Data Clustering via Attribute Segmentation and Element-Chain Grouping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional clustering algorithms often generate clusters that may not be useful to users, as they typically offer a single breakdown of a dataset without considering various user perspectives, such as different sports, countries, or significance, and may not adapt to user interests without initial input.
Innovation Solution
Data factorization techniques that automatically generate multiple attribute sets and element-chain groups based on the most frequent elements in a dataset, using a trained model to calculate features and determine group scores, allowing users to select relevant clusters without initial input, providing multiple views of the data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional clustering algorithms generate a single breakdown of the dataset, then the clustering process is simple and fast, but the resulting clusters may not be useful to users and do not consider various user perspectives
Solution Approach 1:
The patent segments the clustering task into multiple independent attribute sets, where each attribute set represents a different perspective or view of the data. Instead of generating one monolithic clustering, the system creates multiple segmented clusterings based on different attributes (e.g., sport, country, significance), allowing users to explore data from various angles without requiring a single complex algorithm to handle all perspectives simultaneously.
Solution Approach 2:
The patent implements multi-functionality by creating a clustering system that serves multiple purposes through different attribute sets. Each attribute set acts as a universal interface that can be applied to the same dataset to produce different meaningful groupings, making the clustering system adaptable to diverse user needs and perspectives without requiring separate specialized algorithms for each view.
2Adaptability or versatility
If clustering algorithms require initial user input to define clustering criteria, then the clusters can be tailored to user interests, but the process becomes more complex and requires user interaction
Solution Approach 1:
The patent applies preliminary action by pre-defining multiple attribute sets that represent common user perspectives before the user initiates the clustering process. These attribute sets (such as sport, country, significance) are prepared in advance, so when a user wants to cluster data, they can immediately select from these pre-configured perspectives without needing to define clustering criteria from scratch, thus maintaining adaptability while simplifying operation.
3Loss of information
If clustering algorithms partition all data into clusters, then complete coverage of the dataset is achieved, but the result may not highlight potentially interesting subsets that users are most interested in
Solution Approach 1:
The patent applies local quality by allowing different attribute sets to have different levels of importance and relevance for different users or contexts. Each attribute set can be optimized for specific types of interesting subsets (e.g., one attribute set may be better at identifying sport-specific patterns while another excels at geographic patterns), enabling the system to highlight locally relevant interesting subsets rather than applying a uniform clustering approach to the entire dataset.
Data Source
AI summary
An example method is provided in according with one implementation of the present disclosure. The method comprises generating a group of most frequent elements in a dataset, calculating features of each of the most frequent elements in the dataset, applying a trained model to the features of each of the most frequent elements, and generating a list of predicted relevant elements from the list of most frequent elements. The method further comprises determining at least one element-chain group for each predicted relevant element and a group score for the element-chain-group, ordering a plurality of element-chain groups for the dataset based on the group score for each of the element-chain groups, and identifying a predetermined number of element-chain groups to be outputted to a user.