Deduction of analytic context based on text and semantic layer

• US 9,519,636 B2
• Filed: 12/02/2008
• Issued: 12/13/2016
• Est. Priority Date: 10/15/2008
• Status: Active Grant

• Current Assignee: Business Objects SA (SAP SE)
• Original Assignee: Business Objects SA (SAP SE)
• Inventors: Cras, Jean-Yves
• Primary Examiner(s): Pulliam, Christyann
• Assistant Examiner(s): Ohba, Mellissa M
• Application Number:
  US12/326,379
• Publication Number:
  US 20100094844A1
• Time in Patent Office: 2,933 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 16/252 between a Database Management System and a front-end application
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A system includes reception of text, extraction of a plurality of linguistic entities and associated linguistic entity categories based on the text; determination of one or more semantic objects of a semantic layer based on the linguistic entity categories, and generation of a query of the semantic layer based on the plurality of linguistic entities, the associated linguistic entity categories, and the one or more semantic objects. The extraction of the plurality of linguistic entities may include identification of the plurality of linguistic entities from a plurality of semantic object-independent linguistic entity categories and a plurality of semantic object-dependent linguistic entity categories, the plurality of semantic object-dependent linguistic entity categories may be associated with the semantic layer, each of the plurality of semantic object-dependent linguistic entity categories may be associated with a respective semantic object of the semantic layer and each entity of each semantic object-dependent linguistic entity category may be associated with a value of its respective semantic object.

Claims

1. A method comprising:

receiving text in a system that includes a processor;

extracting, by the system, a plurality of linguistic entities and associated linguistic entity categories based on the text;

determining, by the system, one or more semantic objects of a semantic layer based on the linguistic entity categories, wherein each of the one or more semantic objects of the semantic layer associates one or more physical entities stored in a data source with user-friendly names;

determining, by the system, an analysis context based on the text, the linguistic entities, and the associated linguistic entity categories;

after determination of the analysis context and the one or more semantic objects by the system, generating, by the system, a query of the semantic layer based on the analysis context determined by the system and the one or more semantic objects determined by the system, wherein the generating the query of the semantic layer comprises;

determining, by the system, based on the linguistic entity categories that were extracted, whether the one or more semantic objects are to be filtered, and in response to determining to filter the one or more semantic objects, determining how to filter the semantic objects,

wherein when a value from a first linguistic entity category matches a value for one of the one or more semantic objects, using the value from the first linguistic entity category as a query filter, and when only a single entity is mentioned for the first linguistic entity category, including the entity in the query filter and removing the first linguistic entity category from a result of the query of the semantic layer; and

receiving a structured data result in response to the query of the semantic layer.

2. The method according to claim 1, wherein extracting the plurality of linguistic entities comprises:

identifying the plurality of linguistic entities from a plurality of semantic object-independent linguistic entity categories and a plurality of semantic object-dependent linguistic entity categories,

wherein the plurality of semantic object-dependent linguistic entity categories are associated with the semantic layer, and

wherein each of the plurality of semantic object-dependent linguistic entity categories is associated with a respective semantic object of the semantic layer and each particular entity of each particular semantic object-dependent linguistic entity category is associated with a value of the respective semantic object associated with the particular semantic object-dependent linguistic entity category.

3. The method according to claim 1, further comprising:

identifying two or more phrases from the text,

wherein determining an analysis context comprises;

for each of the two or more phrases, identifying an associated linguistic entity.

4. The method according to claim 3, wherein extracting the associated linguistic entity comprises:

identifying the associated linguistic entity category from a plurality of semantic object independent linguistic entity categories and a plurality of semantic object-dependent linguistic entity categories,

wherein the plurality of semantic object-dependent linguistic entity categories are associated with the semantic layer, and

wherein each of the plurality of semantic object-dependent linguistic entity categories is associated with a respective semantic object of the semantic layer and each particular entity of each particular semantic object-dependent linguistic entity category is associated with a value of the respective semantic object associated with the particular semantic object-dependent linguistic entity category.

