Metrics¶
Includes classes for calculating the following:
Classification performance metrics
Unsupervised performance metrics
Pairwise metrics
Classification Metrics¶
Calculates the Precision for either a single or set of binary predictors. |
|
Calculates the Recall for either a single or set of binary predictors. |
|
Calculates the Fbeta score for either a single or set of binary predictors. |
|
Calculates the revenue for either a single or set of binary predictors. |
Pairwise Metrics¶
Computes the cosine similarity between columns in X. |
|
Computes the Jaccard similarity between columns in X. |
Unsupervised Metrics¶
Calculates the negative squared difference between the number of alerts per day in the binary predictor(s) vs the expected. |
|
Calculates the negative squared difference(s) between the percentage of the overall volume that the binary predictor(s) vs the expected. |