METHOD AND SYSTEM FOR ESTIMATING A SENTIMENT FOR AN ENTITY

• US 20090216524A1
• Filed: 02/26/2008
• Published: 08/27/2009
• Est. Priority Date: 02/26/2008
• Status: Active Grant

• Current Assignee: Unify Patente GmbH & Co. KG (Atos Se)
• Original Assignee: Unify GmbH & Co. KG (Atos SE)
• Inventors: Skubacz, Michal, Ziegler, Cai-Nicolas, Suda, Peter
• Granted Patent: US 8,239,189 B2
• Time in Patent Office: Days
• US Class Current: 704/9
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A method for estimating a sentiment conveyed by the content of information sources towards an entity is presented. The sentiment is obtained with respect to a query context that may be specified, e.g. by specific terms or expressions, like a product or service name. A sentiment dictionary having a plurality of sentiment terms is provided, wherein each sentiment term has assigned a sentiment value, and at least one of said sentiment terms is associated to a group context. Text documents are screened for occurrences of sentiment terms that are associated to a group context corresponding to the query context. Calculating a sentiment score value is performed as a function of the occurrences of sentiment terms having a similar or same group context as the query context. The method may be carried out automatically without manual analysis of the actual semantic content of the text documents under consideration.

Claims

1. A method for estimating a sentiment for an entity comprising:

providing a sentiment dictionary having a plurality of sentiment terms, each sentiment term having assigned a sentiment value, at least one of said sentiment terms is associated to a group context;

defining a context for estimating the sentiment of said entity as a query context;

screening at least one text document for occurrences of sentiment terms being associated to a group context corresponding to the query context; and

calculating a sentiment score value as a function of an occurrence of a sentiment term being associated to a group context corresponding to the query context and a match of a context of said text document with the query context.

2. The method of claim 1, wherein calculating the sentiment score value comprises summing the sentiment values of sentiment terms that occur in the screened text documents.

3. The method of claim 1, wherein said sentiment values are assigned to a corresponding sentiment term as a function of a measure for a selectiveness of the group context to which the sentiment term is associated.

4. The method of claim 1, comprising retrieving text documents for screening from an information space as a function of the query context.

5. The method of claim 1, wherein said sentiment terms are at least one of the group of a word, a character string, an expression, a unigram, or an n-gram.

6. The method of claim 1, wherein said query context and said group context is characterized by a context identifier, said context identifier comprising at least one of the group of a word, a character string, an expression, a unigram or an n-gram.

7. The method of claim 6, comprising classifying a sentiment term in the screened text document as being associated to a specific group context if the corresponding group identifier occurs within a predetermined textual distance from the sentiment term.

8. The method of claim 7, wherein the textual distance is a number of words between the specific group identifier and the sentiment term.

9. The method of claim 1, comprising changing the sentiment dictionary as a function of the query context.

10. The method of claim 9, wherein changing the sentiment dictionary comprises removing sentiment terms associated to a group context from the dictionary if the group context of said sentiment term does not correspond to the query context.

11. The method of claim 9, wherein changing the sentiment dictionary comprises assigning sentiment values to said sentiment terms as a function of a match between the group context an the query context and as a function of a measure for a selectiveness of the group context to which a sentiment term is associated.

12. The method of claim 1, wherein said text document is at least one of the group of:

a world wide web page, a newspaper page, speech recognized spoken text, a collection of blog entries, or a paragraph comprising a plurality of words.

13. The method of claim 1, wherein said entity is at least one of the group of a company, a division, a department, a person, a concept, or a party.

14. The method of claim 1, wherein said entity has a denomination as an entity identifier.

15. The method of claim 14, the sentiment score value is calculated as a function of a textual difference between an occurrence of the entity identifier in a text document and an occurrence of a sentiment term.

16. The method of claim 1, further comprising building the sentiment dictionary.

17. A method for building a sentiment dictionary comprising:

providing a plurality of sentiment terms,

classifying said sentiment terms into groups, a group having a common associated group context;

assigning a sentiment value to each sentiment term of said sentiment terms as a function of the group context.

18. The method of claim 17, wherein said group context is characterized by a group context identifier comprising at least one item of the group of a word, a character string, an expression, a unigram or an n-gram.

19. The method of claim 18, wherein sentiment values are assigned to sentiment terms as a function of a selectiveness of the context identifier of the group of sentiment terms, selectiveness being determined as a function of a number of items forming the group identifier.

20. The method of claim 17, wherein said sentiment dictionary is implemented as a relational database.

21. The method of claim 17, comprising generating sentiment terms by extracting adjectives from sample text sources.

22. The method of claim 17, wherein said sentiment terms are n-grams having relationally coupled text items.

23. System for estimating a sentiment for an entity comprising:

a storage means for storing a sentiment dictionary having a plurality of sentiment terms, each sentiment term having assigned a sentiment value, and at least one of said sentiment terms is associated to a group context having a group context identifier comprising at least one character string;

an input means for inputting a query context identifier, said query context identifier comprising at least one character string;

a text retrieving means communicatively coupled to an information space for retrieving at least one text document;

a processing platform for

screening said retrieved text document, determining occurrences of sentiment terms in said text document, and

calculating a sentiment score value as a function of score values of sentiment terms occurring in said text documents and as a function of a similarity measure between the group context identifiers of the occurred sentiment terms and the query context identifier.

24. The system of claim 23, further comprising a user interface for allowing a plurality of users to input and change entries of the sentiment dictionary.