Directed Graph Index for Phrase Expansion in Search Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional information search engines in specialized fields like medicine are inefficient due to their reliance on direct indexing of specific phrases, failing to account for related phrases, synonyms, and derivatives, which can lead to biased search results and limited scope, especially when users query phrases not directly present in the index.
Innovation Solution
An apparatus and method that create an index using a directed graph to connect and expand related phrases, allowing for efficient search on a database by including synonyms, equivalent terms, abbreviations, and derivative words, and providing users with the grounds for search results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional search engines index only specific phrases directly from collected information, then the indexing process is simple and fast, but the search scope is limited and results are biased when users query phrases not directly present in the index
Solution Approach 1:
The system performs preliminary expansion of search phrases by generating related phrases (synonyms, equivalent terms, abbreviations, derivatives) before the actual search operation. This preliminary action ensures that when users query phrases not directly in the index, the system has already prepared alternative search terms to match relevant documents, thereby expanding search scope without compromising indexing speed.
Solution Approach 2:
The system introduces an intermediary layer between the user query and the document index. This intermediary generates and manages a set of related phrases that act as mediators to bridge the gap between exact phrase matches and semantically equivalent expressions, enabling the system to retrieve relevant documents even when the exact query phrase is not present in the index.
2Reliability
If the system expands the index to include related phrases, synonyms, and derivatives, then search completeness improves, but the time and computational resources required for index creation increase
Solution Approach 1:
The system applies local quality by selectively expanding only certain phrases in the index based on their relevance and search frequency, rather than uniformly expanding all phrases. This approach generates related phrases (synonyms, equivalent terms, abbreviations, derivatives) only for key search terms, thereby improving search result completeness for critical queries while minimizing the overall time and computational resources required for index creation.
Solution Approach 2:
The system dynamically adjusts the expansion parameters based on the specific query and context. By changing parameters such as the depth of phrase expansion, the types of related phrases generated, and the selection criteria for which phrases to expand, the system optimizes the balance between search result completeness and index creation efficiency for different search scenarios.
Data Source
AI summary
The present invention is an apparatus for creating an index for conducting a search on a database based on a resource including a text sentence. The apparatus comprises: a dictionary that includes a plurality of phrases associated with each other for each specific concept; an input interface unit that receives an input of the resource; a text processing unit that performs processing for extracting a plurality of words, as tokens, from the text sentence in the received resource; a directed graph generation unit that generates a directed graph representing a connection relationship between the extracted plurality of tokens in accordance with the text sentence; a directed graph search unit that conducts a search on the directed graph based on a search-target-phrase and that, if the search-target-phrase is found in the dictionary, locates a location in the directed graph where the search-target-phrase appeared; and an index creation unit that creates an index indicating association with the resource based on the plurality of tokens in the directed graph.