Query-Goal-Mission Structure for Search Query Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search engines lack the capability to accurately classify user queries, leading to ambiguity and irrelevant recommendations due to complex and exploratory search tasks, such as multitasking, which results in users receiving irrelevant content and needing to submit multiple queries to find desired information.
Innovation Solution
A query-goal-mission structure is generated by evaluating query pairs to determine common goal probabilities, grouping them into goal clusters, and further evaluating these clusters to determine mission probabilities, allowing for the identification of user information needs and providing relevant query recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a session-task approach is used to identify search tasks and provide query recommendations, then the system can process search queries and provide basic recommendations, but the accuracy of query classification deteriorates due to complex and exploratory search tasks involving multitasking
Solution Approach 1:
The patent segments the search task identification process into multiple hierarchical levels: individual query analysis, query pair evaluation with common goal probabilities, goal cluster formation, and mission cluster identification. This segmentation allows the system to handle complex multitasking scenarios by breaking down the classification problem into manageable components that can be analyzed separately and then integrated.
Solution Approach 2:
The patent introduces a new dimensional approach by evaluating queries not just within a single session context but across multiple dimensions including query pairs, goal clusters, and mission clusters. This multi-dimensional analysis enables the system to distinguish between primary and secondary search tasks by examining relationships across different levels of abstraction, thereby improving classification accuracy for exploratory and multitasking searches.
2Device complexity
If queries are evaluated using traditional methods without query-goal-mission structures, then the system operation is simpler, but relevant content is missed and irrelevant results are provided
Solution Approach 1:
The patent implements preliminary action by pre-evaluating query pairs to determine common goal probabilities before final mission classification. The system pre-identifies goal clusters and establishes relationships between queries and potential goals/missions in advance, which enables more reliable and relevant search results when actual queries are submitted, while managing complexity through structured preprocessing.
3Ease of operation
If users submit multiple search queries to locate desired content due to irrelevant recommendations, then the user can eventually find relevant information, but the time and computing resources spent on searches increase
Solution Approach 1:
The patent implements feedback mechanisms by using query Goal-Mission structures to provide more accurate and relevant query recommendations upfront. The system analyzes query patterns, goal clusters, and mission contexts to feedback relevant information to users in the initial search results, reducing the need for multiple iterative queries and thereby decreasing the time and computing resources users spend on search tasks.
Data Source
AI summary
One or more systems and/or methods of generating a query-goal-mission structure for a set of queries are provided. A set of queries may be evaluated to identify query information for the queries within the set of queries. The queries may be evaluated as query pairs to determine common goal probabilities (e.g., likelihood two queries correspond to a particular goal, such as to identify vacation planning information) for the query pairs. Responsive to the common goal probabilities for the query pairs exceeding a goal probability threshold, the query pairs may be grouped into goal clusters. The goal clusters may be evaluated as goal cluster pairs to determine common mission probabilities. Responsive to the common mission probabilities for the goal cluster pairs exceeding a mission probability threshold, the goal clusters may be grouped into mission clusters. The mission clusters and the goal clusters may be utilized to generate a query-goal-mission structure.