Get Groupby Statistics
Calculates statistics for a particular categorical groupby column.
Calculates f1, precision, recall, DEP (if applicable), confidence (if applicable), and count, for each group in the groupby column.
Args: project_id run_id split groupby_col: Required query param. The column to calculate statistics on current_user (User, optional): The current user making the HTTP request. Defaults to Depends(authentication_service.current_user). inference_name (bool, default ""): Optional inference name. Required if split is inference groupby_request (ContentRequest, optional): See ContentRequest. Optional filters for the request
Returns: GroupByMetrics
Authorizations
Path Parameters
training, validation, test, inference Query Parameters
Body
training, validation, test, inference Response
Metrics for a cohort returned in columnar format.
Each field in the class is of equal length (or len 0), mapping to the y-axis of each chart. The labels field contains the x-axis labels which apply to all y-axes for all charts created for this data.
ex: labels: ["apple", "banana", "orange"] precision: [0.95, 0.55, 0.83] recall: [0.93, 0.25, 0.88] ...
Was this page helpful?