tfma.types.ValueWithTDistribution

View source on GitHub

Represents the t-distribution value.

tfma.types.ValueWithTDistribution(
    sample_mean: float,
    sample_standard_deviation: Optional[float] = None,
    sample_degrees_of_freedom: Optional[int] = None,
    unsampled_value: Optional[float] = None
)

It includes sample_mean, sample_standard_deviation, sample_degrees_of_freedom. And also unsampled_value is also stored here to record the value calculated without bootstrapping. The sample_standarddeviation is calculated as: \sqrt{ \frac{1}{N-1} \sum{i=1}^{N}{(x_i - \bar{x})^2} }