tw.anomaly_detection
The AnomalyDetection
class provides a comprehensive set of tools for detecting anomalies in time-series data. It supports operations such as model building, anomaly detection, and root cause analysis (RCA). This class is designed to help users efficiently analyze data and identify unusual patterns or behaviors.
Class Initialization
from tangent_works import TangentWorks
# Initialize the TangentWorks instance
tw = TangentWorks()
# Access the AnomalyDetectionAPI
tw.anomaly_detection
Methods
build_model
tw.anomaly_detection.build_model(configuration=Dict[str, Any], dataset=pd.DataFrame)
Builds an anomaly detection model using the provided configuration and dataset.
Parameters
configuration
(Dict[str, Any]): The configuration parameters for model building.dataset
(pd.DataFrame): The time-series dataset used to train the model.
Returns
AnomalyDetectionModel
: The constructed anomaly detection model.
Examples
model = tw.anomaly_detection.build_model(
configuration = config,
dataset = data
)
detect
tw.anomaly_detection.detect(dataset=pd.DataFrame, model=AnomalyDetectionModel)
Detects anomalies in the provided dataset using the specified anomaly detection model.
Parameters
dataset
(pd.DataFrame): The time-series dataset to analyze for anomalies.model
(AnomalyDetectionModel): The anomaly detection model to use.
Returns
pd.DataFrame
: A DataFrame containing the detected anomalies.
Examples
anomalies = tw.anomaly_detection.detect(
dataset = data,
model = model
)
rca
tw.anomaly_detection.rca(configuration=Dict[str, Any], dataset=pd.DataFrame, model=AnomalyDetectionModel)
Performs root cause analysis (RCA) to identify the causes of anomalies detected in the dataset.
Parameters
configuration
(Dict[str, Any]): Configuration parameters for RCA.dataset
(pd.DataFrame): The time-series dataset to analyze for root causes.model
(AnomalyDetectionModel): The anomaly detection model to use for RCA.
Returns
Dict[int, pd.DataFrame]
: A dictionary mapping root causes to their corresponding model indexes.
Examples
rca_results = tw.anomaly_detection.rca(
configuration = config,
dataset = data,
model = model
)
Key Features
Comprehensive Toolset: Supports anomaly detection, model building, and RCA.
Modular Design: Individual methods for building models, detecting anomalies, and analyzing root causes.
Scalable and Flexible: Works with various time-series datasets and configurations.
Dependencies
pandas
: Used for handling time-series data.Tangent Works core and business logic modules: Used for model building, validation, and anomaly detection.