def processData():
dataLoader.loadTrainingSets(testIndex=-1,
fast=False,
sampleSize=150,
dt_type=config.dataset_type.mnist_hw)
dataLoader.loadTestSets(dt_type=config.dataset_type.mnist_hw)
data = {
'train': {
'img': shd.TrainingDataSet.imagesDataSet,
'lbl': shd.TrainingDataSet.labelsDataSet
},
'test': {
'img': shd.TestSetData.imagesDataSet,
'lbl': shd.TestSetData.labelsDataSet
}
}
shd.TrainingDataSet.imagesDataSet = None
shd.TrainingDataSet.labelsDataSet = None
shd.TestSetData.imagesDataSet = None
shd.TestSetData.labelsDataSet = None
return data
def loadData():
dataLoader.loadTrainingSets(testIndex=-1, fast=False, sampleSize=5000, dt_type=config.dataset_type.mnist_hw)
dataLoader.loadTestSets(dt_type=config.dataset_type.mnist_hw)
sharedData.KNN_Results_5.resultByIndex[testLabel] = rslArr
else:
rslArr[predictedLabel_3] += 1
def findNearestNeighbors(inputVector):
distanceDic = defaultdict(list)
for i in range(len(sharedData.TrainingDataSet.imagesDataSet)):
distance = mh.calculateDistance(inputVector, sharedData.TrainingDataSet.imagesDataSet[i])
distanceDic[distance] = sharedData.TrainingDataSet.labelsDataSet[i] # who care if distance for two or more is equal?
return distanceDic
dataLoader.loadTrainingSets(testIndex=-1, fast=False, sampleSize=100)
dataLoader.loadTestSets()
k_array = [1, 3, 5]
for i in tqdm(range(len(sharedData.TestSetData.imagesDataSet))):
testVector = sharedData.TestSetData.imagesDataSet[i]
testLabel = sharedData.TestSetData.labelsDataSet[i]
distanceDic = findNearestNeighbors(inputVector=testVector)
calculateResults(testLabel=testLabel, resultDic=distanceDic)
print('----1NN--------')
i =0
for key in sorted(sharedData.KNN_Results_1.resultByIndex):
def load_data():
dataLoader.loadTrainingSets(testIndex=-1, fast=True, sampleSize=100, dt_type=config.dataset_type.mnist_fashion)
dataLoader.loadTestSets(dt_type=config.dataset_type.mnist_fashion)