5. The method according to claim 1, wherein generating the query of the semantic layer comprises:

generating a second query based on the analysis context.

6. The method according to claim 1, further comprising:

receiving query results associated with the query; and

determining a visualization based on the query results and on one or more of the associated linguistic entity categories.

7. The method according to claim 6, further comprising:

determining the visualization based on an Analysis Type category associated with the query.

8. The method according to claim 1, wherein the linguistic entity categories include a subject type category, an analysis type category and a dimension type category;

and

wherein the determining an analysis context based on the text, the linguistic entities, and the associated linguistic entity categories comprises;

determining an analysis context including the subject type category, the analysis type category and the dimension type category, wherein the subject type category included in the analysis context specifies a subject of an analysis, said subject of the analysis being a numerical indicator of value, wherein the analysis type category included in the analysis context is different than the subject type category and represents an intent of the analysis, wherein the dimension type category included in the analysis context is different than both the subject type category and the analysis type category and represents an axis of the analysis, wherein values of a quantity are to be disposed along the axis of the analysis.

9. The method according to claim 1, wherein the linguistic entity categories include a subject type category, an analysis type category and a dimension type category;

wherein the method further comprises;

dividing the received text into a plurality of clauses; and

wherein the determining an analysis context based on the text, the linguistic entities, and the associated linguistic entity categories comprises;

determining an analysis context for a first one of the plurality of clauses, the analysis context for the first one of the plurality of clauses including the subject type category, the analysis type category and the dimension type category, wherein the subject type category included in the analysis context for the first one of the plurality of clauses specifies a subject of an analysis, said subject of the analysis being a numerical indicator of value, wherein the analysis type category included in the analysis context for the first one of the plurality of clauses is different than the subject type category and represents an intent of the analysis, wherein the dimension type category included in the analysis context for the first one of the plurality of clauses is different than both the subject type category and the analysis type category and represents an axis of the analysis, wherein values of a quantity are to be disposed along the axis of the analysis.

10. The method according to claim 9, further comprising:

determining an analysis context that is for a second one of the plurality of clauses and separate from the analysis context for the first one of the plurality of clauses, wherein the second one of the plurality of clauses immediately precedes the first one of the plurality of clauses in the received text; and

determining an analysis context that is for a third one of the plurality of clauses and separate from both the analysis context for the first one of the plurality of clauses and the analysis context for the second one of the plurality of clauses, wherein the third one of the plurality of clauses immediately follows the first one of the plurality of clauses in the received text.

11. The method according to claim 10, wherein the first one of the plurality of clauses, the second one of the plurality of clauses and the third one of the plurality of clauses are each a single sentence.

12. The method according to claim 1, wherein the receiving structured data in response to the query of the semantic layer comprises:

receiving structured data from an OLAP data source.

13. The method according to claim 1, wherein the determining, by the system, whether the one or more semantic objects are to be filtered comprises:

if an analysis type category for the analysis context is a first predefined category, not filtering data for values of a semantic object that is explicitly mentioned in a same clause as the analysis type category.

14. A non-transitory computer-readable medium having stored thereon program code, the program code executable by a computer to perform a process comprising:

receiving text;

extracting a plurality of linguistic entities and associated linguistic entity categories based on the text;

determining one or more semantic objects of a semantic layer based on the linguistic entity categories, wherein each of the one or more semantic objects of the semantic layer associates one or more physical entities stored in a data source with user-friendly names;

determining an analysis context based on the text, the linguistic entities, and the associated linguistic entity categories;

after determination of the analysis context and the one or more semantic objects, generating a query of the semantic layer based on the analysis context and the one or more semantic objects, wherein the generating the query of the semantic layer comprises;

determining, based on the linguistic entity categories that were extracted, whether the one or more semantic objects are to be filtered, and in response to determining to filter the one or more semantic objects, determining how to filter the semantic objects,

wherein when a value from a first linguistic entity category matches a value for one of the one or more semantic objects, using the value from the first linguistic entity category as a query filter, and when only a single entity is mentioned for the first linguistic entity category, including the entity in the query filter and removing the first linguistic entity category from a result of the query of the semantic layer; and

receiving a structured data result in response to the query of the semantic layer.

