View source for 18155295. Entity Classification Using Graph Neural Networks simplified abstract (International Business Machines Corporation)

=Entity Classification Using Graph Neural Networks=


==Organization Name==

[[:Category:International Business Machines Corporation|International Business Machines Corporation]]


[[Category:International Business Machines Corporation]]

==Inventor(s)==

[[:Category:Aniket Saxena of Ghaziabad (IN)|Aniket Saxena of Ghaziabad (IN)]][[Category:Aniket Saxena of Ghaziabad (IN)]]

[[:Category:Balaji Ganesan of Bengaluru (IN)|Balaji Ganesan of Bengaluru (IN)]][[Category:Balaji Ganesan of Bengaluru (IN)]]

[[:Category:Muhammed Abdul Majeed Ameen of Kozhikode (IN)|Muhammed Abdul Majeed Ameen of Kozhikode (IN)]][[Category:Muhammed Abdul Majeed Ameen of Kozhikode (IN)]]

[[:Category:Avirup Saha of Kolkata (IN)|Avirup Saha of Kolkata (IN)]][[Category:Avirup Saha of Kolkata (IN)]]

[[:Category:Arvind Agarwal of New Delhi (IN)|Arvind Agarwal of New Delhi (IN)]][[Category:Arvind Agarwal of New Delhi (IN)]]

[[:Category:Abhishek Seth of Deoband (IN)|Abhishek Seth of Deoband (IN)]][[Category:Abhishek Seth of Deoband (IN)]]

[[:Category:Soma Shekar Naganna of Bengaluru (IN)|Soma Shekar Naganna of Bengaluru (IN)]][[Category:Soma Shekar Naganna of Bengaluru (IN)]]

==Entity Classification Using Graph Neural Networks - A simplified explanation of the abstract==


This abstract first appeared for US patent application 18155295 titled 'Entity Classification Using Graph Neural Networks

The abstract describes a computer-implemented method for classifying records using a graph neural network.

*Simplified Explanation: The method involves creating a training dataset of matched records related to an entity, training a graph neural network, and classifying records based on their relationship to the entity.

*Key Features and Innovation:
- Creation of a training dataset comprising subgraphs of matched records
- Importance of attributes in matched records identified
- Records in a subgraph related by a subset of attributes
- Training of a graph neural network by processor units
- Classification of records based on their relationship to the entity

*Potential Applications:
- Data classification in various industries
- Fraud detection in financial transactions
- Customer relationship management in businesses
- Healthcare record management
- Social network analysis

*Problems Solved:
- Efficient classification of records based on relationships
- Identification of important attributes in records
- Improved data organization and analysis

*Benefits:
- Enhanced data classification accuracy
- Streamlined record management processes
- Increased efficiency in data analysis
- Better decision-making based on classified data

*Commercial Applications:
- Data analytics software for businesses
- Fraud detection tools for financial institutions
- Customer relationship management systems
- Healthcare data management platforms

*Questions about Graph Neural Network Classification:
1. How does the graph neural network classify records based on their relationship to the entity?
2. What are the potential limitations of using a graph neural network for record classification?

*Frequently Updated Research:
- Stay updated on advancements in graph neural network technology for record classification.


==Original Abstract Submitted==

A computer implemented method classifies records. A number of processor units creates a training dataset comprising subgraphs of matched records matched to an entity and identifying an importance of attributes in the matched records. The matched records in a subgraph are related to each other by a subset of the attributes. The number of processor units trains a graph neural network using the training dataset. The graph neural network classifies the records as belonging to the entity.



[[Category:G06N3/08]]
[[Category:CPC_G06N3/08]]