osbad.stats¶

Module Contents¶

osbad.stats.ArrayLike¶

Type alias for array-like inputs.

Represents data structures that can be treated as arrays in numerical and analytical operations. This alias is used for type hints to accept both pandas Series and NumPy ndarray objects.

class osbad.stats.OutlierMethodConfig¶

Immutable configuration for a statistical outlier detector.

Stores:

- compute: the detector implementation accepting
  (X_variable, **stats_params_dict)
- params: default statistical parameters

compute: Callable[[Any], Tuple]¶

params: Dict[str, Any]¶

osbad.stats.outlier_method: Dict[str, OutlierMethodConfig]¶

Dictionary mapping outlier-detector identifiers to their configs.

Identifiers:

“sd”: Standard Deviation
“mad”: Median Absolute Deviation
“iqr”: Interquartile range
“zscore”: Z-score
“mod_zscore”: Modified Z-score

Example

# (1): Anomaly detection with standard deviation
# Access the dict of parameters
sd_param_dict = bstats.outlier_method["sd"].params

# Predict the anomalous cycle using standard dev method
# and the corresponding stats parameters
(SD_outlier_dV_index,
 SD_min_limit_dV,
 SD_max_limit_dV) = bstats.outlier_method["sd"].compute(
    df_max_dV["max_diff"],
    **sd_param_dict)

# (2): Anomaly detection with MAD
mad_param_dict = bstats.outlier_method["mad"].params

(MAD_outlier_index_dV,
 MAD_min_limit_dV,
 MAD_max_limit_dV) = bstats.outlier_method["mad"].compute(
    df_max_dV["max_diff"],
    **mad_param_dict)

# To update the statistical parameters or threshold
# Create a copy of the default dict parameter
mad_param_dict_const = bstats.outlier_method["mad"].params.copy()

# Update the dict value to be 1.4826
mad_param_dict_const["mad_factor"] = 1.4826

# Use the updated param_dict in the outlier method
(MAD_outlier_index_dV_const,
 MAD_min_limit_dV_const,
 MAD_max_limit_dV_const) = bstats.outlier_method["mad"].compute(
    df_max_dV["max_diff"],
    **mad_param_dict_const)

osbad.stats.calculate_zscore(df_variable: pandas.Series | numpy.ndarray) → pandas.Series | numpy.ndarray¶

Calculate the Z-score of the selected feature.

Parameters:: df_variable (pd.Series | np.ndarray) – Selected feature.
Returns:: Z-score of selected feature.
Return type:: pd.Series|np.ndarray

osbad.stats.calculate_feature_stats(df_variable: ArrayLike, new_col_name: str = None) → pandas.DataFrame¶

Calculate descriptive statistics for a given feature.

This function computes the mean, minimum, maximum, and standard deviation of the input variable. The results are returned as a pandas DataFrame, optionally labeled with a custom column name.

Parameters:

df_variable (pd.Series | np.ndarray) – Input data series or array for which statistics are calculated.
new_col_name (str, optional) – Optional name for the resulting column in the output DataFrame. Defaults to None.

Returns:

DataFrame with statistics (max, min, mean, std) as rows. If new_col_name is provided, the statistics are stored under that column name.

Return type:

pd.DataFrame