System and method for document categorization

• US 7,496,567 B1
• Filed: 09/28/2005
• Issued: 02/24/2009
• Est. Priority Date: 10/01/2004
• Status: Expired due to Fees

• Current Assignee: Terril John Steichen
• Original Assignee: Terril John Steichen
• Inventors: Steichen, Terril John
• Primary Examiner(s): Vital; Pierre M
• Assistant Examiner(s): Kabir; Sheikh M
• Application Number:
  US11/237,425
• Time in Patent Office: 1,245 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 16/353 into predefined classes
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

The present invention provides methods and systems for automatic categorization of documents. More specifically, the present invention provides for the automatic assignment of a set of pre-defined topics to a set of documents.

Claims

1. A document categorization method implemented by a computer processor for creating associations between one or more documents with a predefined topic, wherein each said predefined topic comprising:

a topic name, a topic threshold and a topic query, wherein said topic threshold comprising a percentage value ranging from zero to one hundred, and wherein said topic query comprising one or more terms, each of said terms comprising;

a word or a phrase, logical and grouping operators that define relationships between said terms, said method comprising;

identifying matching documents using said topic query;

screening said matching documents to produce document-topic associations for each of said documents that also match said topic threshold and said topic query, said screening comprising;

computing a score for each of said matching documents, wherein said score value equals the similarity of each of said document with said topic;

sorting said matching documents in order of said computed score; and

selecting a subset of said matching documents wherein said subset being defined by said topic threshold,

wherein said method implements a bimodal classifier that produces document-topic associations that are consistently accurate in terms of precision and recall over document collections that change over time in size and composition.

2. A computer implemented system for automatically creating associations between one or more documents with a predefined topic, comprising:

a memory; and

a processor;

wherein the predefined topic comprising a topic name, a topic threshold and a topic query, wherein said topic threshold comprising a percentage value ranging from zero to one hundred, and wherein said topic query comprising one or more terms, each of said terms comprising;

a word or a phrase, logical and grouping operators that define relationships between said terms, said system comprising;

identifying matching documents using said topic query;

screening said matching documents to produce document-topic associations for each of said documents that also match said topic threshold and said topic query, said screening comprising;

computing a score for each of said matching documents, wherein said score value equals the similarity of each of said document with said topic;

sorting said matching documents in order of said computed score; and

selecting a subset of said matching documents wherein said subset being defined by said topic threshold,

wherein said system implements a bimodal classifier that produces document-topic associations that are consistently accurate in terms of precision and recall over document collections that change over time in size and composition.

3. A computer implemented system for creating associations between a set of topics and a set of documents, said system including:

a classifier for each topic, each said classifier comprising;

a query expression, each said query expression comprising;

one or more terms, each said term comprising a word, or a phrase, and,

logical and grouping operators, said operators defining relationships between said terms;

a similarity threshold;

a classification engine, said classification engine comprising;

a searcher for determining if a specified document matches a specified query expression,

a scorer for computing the similarity score between each said matching document and said query expression,

a selector for selecting a proportion of said matching documents as specified by said similarity threshold; and

storage means for storing the output of said classification engine,

wherein said system implements a bimodal classifier that produces document-topic associations that are consistently accurate in terms of precision and recall over document collections that change over time in size and composition.

4. A computer-readable medium having computer instructions stored thereon, which, when executed by a processor, causes the processor to perform operations comprising:

creating associations between one or more documents with a predefined topic, wherein said topic comprising;

a topic name, a topic threshold and a topic query, wherein said topic threshold comprising a percentage value ranging from zero to one hundred, and wherein said topic query comprising one or more terms, each of said terms comprising;

a word or a phrase, logical and grouping operators that define relationships between said terms, said operations comprising;

identifying matching documents using said topic query;

screening said matching documents to produce document-topic associations for each of said documents that also match said topic threshold and said topic query, said screening comprising;

computing a score for each of said matching documents, wherein said score value equals the similarity of each of said document with said topic;

sorting said matching documents in order of said computed score; and

selecting a subset of said matching documents wherein said subset being defined by said topic threshold,

wherein said method implements a bimodal classifier that produces document-topic associations that are consistently accurate in terms of precision and recall over document collections that change over time in size and composition.