数据处理

As expected, it can be observed, that for both attributes the distributions of amount values are heavy tailed. In order to approach faster a potential global minimum scaling and normalization of numerical input values is good a practice. Therefore, we first log-scale both variables and second min-max normalize the scaled amounts to the interval [0, 1].

# select "DMBTR" vs. "WRBTR" attribute

numeric_attr_names = ['DMBTR', 'WRBTR']


# add a small epsilon to eliminate zero values from data for log scaling
numeric_attr = ori_dataset[numeric_attr_names] + 1e-7
numeric_attr = numeric_attr.apply(np.log)


# normalize all numeric attributes to the range [0,1]
ori_dataset_numeric_attr = (numeric_attr - numeric_attr.min()) / (numeric_attr.max() - numeric_attr.min())