def getStandardedMatData(rawData):
# 将数据标准化后，再分割为数据集、标签集，然后将格式转换为矩阵
dataSet, labelSet = LoadFile.splitData(rawData)
dataSet = np.mat(dataSet)
labelSet = np.mat(labelSet).T
dataSet = standardize(dataSet)
return dataSet, labelSet
def kFoldCrossValidation(rawData, processFunc, k):
rawData = pd.DataFrame(rawData)
n = np.shape(rawData)[0]
step = n // k
bestModel = 0
leastErr = 0
for i in range(k):
if i == (k-1):
testData = rawData.iloc[step*i: n, :]
trainData = rawData.iloc[: step*i, :]
else:
testData = rawData.iloc[step*i: step*(i+1), :]
trainData = rawData.iloc[: step*i, :]
trainData = trainData.append(rawData.iloc[step*(i+1):])
testData, testLabel = LoadFile.splitData(testData)
trainData, trainLabel = LoadFile.splitData(trainData)
model = getModel(trainData, trainLabel)
err = testModel(model, testData, testLabel)
if err < leastErr:
leastErr = err
bestModel = model
return model
return math.sqrt(z)
def minmaxNormalize(dataSet):
# 数据映射到[0, 1]
dataSet = pd.DataFrame(dataSet)
minDf = dataSet.min()
maxDf = dataSet.max()
normalizedSet = (dataSet - minDf) / (maxDf - minDf)
return normalizedSet
def zscoreStanderize(dataSet):
# 用z-score方法进行数据标准化
dataSet = pd.DataFrame(dataSet)
meanDf = dataSet.mean()
stdDf = dataSet.std()
standerizedSet = (dataSet - meanDf) / stdDf
return standerizedSet
if __name__ == "__main__":
filePath = "F:/2020AI_SummerCamp/dataSet/"
# rawData = LoadFile.loadCSV(filePath + "Pima.csv")
rawData = LoadFile.loadCSV(filePath + "diabetesN.csv")
dataSet, labelSet = LoadFile.splitData(rawData)
# dataSet = solveMissingData(dataSet)
# dataSet = minmaxNormalize(dataSet)
dataSet = zscoreStanderize(dataSet)
print(dataSet)