Entity Categorization Using Suffix Features and Candidate Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for determining the category of an entity in text understanding, such as context-based and knowledge map-based methods, face challenges in accuracy and granularity, especially when context information is insufficient or the entity is not included in the knowledge map, leading to coarse and imprecise categorization.
Innovation Solution
A method that combines suffix features with candidate categories determined by a categorization model to accurately and finely categorize entities, using a suffix category determining module and a machine-learning categorizing module, which enhances the categorization process without relying on additional context information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If context-based or knowledge map-based methods are used to determine entity category, then the categorization can be performed with available information, but the accuracy and granularity of categorization deteriorate when context information is insufficient or the entity is not included in the knowledge map
Solution Approach 1:
The patent segments the entity name into multiple components including suffixes, and processes each segment separately to extract features. This segmentation allows the system to utilize specific linguistic patterns (suffixes) that indicate category information, thereby improving categorization accuracy even when overall context is insufficient.
Solution Approach 2:
The patent introduces a suffix feature extraction mechanism as an intermediary between the entity name and the categorization model. This intermediary component extracts discriminative features from suffixes and feeds them to the categorization model, enabling more accurate categorization without relying on additional context information or knowledge map entries.
2Measurement precision
If additional context information or knowledge map information is used to determine entity category, then categorization accuracy improves, but the complexity of the system increases due to dependency on external information sources
Solution Approach 1:
The patent extracts and utilizes only the suffix portion of entity names as a key feature for categorization. This extraction approach focuses on the most informative part of the entity name that indicates category information, achieving accurate categorization without requiring the system to process or store additional context information or knowledge map data.
Solution Approach 2:
The system performs self-service categorization by extracting features directly from the entity name itself (specifically suffixes), without needing to query external knowledge maps or retrieve additional context information. This self-contained approach reduces system complexity while maintaining categorization accuracy.
Data Source
AI summary
According to embodiments of the present disclosure, a method, an apparatus, a device, and a computer-readable storage medium for determining a category of an entity are provided. The method includes: based on a suffix of the entity, obtaining a suffix feature associated with the suffix; determining one or more candidate categories of the entity based on a name of the entity; and determining a set of categories of the entity based on the one or more candidate categories and the suffix feature.