SYSTEMS AND METHODS FOR PROGRAMATICALLY CLASSIFYING TEXT USING TOPIC CLASSIFICATION

• US 20150052002A1
• Filed: 08/15/2013
• Published: 02/19/2015
• Est. Priority Date: 08/15/2013
• Status: Active Grant

• Current Assignee: Nook Digital, LLC (f/k/a Barnesandnoble.com LLC) (Bahwan CyberTek Pvt Ltd.)
• Original Assignee: Nook Digital, LLC (f/k/a Barnesandnoble.com LLC) (Bahwan CyberTek Pvt Ltd.)
• Inventors: Welch, Michael Jason, Vailaya, Aditya, Rabbat, Ralph Rizkallah, Wu, Jiang
• Granted Patent: US 9,672,556 B2
• Time in Patent Office: Days
• US Class Current: 705/26.7
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06F 16/353 into predefined classes
• CPC Class Codes: G06Q 30/0631 Item recommendations

Abstract

Systems and methods for programmatically classifying text are discussed herein. Some embodiments may provide for a system including circuitry configured to programmatically classify a block of text. For example, the circuitry may be configured to identify topics associated with the block of text and identify one or more categories for each of the topics. The circuitry may be further configured to determine unique categories across the one or more categories for each of the topics. For each unique category, an actual category frequency may be determined based on a number of times each of the topics in the block of text is associated with the unique category. The circuitry may be further configured to associate a unique category with the block of text based on the actual category frequency for each the unique category and one or more other actual category frequencies for one or more other unique categories.

Claims

1. A method of programmatically classifying text comprising:

receiving, from a non-transitory computer readable media, a block of text;

identifying topics associated with the block of text;

identifying one or more categories for each of the topics;

determining unique categories across the one or more categories for each of the topics;

determining, by a processor, an actual category frequency for a unique category based on a number of times each of the topics in the block of text is associated with the unique category; and

associating the unique category with the block of text based on the actual category frequency for the unique category and one or more other actual category frequencies for one or more other unique categories to provide a machine-generated summary of the block of text'"'"'s content'"'"'s meaning.

2. The method of claim 1 further comprising:

identifying an expected category frequency for the unique category;

determining a surprise score for the unique category derived from an algorithmic relationship of the actual category frequency and the expected category frequency; and

associating the unique category with the block of text based on the surprise score for the unique category.

3. The method of claim 2, wherein determining the expected category frequency for the unique category includes determining the expected frequency for the unique category based on a length of the block of text and a number of topics associated with the unique category as defined by classification data, the classification data indicating associations between topics and categories of a text classification system.

4. The method of claim 2 further comprising:

determining classification data indicating an expected category frequency for each unique category of a text classification system; and

determining the expected frequency for the unique category based on the classification data.

5. The method of claim 1 further comprising:

determining a topic frequency for each of the topics based on a number of times each topic is identified in the block of text; and

associating the unique category with the block of text based on the topic frequency for each of the topics.

6. The method of claim 1 further comprising:

associating the unique category with a parent category of the unique category;

determining an actual parent category frequency for the parent category of the unique category;

determining a category score for the unique category derived from an algorithmic relationship of the actual parent category frequency and the actual category frequency; and

associating the unique category with the block of text based on the category score for the unique category and one or more other category scores for the one or more other unique categories.

7. The method of claim 1 further comprising:

associating each unique category with M levels of parent categories of each unique category and N levels of child categories of each unique category;

determining category graph data for the block of text including each unique category, the M levels of parent categories for each unique category, and the N levels of child categories for each unique category;

determining a connected category score for the unique category based on an algorithmic relationship of an actual category frequency for the unique category, an actual parent category frequency for each parent category of the M levels of parent categories, and an actual child category frequency for each child category of the N levels of child categories for the unique category; and

associating the unique category with the block of text based on the connected category score for the unique category and one or more other connected category scores for the one or more other unique categories.

8. The method of claim 1, wherein the block of text is a portion of a text comprising a second block of text and further comprising:

receiving, from the non-transitory computer readable media, the second block of text; and

associating the unique category with the text based on a first category frequency of the unique category in the block of text and a second category frequency for the unique category in the second block of text.

9. The method of claim 1, wherein the block of text is a portion of a text comprising a second block of text and further comprising:

receiving, from the non-transitory computer readable media, the second block of text;

associating a second unique category with the second block of text based on a second actual category frequency for the second unique category; and

associating the unique category and the second unique category with the text.

