Knowledge Base Search Relevance via Multi-Criteria Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current knowledge base search engines often return irrelevant and numerous search results, making it difficult for users to find relevant content, as they lack structured presentation and fail to highlight relationships between items.
Innovation Solution
A method that generates three lists of knowledge base items: one with defined relationships, one in the same category, and one with matching tags, and selects items present in all three lists for display as most relevant, along with optional graphical aids to illustrate relationships and categories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If knowledge base search engines return all items containing keywords, then the quantity of search results increases, but the relevance and quality of results deteriorates
Solution Approach 1:
The patent segments the search result generation process into three distinct lists: (1) items with defined relationships to the selected item, (2) items in the same category, and (3) items with matching tags. This segmentation allows the system to systematically filter and organize results, returning only items that appear in all three lists, thereby ensuring high relevance while maintaining a manageable quantity of results.
Solution Approach 2:
The patent changes the search parameters from simple keyword matching to a multi-criteria evaluation system. Instead of returning any item containing keywords, the system requires items to satisfy three simultaneous conditions: having a defined relationship, belonging to the same category, and sharing matching tags. This parameter change transforms the search from a broad retrieval to a precise filtering operation.
2Quantity of substance
If knowledge base search engines return comprehensive search results, then the completeness of information increases, but the structure and organization of results deteriorates
Solution Approach 1:
The patent applies segmentation by dividing the result set into three conceptually distinct lists based on different relationship types (defined relationships, category membership, and tag matching). This segmentation provides a clear organizational structure that makes the results more comprehensible while maintaining completeness, as users can understand the different ways items are related to the selected item.
Solution Approach 2:
The patent merges the three segmented lists by finding the intersection of all three sets. Items must appear in all three lists to be included in the final results. This merging operation ensures that only highly relevant items are returned, combining the completeness from multiple sources while maintaining a structured, organized result set.
3Quantity of substance
If knowledge base search engines return numerous search results, then the coverage of potential matches increases, but the ease of operation and user experience deteriorates
Solution Approach 1:
The patent changes the search parameters to require simultaneous satisfaction of three conditions (defined relationship, same category, matching tags). This parameter change dramatically reduces the number of results from potentially hundreds to a small, manageable set of highly relevant items, making the results easy to navigate and use while maintaining comprehensive coverage through the multi-criteria approach.
Solution Approach 2:
The patent replaces the mechanical browsing approach (where users must manually scan through long lists of results) with an automated intelligent filtering system. The system automatically applies the three-criteria filter and presents only the most relevant items, substituting user effort with system intelligence and significantly improving ease of operation.
Data Source
AI summary
Embodiments of the present invention provide a novel and non-obvious method, server and computer program product for finding relevant content in a knowledge base. A method for finding items that are related to a user selected item in a knowledge base is provided. The method can include generating a first list of knowledge base items with a defined relationship to the user selected item and generating a second list of knowledge base items that belong to the same category as a category of the user selected item. The method can further include generating a third list of knowledge base items having one or more tags identical to one or more tags of the user selected item and selecting a first set of knowledge base items that are present in the first, second or third lists. The method can further include displaying the first set of knowledge base items as most relevant.