Tag LR Index for Document Search and Update
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing document management and retrieval systems face inefficiencies in searching and updating tagged documents, particularly when common terms and frequently appearing tag names are used, leading to prolonged processing times due to the need to check large numbers of nodes and update complex hierarchical structures.
Innovation Solution
A document management system that employs a word index and a tag LR index storing unit to store appearance positions of words and tags, allowing for efficient search queries by utilizing left-right relations between words and tags, and enabling quick tag updates by modifying only relevant parts of the index.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a hierarchical structure (XML) is used to manage tagged documents, then the system can maintain structured document information, but search processing time increases due to the need to check large numbers of nodes
Solution Approach 1:
The patent pre-calculates and stores the left and right words for each tag in the tag LR index during document indexing. This preliminary action allows the search phase to directly retrieve pre-computed adjacency information without performing complex hierarchical traversals during search operations, thus reducing search processing time while maintaining structured document information
Solution Approach 2:
The patent extracts the word adjacency information from the hierarchical XML structure and stores it separately in the tag LR index. By taking out this frequently accessed information from the complex hierarchical structure, the system can perform rapid lookups without traversing the entire hierarchical tree, thereby reducing search processing time while preserving the structured document representation
2Reliability
If a hierarchical structure (XML) is used to manage tagged documents, then the system can maintain structured document information, but updating tags becomes complex and time-consuming
Solution Approach 1:
The patent segments the tag information storage into two independent components: the hierarchical XML structure for document information and the flat tag LR index for word adjacency. When updating a tag, the system only needs to modify the corresponding entry in the tag LR index rather than recalculating the entire hierarchical structure, thus reducing update complexity while maintaining structured document information
Solution Approach 2:
The patent extracts the word adjacency relationships from the hierarchical structure and stores them in a separate, easily updatable tag LR index. This extraction allows tag updates to be performed by simply modifying index entries without complex hierarchical operations, reducing update complexity while preserving the structured document representation
3Device complexity
If traditional indexing is used without left-right relations, then the index structure is simpler, but search speed decreases due to reading larger document lists
Solution Approach 1:
The patent enhances the tag index by adding local quality information - specifically, the left and right words adjacent to each tag. This localized information allows the search algorithm to quickly filter candidate documents by checking word adjacency without examining entire document lists, thus improving search speed while maintaining a relatively simple index structure
Solution Approach 2:
The patent creates a simplified copy of the hierarchical adjacency information in the tag LR index, storing only the essential left and right word relationships needed for search. This copied information enables rapid search operations without requiring access to the full hierarchical structure, improving search speed while keeping the index structure manageable
Data Source
AI summary
A document management & retrieval system is configured to: store, for each word in a set of words, appearance positions of the each word in a set of documents as a word index; store, for each tag in a set of tags attached to words, a set of words that appear to a right and left of the each tag, and also store, as a tag LR index, appearance positions of the each tag in a set of documents with a combination of the each tag and a word appearing to its right or a combination of the each tag and a word appearing to its left as a key; and, in a tag search where a query phrase contains words and a tag next to each other, refer to the index with a tag and the word to the right or left of the tag as a key, thereby reducing the size of a document list to be read without needing to have a tag name as a secondary key. A tag is updated by just updating two places in the tag LR index.