RegMetric(
source_col,
target_col,
function_name,
function,
normalization_file="external_infos/normalization_info.yml",
)
Bases: DfMetric
Regression metric for dataframes
This class let's you calculate arbitrary
regression metrics. It could be integrated
as DfMetric in the DataframeAnalyzer(ResultAnalyzer).
Parameters
source_col: str
column name in the dataframe where the ground truth is stored.
target_col: str
column name where the predictions are stored
function: str or dict
function to calculate the metric.
If str is given it must be a key of the metric dict.
Otherwise the given dict is initialized with init_object
and additional the key name
must be available to define the
output name in the result dict.
function_name: str
Name of the function
normalization_file: str, default "normalization_info.yml"
This is a image_location relative to the experiment
output folder. If a key exists with the same
name as source_col the metric result is
normalized and additionaly stored.
Source code in niceml/mlcomponents/resultanalyzers/dataframes/regmetric.py
| def __init__( # pylint: disable = too-many-arguments
self,
source_col: str,
target_col: str,
function_name: str,
function: Union[str, Callable],
normalization_file: str = "external_infos/normalization_info.yml",
):
"""
This class let's you calculate arbitrary
regression metrics. It could be integrated
as DfMetric in the DataframeAnalyzer(ResultAnalyzer).
Parameters
----------
source_col: str
column name in the dataframe where the ground truth is stored.
target_col: str
column name where the predictions are stored
function: str or dict
function to calculate the metric.
If str is given it must be a key of the metric dict.
Otherwise the given dict is initialized with `init_object`
and additional the key `name`
must be available to define the
output name in the result dict.
function_name: str
Name of the function
normalization_file: str, default "normalization_info.yml"
This is a image_location relative to the experiment
output folder. If a key exists with the same
name as `source_col` the metric result is
normalized and additionaly stored.
"""
self.source_col = source_col
self.target_col = target_col
self.normalization_file: str = normalization_file
self.func_name = function_name
if isinstance(function, str):
try:
self.function = metric_dict[function]
except KeyError as error:
raise Exception(
f"Function with name {function}"
f" not supported! Available: {list(metric_dict.keys())}"
) from error
else:
self.function = function
|
Functions