handle_missing.py 文件源码

def generate_test_data():
    with open('./test.csv', 'r') as test_file:
        test_csv = csv.reader(test_file, delimiter=',')
        next(test_csv)
        test_data = list(test_csv)
    test_data = numpy.array(test_data)
    # delete id column
    # test_data = numpy.delete(test_data, 0, 1)
    # One of K encoding of categorical data
    encoder = preprocessing.LabelEncoder()
    for j in (1, 2, 3, 4, 5, 6, 7, 8, 9, 14):
        test_data[:, j+1] = encoder.fit_transform(test_data[:, j+1])
    # Converting numpy strings to floats
    test_data = test_data.astype(numpy.float)
    missValueIndex = 7
    Xy_test = test_data[test_data[:, 3+1]==missValueIndex]
    Xy_train = test_data[test_data[:, 3+1]!=missValueIndex]
    X_train = numpy.delete(Xy_train, 3+1 ,1)
    y_train = Xy_train[:, 3+1]
    X_test = numpy.delete(Xy_test, 3+1 ,1)
    market_test_data = MarketingData(X_train, y_train, X_test)
    return market_test_data, test_data


# use knn for impute missing values