Network Estimation Using Machine Learning Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems fail to accurately identify the correct members within an organizational structure for inquiries, as internal organizational information is not publicly available, making it difficult for members of online social networking services to determine whom to contact for specific inquiries without access to private data.
Innovation Solution
A network estimation system that uses supervised and unsupervised machine learning to determine connection strengths and cluster members based on profile similarities, generating connections and recommending peers for inquiries, while allowing members to configure and edit their own estimated organizational structures privately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If internal organizational information is kept private, then organizational security and confidentiality are maintained, but the ability to identify correct members for inquiries is lost
Solution Approach 1:
The system creates a copy of the private organizational structure by training a machine learning model on internal data to generate an estimated organizational chart that can be shared publicly while the original private data remains protected
Solution Approach 2:
The patent introduces an intermediary machine learning model that acts as a mediator between private organizational data and public inquiry routing needs, transforming sensitive internal information into useful public information without exposing the original data
2Measurement precision
If machine learning models are trained on private organizational data, then accurate connection recommendations are generated, but private data must be accessed and processed
Solution Approach 1:
The system performs self-service by automatically training machine learning models on organizational data and generating estimated organizational charts without requiring manual configuration or intervention, reducing the complexity burden on users
3Ease of operation
If estimated organizational structures are shared publicly, then inquiry routing capability is improved, but organizational privacy is compromised
Solution Approach 1:
The system shares a copied version of the organizational structure generated by the machine learning model rather than the original private data, enabling public inquiry routing while maintaining organizational privacy
Data Source
AI summary
This disclosure relates to systems and methods for searching names using name clusters. A method includes training a supervised machine learning system to learn a connection strength between a member and peers of the member; clustering the member with the peers in response to a threshold number of profile similarities between the member and the peers and the connection strength between the member and the peers being above a connection strength threshold value; and applying an unsupervised machine learning system using output from the supervised machine learning system and the clustering to generate a connection between the member and at least one of the peers.