Personalized Query Completions Through Session Co-Occurrence Ranking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing search engines fail to provide personalized query completions by considering a user's previous queries, leading to irrelevant suggestions and search results.
Innovation Solution
A system utilizing a general-purpose action prediction engine ranks query completions based on their likelihood to co-occur with a user's previous queries, incorporating user-specific data from activity sessions to enhance personalization and relevance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the search system provides query completions based on general popularity without considering user history, then the system complexity remains low, but the relevance and personalization of query completions deteriorates
Solution Approach 1:
The system pre-computes and stores co-occurrence statistics between queries in a compressed format during an offline phase. This preliminary action allows the online query completion system to quickly retrieve and use pre-analyzed patterns without performing complex real-time computations, thus improving relevance while maintaining low system complexity during operation
Solution Approach 2:
Instead of storing complete user activity sessions, the system creates a compressed representation by extracting only the essential co-occurrence patterns between queries. This copying approach captures the necessary information for personalization while significantly reducing storage requirements and computational complexity
2Measurement precision
If the search system analyzes complete user activity sessions to generate personalized query completions, then the personalization quality improves, but the processing time and latency increase
Solution Approach 1:
The system extracts only the essential co-occurrence patterns between queries from complete user activity sessions, discarding redundant information such as timestamps, user identifiers, and other non-essential data. This extraction process maintains personalization quality by preserving the key relationships while dramatically reducing the data volume that needs to be processed in real-time
Solution Approach 2:
The system performs the computationally intensive analysis of user activity sessions offline, before real-time query completion is needed. The results are stored in a compressed format that can be quickly retrieved during online operation, thus eliminating the latency that would result from performing full session analysis in real-time
3Measurement precision
If the search system stores and processes complete user activity sessions, then the accuracy of co-occurrence analysis improves, but the storage requirements and computational resources increase
Solution Approach 1:
The system creates a compressed copy of user activity data that contains only the co-occurrence patterns between queries. Instead of storing complete activity sessions with all their details, the system stores a condensed representation that captures the essential relationships needed for accurate query completion while using minimal storage space
Solution Approach 2:
The system extracts only the co-occurrence information from complete user activity sessions, discarding all other data. This extraction process maintains the accuracy needed for generating relevant query completions while reducing storage requirements from terabytes of raw activity data to a compact co-occurrence matrix
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for ranking query completions based on activity session data. One of the methods includes receiving a query prefix from a user. A reference parameter for the user is obtained. One or more likely queries that are likely to co-occur with the reference parameter in user activity sessions are identified, wherein each likely query has an associated prediction score. A ranking of the one or more likely queries is provided in response to receiving the query prefix.