Location sensitive ensemble classifier

1. A computer-implemented method for generating and using a location sensitive ensemble classifier for classifying content, the method performed by a computing system having one or more processors, the computer-implemented method comprising:

obtaining a general ensemble classifier C, which comprises a plurality of classifiers C-N, where N is a positive integer, and is trained to cover multiple regions based on a training data set T;

dividing a validation data set V into a region set;

wherein the region set divides the validation data set V into a plurality of regions, and each region of the plurality of regions comprises a plurality of data points of the validation data set V that fall within the region;

generating a regional ensemble classifier for a region of the plurality of regions, wherein the regional ensemble classifier for the region is an ensemble of one or more of the plurality of classifiers C-N of the general ensemble classifier C and is generated based on the plurality of data points of the validation data set V that fall within the region; and

classifying a content item in at least one of a plurality of classes using the regional ensemble classifier.

2. The computer-implemented method of claim 1 , wherein the classifying the content item in at least one of a plurality of classes using the regional ensemble classifier for the particular region is based on determining that a test data point for the content item belongs to the particular region.

3. The computer-implemented method of claim 1 , wherein dividing the validation data set V into the region set is based on generating a hyperplane set, where the hyperplane set comprises a plurality of generated hyperplanes that divides the validation data set V into the plurality of regions.

4. The computer-implemented method of claim 3 , wherein the generating the hyperplane set is based on a locality sensitive hashing method.

5. The computer-implemented method of claim 1 , wherein the generating the regional ensemble classifier for each region of the plurality of regions is based on applying an ensemble pruning method to the set of classifiers C-N and the plurality of data points that fall within the region.

6. The computer-implemented method of claim 1 , further comprising:

selecting the region set to use to classify content based on a precision score for the regional ensemble classifier for each region of the plurality of regions, where the precision score is computed based on the plurality of data points that fall within the region.

7. The computer-implemented method of claim 1 , further comprising:

selecting the region set to use to classify content based on a recall score for the regional ensemble classifier for each region of the plurality of regions, where the recall score is computed based on the plurality of data points that fall within the region.

8. The computer-implemented method of claim 1 , further comprising:

selecting the region set to use to classify content based on a F1 score for the regional ensemble classifier for each region of the plurality of regions, where the F1 score is computed based on the plurality of data points that fall within the region.

9. One or more non-transitory computer-readable media storing computer-executable instructions which, when executed by a computing system having one or more processors, cause the computing system to perform:

obtaining a general ensemble classifier C, which comprises a plurality of classifiers C-N, where N is a positive integer, and is trained to cover multiple regions based on a training data set T;

dividing a validation data set V into a region set;

wherein the region set divides the validation data set V into a plurality of regions, and each region of the plurality of regions comprises a plurality of data points of the validation data set V that fall within the region;

generating a regional ensemble classifier for a region of the plurality of regions, wherein the regional ensemble classifier for the region is an ensemble of one or more of the plurality of classifiers C-N of the general ensemble classifier C and is generated based on the plurality of data points of the validation data set V that fall within the region; and

classifying a content item as low-quality using the regional ensemble classifier.

10. The one or more non-transitory computer-readable media of claim 9 , wherein the classifying the content item as low-quality using the regional ensemble classifier for the particular region is based on determining that a test data point for the content item belongs to the particular region.

11. The one or more non-transitory computer-readable media of claim 9 , wherein dividing the validation data set V into the region set is based on generating a hyperplane set, where the hyperplane set comprises a plurality of generated hyperplanes that divides the validation data set V into the plurality of regions, wherein the generating the hyperplane set is based on a locality sensitive hashing method.

12. The one or more non-transitory computer-readable media of claim 9 , wherein the generating the regional ensemble classifier for each region of the plurality of regions is based on applying an ensemble pruning method to the set of classifiers C-N and the plurality of data points that fall within the region.

13. The one or more non-transitory computer-readable media of claim 9 , further comprising instructions which, when executed by the computing system, cause the computing system to perform:

selecting the region set to use to classify content based on a precision score for the regional ensemble classifier for each region of the plurality of regions, where the precision score is computed based on the plurality of data points that fall within the region.

14. The one or more non-transitory computer-readable media of claim 9 , further comprising instructions which, when executed by the computing system, cause the computing system to perform:

selecting the region set to use to classify content based on a recall score for the regional ensemble classifier for each region of the plurality of regions, where the recall score is computed based on the plurality of data points that fall within the region.

15. The one or more non-transitory computer-readable media of claim 9 , further comprising instructions which, when executed by the computing system, cause the computing system to perform:

selecting the region set to use to classify content based on a F1 score for the regional ensemble classifier for each region of the plurality of regions, where the F1 score is computed based on the plurality of data points that fall within the region.

16. A computing system comprising:

one or more processors;

storage media; and

instructions stored in the storage media and which, when executed by the computing system, cause the computing system to perform:

obtaining a general ensemble classifier C, which comprises a plurality of classifiers C-N, where N is a positive integer, and is trained to cover multiple regions based on a training data set T;

dividing a validation data set V into a region set;

wherein the region set divides the validation data set V into a plurality of regions, and each region of the plurality of regions comprises a plurality of data points of the validation data set V that fall within the region;

generating a regional ensemble classifier for a region of the plurality of regions, wherein the regional ensemble classifier for the region is an ensemble of one or more of the plurality of classifiers C-N of the general ensemble classifier C and is generated based on the plurality of data points of the validation data set V that fall within the region; and

classifying a content item in at least one of a plurality of classes using the regional ensemble classifier.

17. The computing system of claim 16 , wherein the classifying the content item in at least one of a plurality of classes using the regional ensemble classifier for the particular region is based on determining that a test data point for the content item belongs to the particular region.

18. The computing system of claim 16 , wherein dividing the validation data set V into the region set is based on generating a hyperplane set, where the hyperplane set comprises a plurality of generated hyperplanes that divides the validation data set V into the plurality of regions, wherein the generating the hyperplane set is based on a locality sensitive hashing method.

19. The computing system of claim 16 , wherein the generating the regional ensemble classifier for each region of the plurality of regions is based on applying an ensemble pruning method to the set of classifiers C-N and the plurality of data points that fall within the region.

20. The computing system of claim 16 , further comprising instructions which, when executed by the computing system, cause the computing system to perform:

selecting the region set to use to classify content based on a precision score and a recall score for the regional ensemble classifier for each region of the plurality of regions, where the precision score and the recall score are each computed based on the plurality of data points that fall within the region.