Lattice-Type Data Clustering Apparatus for Reducing Computational Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data clustering methods face challenges with large-scale and scattered data, leading to increased computational complexity, loss of feature information, and inconsistent clustering results, especially when dealing with multi-dimensional data from various sensors in systems like BEMS.
Innovation Solution
A data clustering apparatus and method using a lattice-type segmented space with indexes, allowing for efficient clustering by discriminating the input position of new data and creating or updating clusters based on representative values within lattice unit spaces, reducing computational quantity and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the number of clusters is reduced to decrease computational quantity, then computational complexity is reduced, but feature information of original data is lost
Solution Approach 1:
The patent segments the data space into a hierarchical structure with multiple levels. At each level, the space is divided into regions, and clusters are formed within these regions. This segmentation allows the system to process data in manageable portions while preserving local feature information, avoiding the need to reduce the overall number of clusters and thereby maintaining computational efficiency without losing feature information.
2Productivity
If existing clustering methods are used on scattered data, then clustering is performed, but inconsistent clustering results are obtained and re-clustering is required
Solution Approach 1:
The patent performs preliminary actions by pre-processing scattered data to identify and remove outliers before the main clustering process. This preliminary step ensures that the subsequent clustering operations work with clean, consistent data, preventing the need for re-clustering and ensuring reliable, consistent results throughout the process.
3Device complexity
If data dimension reducing technique is used to decrease computational quantity, then computational complexity is reduced, but data loss occurs and outlier data cannot be discriminated
Solution Approach 1:
The patent segments the high-dimensional data space into multiple lower-dimensional subspaces or regions. By processing data in these segmented regions rather than reducing the overall dimensionality, the system maintains the ability to detect outliers in the original feature space while performing computationally efficient clustering operations within each segment.
4Device complexity
If hierarchical algorithm such as K-D Tree is used to reduce distance computations, then computational quantity is reduced, but number of nodes to be searched increases and computation becomes slow
Solution Approach 1:
The patent segments the data space into a hierarchical structure where each level divides the space into manageable regions. This segmentation strategy reduces the number of nodes that need to be searched by organizing data spatially, allowing the algorithm to quickly eliminate large portions of the search space and find relevant clusters faster, thereby reducing both computational quantity and computation time.
Data Source
AI summary
Provided are a data clustering apparatus and method, which can rapidly and accurately cluster data. The data clustering apparatus includes an index discriminating unit discriminating an index corresponding to an input position of new data input to a space for data clustering, including a lattice-type segmented space having lattice unit spaces set with different indexes, and a clustering unit creating a new cluster in the discriminated index using the input new data as a representative value when a cluster is not created at the discriminated index.