机器学习中常用的处理手段

#方式 1
X = (X - np.mean(X, axis=0)) / np.std(X, axis=0)
#方式 2 缩放到(0,1)
x = (x - np.min(x)) / np.max(x)
# 然后缩放到(a,b)
f(x) = (b-a) x + a
#方式 3
img(img < 0) = 0;
img(img > 255) = 255;
#方式 4
# Problem 1 - Implement Min-Max scaling for grayscale image data
def normalizegrayscale(imagedata):
    """
    Normalize the image data with Min-Max scaling to a range of [0.1, 0.9]
    :param imagedata: The image data to be normalized
    :return: Normalized image data
    """
    a = 0.1
    b = 0.9
    grayscalemin = 0
    grayscalemax = 255
    return a + ( ( (imagedata - grayscalemin)*(b - a) )/( grayscalemax - grayscale_min ) )

#方式1
def onehotencode(x):
    targets = np.array(x).reshape(-1)
    return np.eye(10)[targets]
#方式2
from sklearn.preprocessing import LabelBinarizer 
def onehotencode(x):
    labelbinarizer = LabelBinarizer() 
    labelbinarizer.fit(x)
    return label_binarizer.transform(x)

from sklearn.modelselection import StratifiedShuffleSplit
ss = StratifiedShuffleSplit(nsplits=1, testsize=0.2)
trainidx,validx = next(ss.split(codes, labels))
halfval = int(len(validx)/2)
validx,testidx = validx[:halfval],validx[halfval:]
trainx, trainy = codes[trainidx],labelsvecs[trainidx]
valx, valy = codes[validx],labelsvecs[validx]
testx, testy =  codes[testidx],labelsvecs[testidx]

def visualizeinput(img, ax):
    ax.imshow(img, cmap='gray')
    width, height = img.shape
    thresh = img.max()/2.5
    for x in range(width):
        for y in range(height):
            ax.annotate(str(round(img[x][y],2)), xy=(y,x),
                        horizontalalignment='center',
                        verticalalignment='center',
                        color='white' if img[x][y]<thresh else 'black')
fig = plt.figure(figsize = (12,12)) 
ax = fig.addsubplot(111)
visualizeinput(Xtrain[0], ax)   #X_train[0] 就是一张图片