Active Search Intent Prediction via Browsing Behavior Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search engines fail to accurately predict user search intent by not considering the content that motivated the user's query, as they passively determine search intent after the query is submitted without accounting for previous web page content viewed by the user.
Innovation Solution
The system determines a diverse subset of queries by correlating user queries performed shortly after visiting a web page, using a bipartite graph to rank queries based on their likelihood of being triggered by the web page content, and selecting queries with high search trigger likelihood and diverse topics for presentation as suggestions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If search intent is determined passively after query submission, then the system is simpler to implement, but the prediction accuracy deteriorates because it does not account for previous web page content that motivated the search
Solution Approach 1:
The system performs preliminary action by capturing and storing web page content and user browsing behavior before the search query is submitted. This allows the system to analyze what content motivated the user's search intent, improving prediction accuracy by considering the context that exists before the query is made.
Solution Approach 2:
The system implements feedback by using the captured web page content and browsing behavior to refine and improve search intent predictions. The analyzed content from previously viewed pages provides feedback that enhances the accuracy of subsequent search intent predictions, creating a continuous improvement loop.
2Reliability
If query suggestions are based on correlated queries after web page visits, then search result relevance improves, but the system complexity increases due to bipartite graph construction and query ranking
Solution Approach 1:
The system segments the complex task of search intent prediction into distinct components: capturing web page content, building a bipartite graph with web pages and queries, extracting features, ranking queries, and selecting diverse suggestions. This segmentation allows each component to be optimized independently while working together to improve overall relevance.
Solution Approach 2:
The bipartite graph serves as an intermediary structure that connects web pages to queries, enabling the system to analyze relationships between content and search behavior. This intermediary model facilitates reliable search result relevance by mediating between the raw web page content and the final query suggestions.
3Manufacturing precision
If the system analyzes web page content to predict search intent, then the quality of query suggestions improves, but the data processing requirements and computational resources increase
Solution Approach 1:
The system extracts only the necessary features and information from web page content that are relevant to predicting search intent. By selectively extracting key features rather than processing entire web pages, the system maintains high query suggestion quality while reducing computational resource consumption and energy usage.
Solution Approach 2:
The system changes parameters by transforming web page content into feature representations and using these parameters for ranking and selection. This parameter transformation allows efficient processing by working with condensed feature data rather than raw content, reducing computational requirements while maintaining suggestion quality.
4Adaptability or versatility
If diverse queries are selected from ranked results, then the variety of search suggestions improves, but the selection process becomes more complex requiring objective functions and dissimilarity measurements
Solution Approach 1:
The system implements dynamics by using an objective function that adaptively selects diverse queries based on the ranked results. The dissimilarity measurement dynamically adjusts the selection process to ensure topic diversity, allowing the system to adapt to different search contexts and user preferences while maintaining versatility.
Data Source
AI summary
Many search engines attempt to understand and predict a user's search intent after the submission of search queries. Predicting search intent allows search engines to tailor search results to particular information needs of the user. Unfortunately, current techniques passively predict search intent after a query is submitted. Accordingly, one or more systems and/or techniques for actively predicting search intent from user browsing behavior data are disclosed herein. For example, search patterns of a user browsing a web page and shortly thereafter performing a query may be extracted from user browsing behavior. Queries within the search patterns may be ranked based upon a search trigger likelihood that content of the web page motivated the user to perform the query. In this way, query suggestions having a high search trigger likelihood and a diverse range of topics may be generated and/or presented to users of the web page.