Method and system for generating object classification models

• US 7,996,339 B2
• Filed: 09/17/2004
• Issued: 08/09/2011
• Est. Priority Date: 09/17/2004
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Aklilu, Ameha, Hijer, Raed, Velez, Wilson
• Primary Examiner(s): STARKS, WILBERT L
• Application Number:
  US10/943,343
• Publication Number:
  US 20060074829A1
• Time in Patent Office: 2,517 Days
• Field of Search: 706/12, 706/45, 714/718, 382/125, 600/544, 342/450
• US Class Current: 706/12
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/025 Extracting rules from data

Abstract

A method for generating object classification models is disclosed. Initially, a set of training data is fed into a training algorithm to generate a first object classification model. A set of field data is then applied to the first object classification model to produce a set of field object classifications. The number of data in the set of field data is significantly less than the number of data in the set of training data. Finally, the set of field object classifications and the set of field data are fed into the training algorithm to generate a second object classification model. The second object classification model can be utilized for predicting object classifications.

Claims

1. An object classification system comprising:

a memory unit for storing a computer program for generating object classification models; and

a processor coupled to said memory unit, wherein said processor, responsive to said computer program, comprises;

circuitry for generating a first object classification model by feeding a set of training data related to an object class to be predicted into a training algorithm, wherein said set of training data includes a plurality of objects along with their respective classifications;

circuitry for applying a set of field data related to said object class to be predicted to said first object classification model to produce a set of field object classifications, wherein said set of field data also includes a plurality of objects along with their respective classifications;

circuitry for generating a second object classification model by feeding said set of field data and said set of field object classifications into said training algorithm again; and

circuitry for storing said second object classification model, wherein said second object classification is to be utilized to predict object classifications with a higher prediction accuracy than said first object classification model.

2. The object classification system of claim 1, wherein a number of records in said set of field data is substantially less than a number of records in said set of training data.