pca_augmentation.py 文件源码

def pca_augmentation(data_h5, excel_path):

    f = h5py.File(data_h5, 'r')

    df1 = pd.read_excel(excel_path, sheetname=0, header=1)
    df1 = df1.rename(columns=lambda x: x.strip()).set_index('Image')  # strip whitespace

    df2 = pd.read_excel(excel_path, sheetname=1, header=1)
    df2 = df2.rename(columns=lambda x: x.strip()).set_index('Image')  # strip whitespace
    df = pd.concat([df1, df2])

    X = preprocess_data(f)
    X_mean = np.mean(X, axis=0)
    X = X - X_mean

    # PCA
    pca = PCA().fit(X)