Ontology Search System with Structural Ranking

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing ontology systems face limitations in search capabilities, as they primarily rely on keyword-based searches, which are inadequate in many scenarios, failing to effectively utilize the structural information of nodes and edges for accurate inference and retrieval.

Innovation Solution

An ontology system that accepts search requests specifying nodes and edges of interest and generates a ranking for matching ontologies based on the extent of matching nodes and edges, enabling users to find desired ontologies more accurately by determining and ranking matching ontologies in a knowledge store.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If keyword-based search is used in ontology systems, then the search operation is simple to implement, but the search accuracy and ability to utilize structural information is insufficient

Engineering Contradiction:

Improvesearch accuracyVSAvoidsearch system complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The search system is segmented into multiple independent modules: a parser module that converts natural language queries into structured search requests, a search engine module that processes the structured requests against the ontology, and a result ranking module that scores and orders results. This segmentation allows each module to specialize in one function, improving overall search accuracy while keeping individual module complexity manageable.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

A structured search request format acts as an intermediary between the simple keyword input and the complex ontology structure. The search request includes fields for concept names, relationships, and optional filters, translating user intent into a form that can be efficiently processed against the ontology's nodes and edges without requiring the system to directly parse complex natural language.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If structural information of nodes and edges is utilized for search, then the search capability is enhanced, but the processing time and computational resources increase

Engineering Contradiction:

Improvesearch reliabilityVSAvoidsearch processing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system performs partial matching by allowing users to specify only certain aspects of their search (e.g., concept names and key relationships) while leaving other aspects to be inferred or optionally specified. The result ranking module calculates relevance scores for matching results rather than requiring exact matches, enabling the system to return useful results faster while maintaining reliability through the scoring mechanism.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The search system allows dynamic adjustment of parameters such as the depth of relationship traversal, the types of relationships to consider, and the ranking thresholds. By changing these parameters based on the specific search query and user needs, the system can balance between comprehensive structural analysis and processing speed, reducing time loss when full analysis is not necessary.

Inventive Principle:

Principle #35Parameter changes

3Measurement precision

If ranking of matching ontologies is generated based on extent of matching, then the result relevance is improved, but the computational complexity of determining rankings increases

Engineering Contradiction:

Improveresult relevance precisionVSAvoidranking algorithm complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The ranking algorithm applies different quality criteria to different aspects of the search results. Exact matches on concept names receive higher base scores, while matches on relationships and inferred properties receive progressively lower scores. This local quality approach allows the system to precisely rank results based on their degree of matching without requiring a single complex scoring function, reducing algorithmic complexity while maintaining precision.

Inventive Principle:

Principle #3Local quality

4Adaptability or versatility

If users can search using concepts, relations, instances and properties, then the versatility of search is enhanced, but the difficulty of formulating search requests increases

Engineering Contradiction:

Improvesearch versatilityVSAvoidsearch request formulation ease

Core Design Contradiction:

Adaptability or versatilityVSEase of operation

Solution Approach 1:

Instead of requiring users to learn the ontology's internal structure and formulate complex queries in terms of nodes and edges, the system inverts the approach by accepting natural language concept names and relationships that users already understand. The parser module then translates these familiar terms into the structured search format, making the system versatile while maintaining ease of operation.

Inventive Principle:

Principle #13The other way round (Inversion)

Data Source

PatentUS8468163B2Ontology system providing enhanced search capability with ranking of results

Publication Date: 2013.06.18 ORACLE INT CORP

AI summary

Ontology system providing enhanced search capability receives a search request specifying nodes and edges of interest and determines a set of matching ontologies stored in a knowledge store. The ontology system also generates a ranking for each of the matching ontologies based on the extent of matching. Data indicating the matching ontologies and corresponding rank is sent as a search result.