Search Result Ranking via Anchor-Based User Behavior Data Transfer
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search engines often misallocate user behavior data to secondary sources, leading to inaccurate ranking of primary information sources in search results, as they do not effectively utilize user behavior data from links within documents to improve query relevance.
Innovation Solution
A method is introduced to identify and utilize user behavior data from relevant document anchors, allowing the sharing of this data between documents, thereby promoting primary sources and demoting less relevant ones by deriving quality of result statistics based on user interactions and anchor relevance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If search engines allocate user behavior data to documents based on direct search result interactions, then the ranking reflects immediate user preferences, but user behavior data is misallocated to secondary sources instead of primary information sources
Solution Approach 1:
The patent uses document anchors as intermediaries to transfer user behavior data from secondary source documents to primary information sources. When users interact with search results that contain links to other documents, the system captures these anchor interactions and uses them to infer relevance, thereby mediating the flow of behavioral data to the appropriate primary sources rather than attributing it only to the immediate search result.
Solution Approach 2:
The system implements feedback loops by continuously monitoring user interactions with document anchors and using this information to adjust document rankings. User behavior data collected from anchor clicks and interactions feeds back into the ranking algorithm, allowing the system to learn and improve its allocation of behavioral data to primary sources over time, correcting previous misallocations.
2Productivity
If search engines use traditional ranking methods that focus on direct search result interactions, then the ranking process is simple and fast, but the relevance of query results deteriorates due to inaccurate quality assessment
Solution Approach 1:
The patent performs preliminary analysis of document anchors and their relationships before final ranking determination. By pre-processing and capturing anchor information and user interactions with these anchors, the system prepares quality assessment data in advance, allowing for more accurate document quality evaluation without significantly increasing the time cost of the final ranking process.
3Device complexity
If search engines ignore document anchors and their user interactions, then the ranking system is simpler to implement, but the relevance and accuracy of search results deteriorates
Solution Approach 1:
The patent segments the document quality assessment process into multiple components: direct search result interactions, anchor link interactions, and inferred primary source relationships. By dividing the ranking system into these distinct segments, the patent can incorporate anchor information without requiring a complete redesign of the entire ranking system, thus managing complexity while improving relevance.
Data Source
AI summary
In general, one aspect described can be embodied in a method for providing input to a document ranking process for ranking a plurality of documents, the document ranking process taking as input a quality of result statistic for a query and an individual document. The method can include, for a first document identified as a search result of a query, receiving information regarding an anchor contained within the first document, where the anchor provides a link to a second document; deriving a quality of result statistic for the second document from at least a portion of first data associated with the first document and the query, the first data being indicative of user behavior relative to the first document as a search result for the query; and providing the first quality of result statistic as input to the document ranking process for the second document and the query.