15. The medium according to claim 14, wherein the extracting the plurality of linguistic entities comprises:

identifying the plurality of linguistic entities from a plurality of semantic object-independent linguistic entity categories and a plurality of semantic object-dependent linguistic entity categories,

wherein the plurality of semantic object-dependent linguistic entity categories are associated with the semantic layer, and

wherein each of the plurality of semantic object-dependent linguistic entity categories is associated with a respective semantic object of the semantic layer and each particular entity of each particular semantic object-dependent linguistic entity category is associated with a value of the respective semantic object associated with the particular semantic object-dependent linguistic entity category.

16. The medium according to claim 14, the process further comprising:

identifying two or more phrases from the text,

wherein determination of an analysis context comprises;

for each of the two or more phrases, identification of an associated linguistic entity.

17. The medium according to claim 16, wherein the extracting the associated linguistic entity comprises:

identifying the associated linguistic entity category from a plurality of semantic object independent linguistic entity categories and a plurality of semantic object-dependent linguistic entity categories,

wherein the plurality of semantic object-dependent linguistic entity categories are associated with the semantic layer, and

wherein each of the plurality of semantic object-dependent linguistic entity categories is associated with a respective semantic object of the semantic layer and each particular entity of each particular semantic object-dependent linguistic entity category is associated with a value of the respective semantic object associated with the particular semantic object-dependent linguistic entity category.

18. The medium according to claim 14, wherein the generating the query of the semantic layer comprises:

generating a second query based on the analysis context.

19. The medium according to claim 14, the process further comprising:

receiving query results associated with the query; and

determining a visualization based on the query results and on one or more of the associated linguistic entity categories.

20. The medium according to claim 19, the process further comprising:

determining the visualization based on an Analysis Type category associated with the query.

21. A system comprising:

a semantic layer including a plurality of semantic objects;

a computer comprising;

a linguistic processor to;

receive text; and

extract a plurality of linguistic entities and associated linguistic entity categories based on the text; and

a query builder to;

determine one or more semantic objects of the semantic layer based on the linguistic entity categories, wherein each of the one or more semantic objects of the semantic layer associates one or more physical entities stored in a data source with user-friendly names;

determine an analysis context based on the text, the linguistic entities, and the associated linguistic entity categories;

after determination of the analysis context and the one or more semantic objects, generate a query of the semantic layer based on the analysis context and the one or more semantic objects, wherein the generate a query of the semantic layer based on the analysis context and the one or more semantic objects comprises;

determine, based on the linguistic entity categories that were extracted, whether the one or more semantic objects are to be filtered, and in response to determining to filter the one or more semantic objects, determining how to filter the semantic objects,

wherein when a value from a first linguistic entity category matches a value for one of the one or more semantic objects, using the value from the first linguistic entity category as a query filter, and when only a single entity is mentioned for the first linguistic entity category, including the entity in the query filter and removing the first linguistic entity category from a result of the query of the semantic layer; and

receive a structured data result in response to the query of the semantic layer.

22. The system according to claim 21, the linguistic processor to extract the plurality of linguistic entities by identifying the plurality of linguistic entities from a plurality of semantic object-independent linguistic entity categories and a plurality of semantic object-dependent linguistic entity categories,

wherein the plurality of semantic object-dependent linguistic entity categories are associated with the semantic layer, and

wherein each of the plurality of semantic object-dependent linguistic entity categories is associated with a respective semantic object of the semantic layer and each particular entity of each particular semantic object-dependent linguistic entity category is associated with a value of the respective semantic object associated with the particular semantic object-dependent linguistic entity category.

23. The system according to claim 21, further comprising:

a visualization engine to;

receive query results associated with the query; and

determine a visualization based on the query results and on one or more of the associated linguistic entity categories.

24. The system according to claim 23, the visualization engine further to determine the visualization based on an Analysis Type category associated with the query.

