Query Recommendation Clustering via Hierarchical Intent Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Search engines often fail to provide relevant search query recommendations that span the user's intent beyond the initial search term, leading to inefficient search sessions and increased computation during runtime.
Innovation Solution
A method that pre-computes query clusters by forming a vector for each query, determining query similarity, and grouping them into clusters, allowing for the identification of a baseline and neighboring query clusters, which are then organized into a hierarchical arrangement and sent to the user's device for efficient query suggestions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If search engines compute query recommendations in real-time based on user input, then the recommendations can be highly relevant to current search intent, but the computation time and processing load increase significantly
Solution Approach 1:
The patent pre-computes query clusters and their hierarchical relationships during off-peak hours, organizing queries into structured clusters with pre-determined relationships. When a user submits a search query, the system simply retrieves the relevant pre-computed cluster and its neighbors, dramatically reducing real-time computation while maintaining recommendation quality
Solution Approach 2:
The patent segments the vast search query space into discrete, manageable query clusters. Each cluster represents a coherent topic or intent category, and the system pre-computes relationships between neighboring clusters. This segmentation allows efficient navigation through query recommendations without processing the entire query space in real-time
2Adaptability or versatility
If search engines provide comprehensive query recommendations covering all potential user intents, then the user can explore more search directions, but the system complexity and data processing requirements increase
Solution Approach 1:
The patent divides the comprehensive query recommendation space into structured query clusters, where each cluster represents a coherent topic. The system pre-computes hierarchical relationships between clusters, allowing users to navigate through organized groups of related queries rather than presenting all possible queries at once, thus managing complexity while maintaining comprehensive coverage
Solution Approach 2:
The patent introduces a hierarchical dimension to query recommendations by organizing queries into clusters and sub-clusters. Instead of presenting a flat list of all possible queries, the system creates a multi-level structure where users can drill down from broad topic clusters to specific query suggestions, making the comprehensive recommendation system more manageable and navigable
3Ease of operation
If search engines organize query recommendations in a detailed hierarchical structure, then the user can better navigate search intent, but the computation and data organization requirements increase
Solution Approach 1:
The patent performs the computationally intensive task of organizing queries into hierarchical clusters during pre-computation phases, storing the structured relationships in advance. During user interactions, the system only needs to retrieve and display pre-organized clusters, significantly reducing real-time energy consumption while providing detailed hierarchical navigation
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Examples are disclosed that relate to grouping query clusters based on intent of a search query. One example provides a method for generating recommended queries related to a search query, the method performed by a system including a logic subsystem, and memory holding instructions executable by the logic subsystem to receive the search query, execute the search query on an index of pre-computed query clusters to identify a baseline query cluster and one or more neighboring query clusters, form a data structure including queries from within the baseline query cluster and the one or more neighboring query clusters, traverse the data structure to form a hierarchical arrangement of queries, and send the hierarchical arrangement of queries to a computing device from which the search query was received.