Query Subsumption Calculus for Document Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for selecting the best queries for classifying textual documents are inefficient and prone to over-fitting, especially when dealing with large automatically generated query sets, which can lead to increased computation time and errors.
Innovation Solution
The implementation of a query subsumption calculus to determine whether one query is more general than another, allowing for the generation of an optimal classifying query set by analyzing text within a document and updating the query set based on subsumption relationships, thereby reducing the number of queries needed to be run against documents.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a large number of automatically generated queries are used for document classification, then classification coverage is improved, but computation time and resource intensity increase
Solution Approach 1:
The patent extracts and removes redundant queries from the automatically generated query set by applying subsumption relationships. Specifically, it identifies queries that are subsumed by more general queries and removes them, keeping only the most general queries that provide equivalent or better classification coverage with reduced computation.
Solution Approach 2:
The patent creates a hierarchical query structure where general queries serve multiple classification purposes. A general query can subsume multiple specific queries, allowing one query to perform the classification function that would otherwise require multiple separate queries, thereby reducing overall computation time while maintaining classification accuracy.
2Measurement precision
If multiple specific queries are used for fine-grained classification, then classification precision is improved, but the system becomes prone to over-fitting
Solution Approach 1:
The patent implements a dynamic query selection mechanism that adapts the level of query specificity based on the document being classified. The system can switch between using general queries for broad categorization and more specific queries when needed, preventing over-fitting by not always applying the most specific queries while maintaining precision when appropriate.
Solution Approach 2:
The patent changes the parameter of query generality by establishing subsumption relationships between queries with different specificity levels. This allows the system to transform a fixed set of specific queries into a flexible hierarchy where the effective query set can be dynamically adjusted by selecting appropriate levels of generality based on the classification task at hand.
3Reliability
If all automatically generated queries are executed against documents, then complete classification coverage is achieved, but device complexity increases
Solution Approach 1:
The patent segments the query set into hierarchical levels based on subsumption relationships, organizing queries from general to specific. This segmentation allows the system to process only the necessary portion of the query hierarchy for each document, reducing complexity while maintaining complete coverage through the structured organization of queries by generality level.
4Productivity
If redundant queries are removed from the query set, then computation efficiency is improved, but query set optimality becomes harder to determine
Solution Approach 1:
The patent implements a feedback mechanism where the subsumption relationship analysis provides information about query redundancy. By evaluating which queries are subsumed by others, the system receives feedback on which queries can be safely removed, enabling efficient query set optimization while maintaining classification effectiveness through the structured feedback from subsumption analysis.
Data Source
AI summary
In general, embodiments of the present invention provide systems, methods and computer readable media for generating an optimal classifying query set for categorizing and/or labeling textual data based on a query subsumption calculus to determine, given two queries, whether one of the queries subsumes another. In one aspect, a method includes generating a group of determining queries based on analyzing text within a document; receiving a group of classifying queries; and, for each determining query within the group of determining queries, determining whether at least one of the classifying queries is subsumed by the determining query; and updating the group of classifying queries in an instance in which the classifying query is subsumed by the determining query.