Tri-Point Arbitration for Unsupervised Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing clustering algorithms rely on user input for the number of clusters, leading to suboptimal groupings if the optimal number is unknown, and rely on external analysts for similarity determination, introducing subjectivity and missing similar data points due to two-point distance measures.
Innovation Solution
The implementation of tri-point arbitration, which uses an internal arbiter to determine similarity and automatically determine the number of clusters through a tri-point arbitration similarity matrix, allowing for unsupervised clustering and reducing subjectivity by evaluating data points based on distances to multiple arbiter points.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If traditional two-point distance measures are used for clustering, then the clustering process is simple to implement, but subjectivity is introduced and similar data points are missed
Solution Approach 1:
The patent introduces an internal arbiter data point as an intermediary to mediate the similarity determination between data points. Instead of directly comparing two points, the arbiter point serves as a reference to objectively determine whether two points are similar, eliminating subjectivity and improving measurement precision while maintaining operational simplicity
Solution Approach 2:
The system uses the data points themselves (through the arbiter mechanism) to determine their own similarity relationships without requiring external analyst input. The arbiter point is selected from the data set and automatically determines similarity based on distance metrics, enabling the system to self-evaluate similarity objectively
2Adaptability or versatility
If user input is required for the number of clusters, then the clustering algorithm can be controlled, but suboptimal groupings occur when the optimal number is unknown
Solution Approach 1:
The patent implements a feedback mechanism where the internal arbiter continuously evaluates similarity relationships and provides information about the data structure. This feedback enables the algorithm to automatically determine when optimal clustering is achieved without requiring external specification of the number of clusters, improving reliability while maintaining adaptability through the arbiter-driven decision process
Solution Approach 2:
The clustering algorithm determines the optimal number of clusters autonomously using the internal arbiter mechanism. The arbiter evaluates similarity relationships and automatically identifies when clustering should stop, eliminating the need for user input about the number of clusters and preventing suboptimal groupings while maintaining system controllability
3Reliability
If external analysts determine similarity, then domain expertise can be applied, but subjectivity and bias are introduced
Solution Approach 1:
The system replaces external analyst judgment with an internal arbiter mechanism that objectively determines similarity using distance metrics. The arbiter point, selected from the data set, automatically evaluates similarity relationships without human intervention, eliminating subjectivity and bias while maintaining reliability through the systematic application of distance-based criteria
Solution Approach 2:
The patent replaces the mechanical process of external analyst evaluation with an automated computational mechanism. The internal arbiter uses mathematical distance calculations to determine similarity, substituting human judgment with an objective, reproducible computational process that eliminates subjectivity while preserving domain knowledge through the data-driven approach
Data Source
AI summary
Techniques are described for performing cluster analysis on a set of data points using tri-point arbitration. In one embodiment, a first cluster that includes a set of data points is generated within volatile and/or non-volatile storage of a computing device. A set of tri-point arbitration similarity values are computed where each similarity value in the set of similarity values corresponds to a respective data point pair and is computed based, at least in part, on a distance between the respective data point pair and a set of one or more arbiter data points. The first cluster is partitioned within volatile and/or non-volatile storage of the computing device into a set of two or more clusters. A determination is made, based at least in part on the set of similarity values in the tri-arbitration similarity matrix, whether to continue partitioning the set of data points.