Ontology-Based Query Relevance Weighting for Unrecognized Elements
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for searching and identifying relevant information in repositories are inefficient due to the difficulty in matching and comparing reference documents with unknown documents, leading to unsatisfactory retrieval results.
Innovation Solution
A method and system that map queries with knowledge references to identify related words using parent and sibling relationships, determine the number of occurrences of these elements, and assign weights based on occurrences to determine query relevance, enabling effective retrieval of elements in a hierarchical manner.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If reference words are extracted and mapped with results stored in the database using simple word comparison, then the search process is simple and fast, but the retrieval results are unsatisfactory when comparing word lists of reference documents with unknown documents
Solution Approach 1:
The patent introduces an intermediary ontology layer between the query and the repository. The query is mapped to ontology concepts (classes, properties, values) which then serve as intermediaries to search the repository. This mediator layer transforms simple word comparison into semantic concept matching, improving retrieval accuracy while managing complexity through structured ontology relationships.
Solution Approach 2:
The patent replaces the mechanical word-by-word comparison system with a semantic reasoning system based on ontology. Instead of mechanically comparing word lists, the system uses semantic relationships (parent-child, sibling relationships in ontology) to infer meaning and relevance, substituting brute-force comparison with intelligent semantic matching.
2Productivity
If the vast amount of information stored in the repository is searched using traditional methods, then all information is available for retrieval, but the difficulty in searching relevant documents increases significantly
Solution Approach 1:
The patent segments the vast repository information into structured ontology components (classes, properties, values, relationships). This segmentation organizes the information hierarchy, allowing efficient navigation and search through predefined ontological structures rather than searching through unstructured text, thereby improving search efficiency while maintaining ease of operation.
3Adaptability or versatility
If word occurrence frequency is used to determine relevance, then the method is simple to implement, but it fails to capture the semantic relationships between words
Solution Approach 1:
The patent adds another dimension to word analysis by introducing ontological relationships (parent-child, sibling relationships) beyond simple occurrence frequency. This dimensional transformation from scalar frequency counting to multi-dimensional semantic relationship mapping enables the system to understand semantic connections while managing complexity through structured ontological frameworks.
Data Source
AI summary
In one embodiment, a method is disclosed, comprising: receiving a query comprising at least one unrecognized element in a repository; mapping the query with a plurality of knowledge references to identify one or more words comprising one or more elements related to the unrecognized element of the query; identifying the relationship of the unrecognized element and the words, wherein the relationship is derived using a parent and sibling relationship; determining, by the one or more hardware processors, a number of occurrences of the one or more elements related to the unrecognized element; assigning, by the one or more hardware processors, a weight to an element from the one or more elements, wherein the one or more elements are related to the unrecognized element based upon the number of occurrences; and determining, by the one or more hardware processors, a relevance of the query using the weight assigned to the element.