10. The method of claim 1, wherein identifying the topics includes identifying a first topic by comparing a portion of the block of text with a topic list of topics and determining the portion of the block of text matches at least the first topic.

11. A system configured to programmatically classify text, comprising:

circuitry configured to;

receive, from a non-transitory computer readable media, a block of text;

identify topics associated with the block of text;

identify one or more categories for each of the topics;

determine unique categories across the one or more categories for each of the topics;

determine an actual category frequency for a unique category based on a number of times each of the topics in the block of text is associated with the unique category; and

associate the unique category with the block of text based on the actual category frequency for the unique category and one or more other actual category frequencies for one or more other unique categories.

12. The system of claim 11, wherein the circuitry is further configured to:

identify an expected category frequency for the unique category;

determine a surprise score for the unique category derived from an algorithmic relationship of the actual category frequency and the expected category frequency; and

associate the unique category with the block of text based on the surprise score for the unique category.

13. The system of claim 12, wherein the circuitry configured to determine the expected category frequency for the unique category includes the circuitry being configured to determine the expected frequency for the unique category based on a length of the block of text and a number of topics associated with the unique category as defined by classification data, the classification data indicating associations between topics and categories of a text classification system.

14. The system of claim 12, wherein the circuitry is further configured to:

determine classification data indicating the expected category frequency for each unique category of a text classification system; and

determine the expected frequency for the unique category based on the classification data.

15. The system of claim 11, wherein the circuitry is further configured to:

determine a topic frequency for each of the topics based on a number of times each topic is identified in the block of text; and

associate the unique category with the block of text based on the topic frequency for each of the topics.

16. The system of claim 11, wherein the circuitry is further configured to:

associate the unique category with a parent category of the unique category;

determine an actual parent category frequency for the parent category of the unique category;

determine a category score for the unique category derived from an algorithmic relationship of the actual parent category frequency and the actual category frequency; and

associate the unique category with the block of text based on the category score for the unique category and one or more other category scores for the one or more other unique categories.

17. The system of claim 11, wherein the circuitry is further configured to:

associate each unique category with M levels of parent categories of each unique category and N levels of child categories of each unique category;

determine category graph data for the block of text including each unique category, the M levels of parent categories for each unique category, and the N levels of child categories for each unique category;

determine a connected category score for the unique category based on an algorithmic relationship of an actual category frequency for the unique category, an actual parent category frequency for each parent category of the M levels of parent categories, and an actual child category frequency for each child category of the N levels of child categories for the unique category; and

associate the unique category with the block of text based on the connected category score for the unique category and one or more other connected category scores for the one or more other unique categories.

18. The system of claim 11, wherein:

the block of text is a portion of a text comprising a second block of text; and

the circuitry is further configured to;

receive, from the non-transitory computer readable media, the second block of text; and

associate the unique category with the text based on a first category frequency of the unique category in the block of text and a second category frequency for the unique category in the second block of text.

19. The system of claim 11, wherein:

the block of text is a portion of a text comprising a second block of text; and

the circuitry is further configured to;

receive, from the non-transitory computer readable media, the second block of text;

associate a second unique category with the second block of text based on a second actual category frequency for the second unique category; and

associate the unique category and the second unique category with the text.

20. The system of claim 11, wherein the circuitry configured to identify the topic includes the circuitry being configured to identify a first topic by comparing a portion of the block of text with a topic list of topics and determining the portion of the block of text matches at least the first topic.

21. A computer program product comprising a non-transitory computer readable storage medium and computer program instructions stored therein, the computer program instructions comprising program instructions for:

receiving, from a non-transitory computer readable media, a block of text;

identifying topics associated with the block of text;

identifying one or more categories for each of the topics;

determining unique categories across the one or more categories for each of the topics;

determining, by a processor, an actual category frequency for a unique category based on a number of times each of the topics in the block of text is associated with the unique category; and

associating the unique category with the block of text based on the actual category frequency for the unique category and one or more other actual category frequencies for one or more other unique categories.

22. The computer program product of claim 21, wherein the computer program instructions comprising program instructions for:

identifying an expected category frequency for the unique category;

determining a surprise score for the unique category derived from an algorithmic relationship of the actual category frequency and the expected category frequency; and

associating the unique category with the block of text based on the surprise score for the unique category.