Predicting Future Queries Using Aggregate User Behavior
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional short-term intent detection applications only predict subsequent queries based on individual user query history within a search session, failing to account for broader trends and aggregated user behaviors, which limits their accuracy and relevance.
Innovation Solution
A computing system that utilizes query logs to generate language models combining a background model and a lifetime model, predicting future queries based on aggregate user behaviors and query histories, allowing for the selection of future queries and corresponding advertisements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional short-term intent detection applications use only individual user query history within a search session, then the system complexity remains low, but the prediction accuracy and relevance of suggested queries deteriorate
Solution Approach 1:
The patent combines multiple data sources including individual user query history, aggregate query logs from multiple users, and contextual information to form a comprehensive prediction model. This merging of data sources improves prediction accuracy by considering both individual user intent and broader usage patterns, while the modular architecture manages the increased complexity through organized data integration.
2Measurement precision
If the system analyzes only queries within the current search session, then the processing time remains short, but the relevance and usefulness of predicted queries deteriorate
Solution Approach 1:
The system performs preliminary analysis by pre-processing and indexing query logs from multiple users before prediction is needed. This allows the prediction model to quickly query pre-computed patterns and statistics during actual use, improving query relevance through comprehensive data analysis while minimizing processing time during the prediction moment itself.
3Reliability
If the system uses aggregate query logs from multiple users, then the prediction model becomes more robust, but the data processing and storage requirements increase
Solution Approach 1:
The system extracts only the essential features and patterns from aggregate query logs that are relevant to prediction, rather than processing all raw data. By extracting key statistical patterns, query transitions, and usage behaviors from the large volume of query logs, the system builds a robust prediction model while managing data storage and processing requirements through selective feature extraction.
Data Source
AI summary
A system, media, and method for selecting future queries are provided. The selected future queries are used to transmit appropriate online advertising to a user that issues queries to a search engine. The search engine is coupled to a prediction component that predicts what subject the user is going to be interested in and when the user will be interested in the subject. The prediction component returns a future query using statistical language models representing a query history of the user and aggregate query histories for a community of users.