Messaging Platform Message Text for Search Relevance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional search techniques face challenges in selecting efficient and relevant signals, particularly with hyperlink anchor text, which is slow to acquire and lacks sentiment indication, making it inadequate for real-time applications.
Innovation Solution
A method and system that identifies message text from a messaging platform referencing documents, calculates a relevance score based on search terms and message text, and provides document references based on these scores, utilizing message text as a relevance signal to enhance search result relevance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If hyperlink anchor text is used as a relevance signal, then search result relevance is improved, but acquisition speed deteriorates due to the time required for new web pages to accumulate inbound links
Solution Approach 1:
The patent introduces message platform data as an intermediary signal source between the document and the search relevance calculation. Instead of directly relying on hyperlink anchor text from web pages, the system uses messages referencing documents to provide relevance signals that are both timely and relevant, resolving the contradiction between relevance quality and acquisition speed
Solution Approach 2:
The system performs preliminary analysis of message platform data to extract and store relevance signals in advance. By pre-processing message text and associating it with referenced documents before search queries arrive, the system prepares relevance information ahead of time, enabling fast response to search requests while maintaining high relevance quality
2Measurement precision
If hyperlink anchor text is used as a relevance signal, then contextual relevance is improved, but sentiment indication capability deteriorates due to the descriptive nature of anchor text
Solution Approach 1:
The patent makes the relevance signal system multi-functional by using message text that can simultaneously provide both contextual relevance information and sentiment indication. Messages naturally contain both descriptive elements (for relevance) and emotional expressions (for sentiment), allowing a single signal source to fulfill multiple functions that previously required separate mechanisms
Solution Approach 2:
The system changes the parameter being measured from purely descriptive anchor text to message text that includes sentiment dimensions. By analyzing message text which inherently contains emotional and contextual information, the system transforms the relevance signal from a single-dimensional descriptive metric to a multi-dimensional signal that captures both relevance and sentiment
3Device complexity
If traditional search signals are used, then system complexity is reduced, but search result relevance deteriorates for real-time applications
Solution Approach 1:
The patent introduces dynamic signal selection that adapts to different application requirements. The system can dynamically adjust which relevance signals are used based on whether real-time performance or comprehensive analysis is prioritized, allowing the complexity level to match the specific application needs while maintaining high relevance quality
Data Source
AI summary
A method and system for searching documents. The method can include: identifying a first set of messages of a messaging platform referencing a document; identifying message text of the first set of messages; associating the message text with the document in a storage repository; receiving, by a computer processor, a search request including a search term provided by a client; calculating, by the computer processor, a relevance score for the document based at least on the search term and the message text; and providing, to the client and based on the relevance score, a reference to the document in response to the search request.