attributes
qoa4ml.lang.attributes
¶
Classes¶
DataQualityEnum
¶
Attributes¶
ACCURACY = 'data_accuracy'
class-attribute
instance-attribute
¶
The ratio between correct and total data the service received (%).
The string value is namespaced (data_accuracy) so it cannot collide
with MLModelQualityEnum.ACCURACY when both feed the MetricNameEnum
union during Pydantic coercion.
COMPLETENESS = 'completeness'
class-attribute
instance-attribute
¶
The ratio between received and expected number of data attributes sent to the service.
User-populated only: no built-in eva_* function produces this key;
clients compute completeness against their own expected-schema and
report it via QoaClient.observe_metric.
DUPLICATE_RATIO = 'duplicate_ratio'
class-attribute
instance-attribute
¶
Ratio of duplicate entries in the data
ERROR_RATIOS = 'error_ratios'
class-attribute
instance-attribute
¶
Ratio of errors in the data
NONE_RATIO = 'none_ratio'
class-attribute
instance-attribute
¶
Ratio of none or empty entries in the data
NULL_CORRELATIONS = 'null_correlations'
class-attribute
instance-attribute
¶
Correlation between null or undefined entries in the data
NULL_COUNT = 'null_count'
class-attribute
instance-attribute
¶
Count of null or undefined entries in the data
TOTAL_DUPLICATE = 'total_duplicate'
class-attribute
instance-attribute
¶
Total number of duplicate entries in the data
TOTAL_ERRORS = 'total_errors'
class-attribute
instance-attribute
¶
Total number of errors as given by the user in the data
TOTAL_NONE = 'total_none'
class-attribute
instance-attribute
¶
Total number of none or empty entries in the data
TOTAL_VALID = 'total_valid'
class-attribute
instance-attribute
¶
Total number of valid entries in the data
MLModelQualityEnum
¶
Attributes¶
ACCURACY = 'model_accuracy'
class-attribute
instance-attribute
¶
Can be measured in different ways such as confidence score in classification models.
The string value is namespaced (model_accuracy) to keep it distinct
from DataQualityEnum.ACCURACY (data_accuracy) when both feed the
MetricNameEnum union.
AUC = 'auc'
class-attribute
instance-attribute
¶
The measure of the ability of a classifier to distinguish between classes and is used as a summary of the ROC curve
MSE = 'mse'
class-attribute
instance-attribute
¶
Mean square error used for regression models
PRECISION = 'precision'
class-attribute
instance-attribute
¶
The fraction of true positive responses over total number of positive responses
RECALL = 'recall'
class-attribute
instance-attribute
¶
The fraction of true positive responses over total number of correct responses
ServiceQualityEnum
¶
Attributes¶
AVAILABILITY = 'availability'
class-attribute
instance-attribute
¶
The ratio between up time and down time of the service (%)
RELIABILITY = 'reliability'
class-attribute
instance-attribute
¶
The ratio between correct and total service responses (%)
RESPONSE_TIME = 'response_time'
class-attribute
instance-attribute
¶
The response time of each microservice, measured by the time period between receiving and replying the request (s)