Density-Based Clustering Radius Search for Large-Scale Datasets

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Density-based clustering algorithms face challenges in parameter tuning, particularly in high-dimensional and large-scale datasets, leading to over-segmentation or under-segmentation, which affects the quality and interpretability of clustering results.

Innovation Solution

A modified ternary search algorithm is employed to efficiently tune density-based clustering algorithms by leveraging the near-unimodality of the neighborhood radius parameter, using initial bounds that are closer to the optimal value, reducing computational time and resources while achieving the desired number of clusters.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional parameter tuning methods are used for density-based clustering, then the algorithm can handle arbitrary shapes and sizes, but the computational time and resources increase significantly

Engineering Contradiction:
Improveclustering qualityVSAvoidcomputational time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent applies preliminary action by performing ternary search to find optimal parameter bounds before executing the main density-based clustering algorithm. The system pre-calculates upper and lower bounds for the neighborhood radius parameter using sampled data, which narrows the search space and accelerates the subsequent clustering process while maintaining high clustering quality

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent utilizes parameter changes by dynamically adjusting the neighborhood radius parameter within calculated bounds during the ternary search process. The system modifies parameters such as sampling rates and parameter bounds based on data characteristics, enabling efficient adaptation to different datasets while reducing computational overhead

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If exhaustive search is used to find optimal parameter values, then clustering accuracy improves, but computational resources are excessively consumed

Engineering Contradiction:
Improveparameter tuning accuracyVSAvoidcomputational resources
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies segmentation by dividing the parameter search space into three equal segments using ternary search. Instead of exhaustively searching the entire parameter space, the algorithm segments it and systematically eliminates two-thirds of the search space in each iteration, achieving high parameter tuning accuracy with significantly reduced computational resources

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent uses copying by working with sampled copies of the dataset rather than the complete dataset during parameter tuning. The system creates sampled representations of the data to estimate optimal parameter bounds, which then guide the clustering process on the full dataset, reducing computational resource requirements while maintaining accuracy

Inventive Principle:
Principle #26Copying

3Productivity

If the neighborhood radius parameter is not properly tuned, then the algorithm runs faster, but over-segmentation or under-segmentation occurs

Engineering Contradiction:
Improveclustering speedVSAvoidclustering accuracy
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The patent implements feedback by using the results of sampled clustering runs to inform and adjust the parameter search bounds. The system evaluates clustering outcomes at different parameter values and uses this feedback to refine the upper and lower bounds, ensuring optimal parameter selection that balances clustering speed and accuracy

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent applies dynamics by making the parameter bounds adaptive rather than static. The upper and lower bounds for the neighborhood radius parameter are dynamically adjusted based on data characteristics and sampling results, allowing the algorithm to optimize both speed and accuracy for different datasets and clustering scenarios

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4657284A1Large-scale density-based clustering
Publication Date: 2025.12.03 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP4657284A1 patent drawingFigure 1
  • EP4657284A1 patent drawingFigure 2A
  • EP4657284A1 patent drawingFigure 2B~2C

AI summary

Systems and methods are provided for implementing large-scale density-based clustering functionalities. In examples, a system selects, for a dataset (which may be sampled at 100 % or less), an upper bound value and a lower bound value of a neighborhood radius parameter of a density-based clustering algorithm. The system identifies, using a modified ternary search algorithm, an optimal neighborhood radius parameter value, based on the upper and lower bound values, outputs the optimal neighborhood radius parameter value and/or a corresponding optimal number of clusters within the dataset. The modified ternary search algorithm leverages the near-unimodality of the neighborhood radius parameter, while selection of the upper bound value leverages a characteristic in which the neighborhood radius parameter value increases as the sampling rate decreases, and selection of the lower bound value uses ternary search that takes the number of clusters as a parameter instead of the neighborhood radius parameter.