25. A method comprising:

receiving text;

extracting a plurality of linguistic entities and associated linguistic entity categories based on the text;

determining one or more semantic objects of a semantic layer based on the linguistic entity categories;

dividing the received text into a plurality of clauses;

determining an analysis context for a first one of the plurality of clauses based on the text, the linguistic entities, and the associated linguistic entity categories;

generating a query of the semantic layer based on the analysis context and the one or more semantic objects, wherein the generating the query of the semantic layer comprises determining, by the system, based on the linguistic entity categories that were extracted, whether the one or more semantic objects are to be filtered, and in response to determining to filter the one or more semantic objects, determining how to filter the semantic objects, wherein when a value from a first linguistic entity category matches a value for one of the one or more semantic objects, using the value from the first linguistic entity category as a query filter, and when only a single entity is mentioned for the first linguistic entity category, including the entity in the query filter and removing the first linguistic entity category from a result of the query of the semantic layer, and receiving a structured data result in response to the query of the semantic layer; and

receiving a row-set in response to the query of the semantic layer;

wherein the linguistic entity categories include a subject type category, an analysis type category and a dimension type category;

wherein the determining an analysis context based on the text, the linguistic entities, and the associated linguistic entity categories comprises;

determining an analysis context for a first one of the plurality of clauses, the analysis context for the first one of the plurality of clauses including the subject type category, the analysis type category and the dimension type category, wherein the subject type category included in the analysis context for the first one of the plurality of clauses specifies a subject of an analysis, said subject of the analysis being a numerical indicator of value, wherein the analysis type category included in the analysis context for the first one of the plurality of clauses is different than the subject type category and represents an intent of the analysis, wherein the dimension type category included in the analysis context for the first one of the plurality of clauses is different than both the subject type category and the analysis type category and represents an axis of the analysis, wherein values of a quantity are to be disposed along the axis of the analysis;

wherein the analysis type category included in the analysis context for the first one of the plurality of clauses has a name; and

wherein the method further comprises;

using the name of the analysis type category in the analysis context for the first one of the plurality of clauses as a parameter for a rule-based determination of a type of chart most suited to present the row-set that is received in response to the query that is generated based on the analysis context that includes the analysis type category.

26. The method according to claim 25, wherein the generating the query of the semantic layer based on the analysis context and the one or more semantic objects comprises:

generating the query of the semantic layer based on the analysis context for the first one of the plurality of clauses and the one or more semantic objects without considering the analysis type category included in the analysis context for the first one of the plurality of clauses.

27. The method according to claim 25, wherein each of the one or more semantic objects of the semantic layer associates one or more physical entities stored in a data source with user-friendly names.

28. A method comprising:

receiving text in a system that includes a processor;

generating, by the system, a query of a semantic layer in response to the receiving the text; and

receiving, by the system, structured data in response to the query of the semantic layer;

wherein the generating, by the system, a query of the semantic layer comprises;

after receiving the text, extracting, by the system, a plurality of linguistic entities and associated linguistic entity categories based on the text;

determining, by the system, one or more semantic objects of the semantic layer based on the linguistic entity categories;

determining, by the system, an analysis context based on the text, the linguistic entities, and the associated linguistic entity categories; and

after determination of the analysis context and the one or more semantic objects by the system, generating, by the system, the query of the semantic layer based on the analysis context determined by the system and the one or more semantic objects determined by the system, wherein the generating the query of the semantic layer comprises;

determining, by the system, based on the linguistic entity categories that were extracted, whether the one or more semantic objects are to be filtered, and in response to determining to filter the one or more semantic objects, determining how to filter the semantic objects,

wherein when a value from a first linguistic entity category matches a value for one of the one or more semantic objects, using the value from the first linguistic entity category as a query filter, and when only a single entity is mentioned for the first linguistic entity category, including the entity in the query filter and removing the first linguistic entity category from a result of the query of the semantic layer; and

receiving a structured data result in response to the query of the semantic layer.

29. The method according to claim 28, wherein each of the one or more semantic objects of the semantic layer associates one or more physical entities stored in a data source with user-friendly names.