Semantic search apparatus and method using mobile terminal

• US 10,289,717 B2
• Filed: 02/04/2016
• Issued: 05/14/2019
• Est. Priority Date: 03/18/2015
• Status: Active Grant

• Current Assignee: Industry Academic Cooperation Foundation
• Original Assignee: Industry Academic Cooperation Foundation
• Inventors: Lee, Kyong Ho, Song, Min Jae, Shin, Sang Jin, Eom, Sung Kwang
• Primary Examiner(s): Truong, Dennis
• Application Number:
  US15/016,213
• Publication Number:
  US 20160275196A1
• Time in Patent Office: 1,195 Days
• Field of Search: 707769, 707E17015, 707760, 707E17014, 707E17078, 707E17087, 707E17099, 707706, 707771, 704 9, 704 10, 704 4
• CPC Class Codes: G06F 16/24522 Translation of natural language queries to structured queries
• CPC Class Codes: G06F 16/3338 Query expansion

Abstract

Disclosed are an apparatus and a method of searching for information by a mobile device. The present invention provides simplified ontology to be applicable to a mobile environment having a limited resource, and provides ontology capable of providing a combined search environment by combining DBs used by various applications within a mobile device, respectively. Further, the present invention provides a semantic search engine providing a function of searching a local database within a mobile device, and expanding a search to a web as necessary and performing the search.

Claims

1. A search apparatus using a mobile terminal, the search apparatus comprising:

an ontology managing unit configured to manage ontology of data searchable by the mobile terminal; and

a semantic searching unit configured to receive a search query, generate a keyword graph, in which keywords included in the search query are connected, by analyzing a structure of the search query, generate a response graph corresponding to the keyword graph by searching for the keyword included in the keyword graph in the ontology, and search for data corresponding to the search query in a local database by using the response graph,

wherein the semantic searching unit extracts the keywords from words included in the search query by analyzing a structure of the received search query, and classifies the extracted keywords into any one of an object keyword representing a search object of the search query, a relationship keyword representing a relationship of the search query, and an information keyword representing information of the search query, and generates the keyword graph, in which the keywords are connected, and

wherein when the semantic searching unit generates one or more response graphs, the semantic searching unit calculates a score by using similarity or a correlation between each response graph and the information keyword included in the keyword graph and selects the response graph having the highest calculated score as a final response graph, and when the semantic searching unit generates one response graph, the semantic searching unit selects the response graph as the final response graph.

2. The search apparatus of claim 1, wherein the ontology includes concepts, which are abstract concepts for the data stored in the local database, and connection edges, which connect the concepts with each other by using the concepts as nodes, and

the concepts connected by one connection edge are superordinate concepts representing a higher concept or subordinate concepts representing a lower concept according to a superordinate and subordinate relationship of the concepts therebetween.

3. The search apparatus of claim 2, wherein the keyword graph has the object keyword as a start node and is a graph, in which the relationship keyword node and the information keyword node are connected.

4. The search apparatus of claim 2, wherein the semantic searching unit selects the concepts corresponding to the keywords by searching for the keywords included in the keyword graph in the ontology, and generates the response graph by connecting the selected concepts.

5. The search apparatus of claim 4, wherein the semantic searching unit generates the response graph by, when the keyword corresponds to the object keyword, selecting the concept corresponding to the keyword in the ontology and setting the selected concept as a start node of the response graph, by, when the keyword corresponds to the relationship keyword, selecting the connection edge corresponding to the keyword from the connection edges included in a subordinate level of the concept set as the start node of the response graph in the mobile ontology and adding the selected connection edge to the response graph as an edge, and by, when the keyword corresponds to the information keyword, selecting the concept corresponding to the keyword from the concepts, which are included in a subordinate level of the connection edge added to the response graph as the edge, in the ontology and adding the selected concept to the response graph as a node.

6. The search apparatus of claim 5, wherein when the semantic searching unit selects the connection edge corresponding to the keyword in the ontology in the case where the keyword corresponds to the relationship keyword, the semantic searching unit selects the connection edge having a synonym relationship with the keyword as the connection edge corresponding to the keyword.

