Recursive Directory Tree for Information Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information retrieval technologies face challenges in providing accurate, concise, and logically structured search tips, often relying on user behavior and generating large amounts of unorganized data, which can lead to disorientation and inefficiency in finding relevant information.
Innovation Solution
A method and system that refine semantics by creating a directory table of retrieval words, establishing kinship between input words and associated data items, and presenting search results in a recursive directory tree structure, independent of user behavior, ensuring accurate and logically organized information retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If statistical methods are used to generate search tips based on user behavior, then the quantity of search tips increases, but the quality and logical structure of the tips deteriorate
Solution Approach 1:
The patent segments search tips into two distinct categories: semantic tips (derived from knowledge graphs and semantic relationships) and statistical tips (derived from user behavior). This segmentation allows the system to maintain high-quality, logically-structured semantic tips while optionally supplementing with statistical tips, thereby resolving the contradiction between quantity and quality.
Solution Approach 2:
The patent introduces a knowledge graph as an intermediary layer between user queries and search tips. The knowledge graph provides semantically-rich, logically-structured relationships that mediate between raw statistical data and final tip generation, ensuring high quality and logical coherence while enabling comprehensive tip coverage.
2Quantity of substance
If unorganized search tip data is provided to users, then the quantity of information increases, but the ease of finding valuable information decreases
Solution Approach 1:
The patent segments and hierarchically organizes search tips using parent-child relationships in the knowledge graph. Tips are structured from general to specific, with parent nodes representing broader concepts and child nodes representing more specific information. This hierarchical segmentation enables users to easily navigate and find valuable information without being overwhelmed by unorganized data.
Solution Approach 2:
The patent adds a semantic dimension to search tip organization by incorporating semantic relationships (is-a, part-of, associated-with) from the knowledge graph. This transforms flat, one-dimensional tip lists into multi-dimensional semantic structures, enabling users to explore information along multiple semantic pathways and easily locate relevant information.
3Adaptability or versatility
If search tips are generated based on massive user behavior statistics, then the coverage of tips increases, but the precision and accuracy of tips decrease
Solution Approach 1:
The patent performs preliminary action by pre-building a comprehensive knowledge graph with accurate semantic relationships before search operations. This pre-established semantic framework ensures high precision and accuracy in tip generation, while the knowledge graph's comprehensive coverage of domain knowledge provides extensive tip coverage, resolving the contradiction between coverage and precision.
Solution Approach 2:
The patent implements feedback mechanisms where search results and user interactions continuously refine and update the knowledge graph. This feedback loop enhances both the coverage (by incorporating new domains and relationships) and precision (by validating and correcting semantic relationships) of search tips over time.
Data Source
AI summary
A retrieval method and system are disclosed. The method involves: A). searching a catalog table of retrieval keywords based on a retrieval keyword inputted by a user at a terminal, and acquiring a first data terms set including the inputted retrieval keyword; B). searching a data table of information index based on each data term of the first data terms set related to the inputted retrieval keyword, and acquiring a second data terms set; C). combining and sending the first data terms set to the terminal wherein the first data terms set is combined in a way of recursion, and sending the second data terms set to the terminal. The solution extracts semantics by formalizing retrieval character strings, and guides and returns retrieval information, thereby improving the deterministic, logic and efficiency of retrieval and user experience.