LOD Node Classification via Tagged Type Matching

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Traditional named entity classification methods require manual determination of types, leading to a heavy workload and are challenged by the increasing volume of web data, necessitating an efficient automatic classification solution.

Innovation Solution

A method utilizing a processor to read LOD nodes corresponding to named entities, determine tagged types, and compute the possibility of belonging to candidate types based on attribute matching scores derived from an intermediate ontology, facilitating automatic classification.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual determination of types is used for named entity classification, then classification accuracy can be maintained, but the workload increases heavily

Engineering Contradiction:

Improveclassification accuracyVSAvoidworkload

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The system enables automatic classification of named entities by utilizing web data and machine learning algorithms, allowing the system to classify entities autonomously without requiring manual determination of types for each entity, thus reducing the workload while maintaining classification accuracy

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent creates a training set by copying and utilizing existing web data about named entities, including their types and attributes, to train the classification system. This allows the system to learn from pre-existing data rather than requiring manual annotation of every entity

Inventive Principle:

Principle #26Copying

2Extent of automation

If traditional machine learning methods are used for named entity classification, then automation is achieved, but the system struggles with the increasing volume of web data

Engineering Contradiction:

ImproveautomationVSAvoidhandling web data volume

Core Design Contradiction:

Extent of automationVSAdaptability or versatility

Solution Approach 1:

The patent introduces a new dimension by integrating web data as an additional data source beyond traditional training sets. This allows the system to leverage the vast information available on the web, including entity types, attributes, and relationships, to improve classification performance on large-scale data

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The system is designed to handle multiple types of data inputs (traditional training data and web data) and can adapt to classify various types of named entities. The classification model is made universal to work with diverse data sources and entity types, enhancing its versatility in handling increasing web data volume

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS10108705B2Method to improve the named entity classification

Publication Date: 2018.10.23 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

A method is described for providing a named entity classification in a computing system having a processor, comprising the steps of the processor reading, from an LOD (Linking Opening Data) set, an LOD node corresponding to a to-be-classified named entity. The processor also determining a type attribute of the LOD node corresponding to the to-be-classified named entity as a tagged type of the to-be-classified named entity and further reading a candidate type. Finally, the processor computing, based on the tagged type, a possibility of the to-be-classified named entity belonging to the candidate type.