7. The search apparatus of claim 5, wherein when the semantic searching unit selects the concept corresponding to the keyword in the ontology in the case where the keyword corresponds to the information keyword, the semantic searching unit selects the concept as the concept corresponding to the keyword when data including the keyword as a partial structure exists in the data of the local database corresponding to the concept.

8. The search apparatus of claim 5, wherein the semantic searching unit connects the start node of the response graph with a supplement concept through the connection edge, and searches for data corresponding to the supplement concept in an external web connected with the semantic searching unit by using a web application program interface (API), and obtains the data.

9. The search apparatus of claim 8, wherein the semantic searching unit searches for the data corresponding to the supplement concept by using the concept included in the ontology as an input variable of the web API, and connects the concept, which becomes the input variable of the web API, with the supplement concept through the connection edge and adds the concept to the response graph.

10. The search apparatus of claim 4, wherein the semantic searching unit calculates the score by weight-summing the similarity and the correlation.

11. The search apparatus of claim 4, wherein the similarity is a value obtained by measuring character string similarity between the information keyword included in the keyword graph and the data of the local database corresponding to the concept of the response graph corresponding to the information keyword, and

the correlation is a value obtained by calculating a degree, by which the data corresponding to the information keyword occupies a position in a field.

12. The search apparatus of claim 4, wherein the semantic searching unit generates a structured query language (SQL) query corresponding to the final response graph, and searches for data corresponding to the final response graph in the local database by executing the generated SQL query, and obtains the data, and

the SQL query includes sub-SQL queries having an inclusion relationship with each other.

13. A search apparatus using a mobile terminal, the search apparatus comprising:

a query organizing unit configured to receive a search query, analyze a structure of the search query, and generate a keyword graph, in which keywords included in the search query are connected;

a response graph generating unit configured to search for the keywords included in the keyword graph in the ontology for data stored in a local database and generate a response graph corresponding to the keyword graph; and

a local data searching unit configured to search for data corresponding to the search query in the local database by using the response graph,

wherein the query organizing unit extracts the keywords from words included in the search query by analyzing a structure of the received search query, and classifies the extracted keywords into any one of an object keyword representing a search object of the search query, a relationship keyword representing a relationship of the search query, and an information keyword representing information of the search query, and generates the keyword graph, in which the keywords are connected,

wherein when the response graph generating unit generates one or more response graphs, the response graph generating unit calculates a score by using similarity or a correlation between each response graph and the information keyword included in the keyword graph and selects the response graph having the highest calculated score as a final response graph, and when the response graph generating unit generates one response graph, the response graph generating unit selects the response graph as the final response graph.

14. The search apparatus of claim 13, further comprising:

a web searching unit configured to connect a start node of the response graph with a supplement concept through the connection edge, and search for data corresponding to the supplement concept in an external web connected with the search apparatus using the mobile terminal by using a web application program interface (API), and obtain the data.

15. A search method using a mobile terminal, the search method comprising:

a query organizing step of receiving a search query, analyzing a structure of the search query, and generating a keyword graph, in which keywords included in the search query are connected;

a response graph generating step of searching for the keywords included in the keyword graph in the ontology for data stored in a local database and generating a response graph corresponding to the keyword graph; and

a local data searching step of searching for data corresponding to the search query in the local database by using the response graph,

wherein the query organizing step comprises extracting the keywords from words included in the search query by analyzing a structure of the received search query, and classifying the extracted keywords into any one of an object keyword representing a search object of the search query, a relationship keyword representing a relationship of the search query, and an information keyword representing information of the search query, and generating the keyword graph, in which the keywords are connected,

wherein the response graph generating step comprises, when one or more response graphs are generated, calculating a score by using similarity or a correlation between each response graph and the information keyword included in the keyword graph and selecting the response graph having the highest calculated score as a final response graph, and one response graph generated, selecting the response graph as the final response graph.