Clustering Framework Using User Preference Degrees
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional clustering techniques fail to accurately group objects based on user preferences, leading to inaccurate cluster formations.
Innovation Solution
A framework that incorporates user preference information to calculate preference degrees and similarity measures between objects, using a clustering system with an input module, preference degree module, similarity module, and clustering module to generate ordered clusters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional clustering techniques are used to group objects based on intrinsic characteristics and distances, then the clustering process is simple and objective, but the cluster formations are inaccurate according to user objectives
Solution Approach 1:
The patent introduces preference degrees as an intermediary parameter that mediates between user objectives and object characteristics. The preference degree module calculates this intermediary value by comparing user preferences with object attributes, allowing the clustering system to incorporate subjective user requirements without fundamentally redesigning the entire clustering architecture. This intermediary enables accurate cluster formations aligned with user objectives while maintaining relative system simplicity.
2Measurement precision
If user preference information is incorporated into clustering to achieve accurate groupings, then cluster formations align with user objectives, but the calculation and processing complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing preference degrees for all object pairs before the actual clustering process. The preference degree module processes user preference information and computes these values in advance, allowing the clustering module to directly utilize pre-computed preference degrees during cluster formation. This preliminary processing reduces the computational burden during clustering, minimizing overall processing time while maintaining high accuracy in aligning clusters with user objectives.
Data Source
AI summary
A framework for clustering is described herein. In accordance with one aspect, a data set having x number of objects, and preference information of a user is provided to a clustering tool. The clustering tool may calculate preference degrees between objects in the data set. In addition, similarity measures of objects in the data set may be calculated. Clusters of objects may then be generated from the data set.