US 2019/0057321 A1 · CLASSIFICATION

Abstract

An example method comprises performing for each class from a plurality of classes: constructing binary training set for the class, the binary training set including labeled cases for that class from the main training set other labeled cases from the main training set; training classifier for the class on the binary training set; computing a local calibration threshold using scores of the labeled cases in the binary training set; and adjusting all scores of the label cases in the binary training set with the local calibration threshold to meet a global decision threshold. The method also comprises determining, with the processor, a global hierarchical calibration threshold by using the adjusted scores for all classes to optimize a performance measurement of all trained classifiers. The method further comprises classifying, with the processor, a new case by using a previously trained classifier, a local calibration threshold, and the global hierarchical calibration threshold.

Claims15

Dependent claims

A method comprising, by a processor:

performing, with the processor, for each class from a plurality of classes, wherein each class is associated with a set of labeled cases in a main training set:
constructing a binary training set for the class, the binary training set including labeled cases for that class from the main training set and other labeled cases from the main training set,
training a classifier for the class on the binary training set,
computing a local calibration threshold using scores of the labeled cases in the binary training set, and
adjusting all scores of the labeled cases in the binary training set with the local calibration threshold to meet a global decision threshold;
determining, with the processor, a global hierarchical calibration threshold by using the adjusted scores for all classes to optimize a performance measurement of all trained classifiers;
and
classifying, with the processor, a new case by using a previously trained classifier, a local calibration threshold, and the global hierarchical calibration threshold.

5 more elements

A system comprising:

a classifier generating engine to perform for each class from a plurality of classes, wherein each class is associated with a set of labeled cases in a main training set:
generate a binary training set for the class, the binary training set including
labeled cases for that class from the main training set and other labeled cases from the main training set,
train a classifier for the class on the binary training set,
score the cases in the binary training set excluding any unlabeled cases,
compute a local calibration threshold using scores of the labeled cases in the binary training set, and
adjust all scores of the labeled cases in the binary training set with the local calibration threshold to meet a global decision threshold;
a global hierarchical calibration threshold engine to determine a global hierarchical calibration threshold by using the adjusted scores for all classes and to optimize a performance measurement of all trained classifiers;
and
a classification engine to classify a new case by using a previously trained classifier, a local calibration threshold, and the global hierarchical calibration threshold.

7 more elements

A non-transitory machine-readable storage medium encoded with instructions executable by at least one processor, the machine-readable storage medium comprising instructions to:

perform for each class from a plurality of classes, wherein each class is associated with a set of labeled cases in a main training set:
construct a binary training set for the class, the binary training set including labeled cases for that class from the main training set and other labeled cases from the main training set,
train a classifier for the class on the binary training set,
compute a local calibration threshold using scores of the labeled cases in the binary training set,
adjust all scores of the labeled cases in the binary training set with the local calibration threshold to meet a global decision threshold;
determine, for each class, a set of other classes that are implied by that class;
determine a global hierarchical calibration threshold by using the adjusted scores for all classes to optimize a performance measurement of all trained classifiers;
and
classify a new case by using a previously trained classifier, a local calibration threshold, and the global hierarchical calibration threshold.

6 more elements

Dependencies are inferred from the claim text.

Classification

Life of this patent

Classification

Life of this patent

Abstract

Claims15