digit = None;

X_train, X_test = [], [];

folderPath = r"data\digits\trainingDigits";

for fileName in os.listdir(folderPath):

with open(folderPath + "\\" + fileName, "rt") as file:

digit = list(np.array([[int(d) for d in line if d != "\r" and d != "\n"] for line in file.readlines()]).flatten());

digit.append(int(fileName[0]));

X_train.append(digit);

folderPath = r"data\digits\testDigits";

for fileName in os.listdir(folderPath):

with open(folderPath + "\\" + fileName, "rt") as file:

digit = list(np.array([[int(d) for d in line if d != "\r" and d != "\n"] for line in file.readlines()]).flatten());

digit.append(int(fileName[0]));

X_test.append(digit);

X_train = np.mat(X_train);

X_test = np.mat(X_test);

knn = KNN.KNN(X_train, X_train.shape[1] - 1,

distanceCalculator = lambda X, v: 1 - DataHelper.calcJaccardCoefficient(X, v),

isNormalizeFeatures = False);

errorCount = 0;

for row in X_test:

if knn.getMostFrequentlyClass(knn.findKNN(row, 3)[0]) != row[0, -1]:

errorCount += 1;

print(errorCount);

print(errorCount * 100 / X_test.shape[0]);