US 8,275,726 B2 · Object classification using taxonomies

Abstract

As provided herein objects from a source catalog, such as a provider's catalog, can be added to a target catalog, such as an enterprise master catalog, in a scalable manner utilizing catalog taxonomies. A baseline classifier determines probabilities for source objects to target catalog classes. Source objects can be assigned to those classes with probabilities that meet a desired threshold and meet a desired rate. A classification cost for target classes can be determined for respective unassigned source objects, which can comprise determining an assignment cost and separation cost for the source objects for respective desired target classes. The separation and assignment costs can be combined to determine the classification cost, and the unassigned source objects can be assigned to those classes having a desired classification cost.

Claims20

Dependent claims

A method for classifying an object from a source catalog to a target catalog, comprising:

determining a probability that an object from a source catalog belongs to a class in a target catalog using a baseline statistical classification algorithm;
assigning the object to the class if the probability meets a desired threshold;
and
responsive to the probability not meeting the desired threshold:
determining a classification cost for the object, comprising:
determining an assignment cost for assigning the object to a second class in the target catalog, the assignment cost a function of a probability that the object belongs to the second class;
determining a separation cost for the object based on:
a distance between the second class and a third class, in the target catalog, of a second object;
and
a similarity between the object and the second object;
and
combining the assignment cost and the separation cost to determine the classification cost;
and
determining whether to assign the object to the second class based on the classification cost.

11 more elements

A system for classifying an object from a source catalog to a target catalog, comprising:

a baseline classifier configured to determine a probability that an object from a source catalog belongs to a class in a target catalog;
an object assignment component configured to assign the object to the class if the probability meets a desired threshold;
a classification cost determination component configured to determine a classification cost for the object, comprising:
an assignment cost determination component configured to determine an assignment cost for assigning the object to a second class in the target catalog, the assignment cost a function of a probability that the object belongs to the second class;
a separation cost determination component configured to determine a separation cost for the object based on:
a distance between the second class and a third class, in the target catalog, of a second object;
and
a similarity between the object and the second object;
and
a combination component configured to combine the assignment cost and the separation cost to determine the classification cost;
and
an assignment component configured to determine whether to assign the object to the second class based on the classification cost, at least some of at least one of the baseline classifier, the object assignment component, the classification cost determination component, or the assignment component implemented at least in part via a processing unit.

9 more elements

Dependencies are inferred from the claim text.

Object classification using taxonomies

Life of this patent

Object classification using taxonomies

Life of this patent

Abstract

Claims20