Granted Patent B2

US 12,488,285 · App. 18/059,852 · Granted Dec 2, 2025

Inventor: Paul O′Hara (Dublin, IE)

Assignee: Business Objects Software LTD

G06N20/00G06F18/214

Abstract

Claims (61)

1 . One or more non-transitory computer-readable media storing computer-executable instructions that, when executed by at least one processor, perform a method for applying category classification labels to a new feature contribution score dataset, the method comprising:

obtaining a plurality of historical feature contribution score (FCS) datasets, each historical FCS dataset comprising a first plurality of feature contribution scores and a size of the historical FCS dataset,

applying default feature contribution category classification (FCCC) parameters to the plurality of historical FCS datasets;

optimizing the default FCCC parameters to produce a plurality of optimized FCCC parameters;

producing a training dataset comprising the optimized FCCC parameters;

using the training dataset to train a machine learning model to apply the category classification labels;

applying the new FCS dataset to the machine learning model, the new FCS dataset comprising a second plurality of feature contribution scores and a size of the new FCS dataset; and

providing the category classification labels for the new FCS dataset to a user interface.

2 . The non-transitory computer-readable media of claim 1 , wherein when the number of historical datasets is less than a threshold for a given size of the historical FCS dataset, further comprising:

creating a plurality of materialized feature contribution score datasets, each materialized FCS dataset comprising a third plurality of scores and a size of the materialized FCS dataset;

combining the historical FCS datasets with the materialized FCS datasets to produce a plurality of augmented FCS datasets; and

applying the default FCCC parameters to the plurality of augmented FCS datasets.

3 . The non-transitory computer-readable media of claim 1 , wherein a feature contribution score indicates the importance of an input feature to a target feature of the machine-learning model.

4 . The non-transitory computer-readable media of claim 1 , wherein optimizing the default feature contribution category classification parameters comprises:

retrieving base category classification thresholds;

sampling a plurality of FCS datasets to determine the size of the FCS dataset; and

producing augmented category classification thresholds.

5 . The non-transitory computer-readable media of claim 1 , wherein the category classification labels are customizable by a user for display on the user interface.

6 . The non-transitory computer-readable media of claim 1 , wherein the category classification labels are model agnostic.

7 . The non-transitory computer-readable media of claim 1 , wherein the size of the new FCS dataset is in the range of 2 to 200.

8 . A method for applying category classification labels to a new feature contribution score dataset, the method comprising:

obtaining a plurality of historical feature contribution score (FCS) datasets, each historical FCS dataset comprising a first plurality of feature contribution scores and a size of the historical FCS dataset,

applying default feature contribution category classification (FCCC) parameters to the plurality of historical FCS datasets;

optimizing the default FCCC parameters to produce a plurality of optimized FCCC parameters;

producing a training dataset comprising the optimized FCCC parameters;

using the training dataset to train a machine learning model to apply the category classification labels;

applying the new FCS dataset to the machine learning model, the new FCS dataset comprising a second plurality of feature contribution scores and a size of the new FCS dataset; and

providing the category classification labels for the new FCS dataset to a user interface.

9 . The method of claim 8 , wherein when the number of historical datasets is less than a threshold for a given size of the historical FCS dataset, further comprising:

creating a plurality of materialized feature contribution score datasets, each materialized FCS dataset comprising a third plurality of scores and a size of the materialized FCS dataset;

combining the historical FCS datasets with the materialized FCS datasets to produce a plurality of augmented FCS datasets; and

applying the default FCCC parameters to the plurality of augmented FCS datasets.

10 . The method of claim 8 , wherein a feature contribution score indicates the importance of an input feature to a target feature of the machine-learning model.

11 . The method of claim 8 , wherein optimizing the default feature contribution category classification parameters comprises:

retrieving base category classification thresholds;

sampling a plurality of FCS datasets to determine the size of the FCS dataset; and

producing augmented category classification thresholds.

12 . The method of claim 8 , wherein category classification labels are customizable by a user for display on the user interface.

13 . The method of claim 8 , wherein the plurality of category classification labels are model agnostic.

14 . The method of claim 8 , wherein the size of the new FCS dataset is in the range of 2 to 200.

15 . A system for applying category classification labels to a new feature contribution score dataset, the system comprising:

at least one processor;

and at least one non-transitory memory storing computer executable instructions that when executed by the at least one processor cause the system to carry out actions comprising:

obtaining a plurality of historical feature contribution score (FCS) datasets, each historical FCS dataset comprising a first plurality of feature contribution scores and a size of the historical FCS dataset,

applying default feature contribution category classification (FCCC) parameters to the plurality of historical FCS datasets;

optimizing the default FCCC parameters to produce a plurality of optimized FCCC parameters;

producing a training dataset comprising the optimized FCCC parameters;

using the training dataset to train a machine learning model to apply the category classification labels;

applying the new FCS dataset to the machine learning model, the new FCS dataset comprising a second plurality of feature contribution scores and a size of the new FCS dataset; and

providing the category classification labels for the new FCS dataset to a user interface.

16 . The system of claim 15 , wherein when the number of historical datasets is less than a threshold for a given size of the historical FCS dataset, further comprising:

creating a plurality of materialized feature contribution score datasets, each materialized FCS dataset comprising a third plurality of scores and a size of the materialized FCS dataset;

combining the historical FCS datasets with the materialized FCS datasets to produce a plurality of augmented FCS datasets; and

applying the default FCCC parameters to the plurality of augmented FCS datasets.

17 . The system of claim 15 , wherein a feature contribution score indicates the importance of an input feature to a target feature of the machine-learning model.

18 . The system of claim 15 , wherein optimizing the default feature contribution category classification parameters comprises:

retrieving base category classification thresholds;

sampling a plurality of FCS datasets to determine the size of the FCS dataset; and

producing augmented category classification thresholds.

19 . The system of claim 15 , wherein the category classification labels are model agnostic.

20 . The system of claim 15 , wherein the size of the new FCS dataset is in the range of 2 to 200.

Assignments (2)

Continuity (1)

References Cited (10)