorted(kMaxs.getArrayList(),
reverse=True,
key=lambda tup: tup[2])
kMaxs.resetArrayList(
) # Reset arrayList before going to the next testSet.
for i in range(0, len(sorted_subSequences)):
if i == 5: break
print("Train " + sorted_subSequences[i][0].__str__() + ") PatternID: " \
+ journeyPatternIDs[sorted_subSequences[i][0]].__str__() \
+ ", MatchingPoints: " + sorted_subSequences[i][2].__str__() + ".html")
curSubSeqTrajectory = trainTrajs[sorted_subSequences[i][0]]
full_longitudes, full_latitudes = GetCoordinates.getCoordinates(
curSubSeqTrajectory)
curSubSeqTrajectory = sorted_subSequences[i][1]
sub_longitudes, sub_latitudes = GetCoordinates.getCoordinates(
curSubSeqTrajectory)
fileName = "lcss" + testNum.__str__() + "-train" \
+ sorted_subSequences[i][0].__str__() + "_PatternID_" \
+ journeyPatternIDs[sorted_subSequences[i][0]].__str__() \
+ "-MatchingPoints_" + sorted_subSequences[i][2].__str__() + ".html"
GmPlot.gmPlotOfColours(full_latitudes, full_longitudes,
sub_latitudes, sub_longitudes,
os.path.join(storeMapsDir, fileName))
print("\nElapsed time of KNNwithLCSS for 'test_set_a2': ",
time.strftime("%H:%M:%S", time.gmtime(time.time() - start_time)),
'mins')
storeMapsDir + "/dtw" + testNum.__str__() + "-train" +
sorted_nearestNeighbors[i][0].__str__() + "-PatternID_" +
sorted_nearestNeighbors[i][1].__str__() + "-DTW_" +
sorted_nearestNeighbors[i][2].__str__() + ".html",
zoom=13)
if makeListOfAllNeighbors:
# Make a list with all the neighbours for all the tests
sorted_nearestNeighbors_forAllTests.append(
sorted_nearestNeighbors_forTest)
print("\nElapsed time of KNNwithDTW for 'test_set_a1': ",
time.strftime("%H:%M:%S", time.gmtime(time.time() - start_time)))
return sorted_nearestNeighbors_forAllTests
if __name__ == '__main__':
dataSets = readDatasets.read_dataset(True, True, False)
trainSet = dataSets[0]
testSetA1 = dataSets[1]
K = 5
plotPatterns = True
makeListOfAllNeighbors = False
maxWarpingWindowPercentage = 0.11
findKnearestNeighbors(K, maxWarpingWindowPercentage, plotPatterns,
makeListOfAllNeighbors, trainSet, testSetA1)
os.path.join(storeMapsDir, fileName),
zoom=13)
if makeListOfAllNeighbors:
# Make a list with all the neighbours for all the tests
sorted_nearestNeighbors_forAllTests.append(
sorted_nearestNeighbors_forTest)
print("\nElapsed time of KNNwithDTW for 'test_set_a1': ",
time.strftime("%H:%M:%S", time.gmtime(time.time() - start_time)))
return sorted_nearestNeighbors_forAllTests
if __name__ == '__main__':
dynamic_datasets_path = os.path.join('..', '..')
dataSets = readDatasets.read_dataset(True, True, False,
dynamic_datasets_path)
trainSet = dataSets[0]
testSetA1 = dataSets[1]
K = 5
plotPatterns = True
makeListOfAllNeighbors = False
maxWarpingWindowPercentage = 0.11
findKnearestNeighbors(K, maxWarpingWindowPercentage, plotPatterns,
makeListOfAllNeighbors, trainSet, testSetA1)
exit()