High-Dimensional Data Visualization via Evaluation Index Matrix
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data visualization techniques, such as principal component analysis and multidimensional scaling, struggle to effectively interpret and overview high-dimensional data structures, becoming complex as data volume and dimensions increase, and are limited in displaying attribute relationships in scatter plots.
Innovation Solution
A visualization device and method that calculates an evaluation index for attribute combinations in high-dimensional data, narrowing down characteristic structures and generating image information for visualization, using techniques like correlation coefficients or mutual information to present significant combinations in a tree view format.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If principal component analysis or multidimensional scaling is used to compress high-dimensional data into lower dimensions for visualization, then the data structure can be visualized in scatter plots, but it becomes difficult to interpret the visualized content and the plot becomes complicated as the number of data and dimensions increase
Solution Approach 1:
The patent segments high-dimensional data visualization into multiple two-dimensional scatter plots arranged in a matrix, where each plot displays relationships between specific attribute pairs. This segmentation allows users to focus on individual attribute relationships without being overwhelmed by the complexity of visualizing all dimensions simultaneously, thereby maintaining interpretability while handling high-dimensional data.
Solution Approach 2:
The patent uses the spatial arrangement of scatter plots in a matrix layout to represent the dimensional relationships of high-dimensional data. Instead of compressing all dimensions into a single low-dimensional plot, the system uses the two-dimensional matrix structure to organize multiple attribute pairs, effectively utilizing spatial dimensions to preserve information about attribute relationships.
2Adaptability or versatility
If scatter plots are used to display fixed attribute combinations (objective variable and certain explanatory variable), then the data distribution feature can be visualized, but the attributes to be displayed cannot be set for each node and data structures capable of being overviewed are limited
Solution Approach 1:
The patent implements dynamic attribute selection where users can freely choose different attribute pairs to display in scatter plots. The system allows changing the explanatory variable and objective variable combinations based on analysis needs, making the visualization adaptable to different analytical scenarios rather than being fixed to predetermined attribute combinations.
Solution Approach 2:
The scatter plot visualization system is designed to handle multiple attribute combinations universally. Each scatter plot can display any pair of attributes from the high-dimensional dataset, making the visualization tool versatile for exploring different data relationships and structures without being limited to specific predetermined combinations.
Data Source
AI summary
A visualization device includes: an evaluation index calculation unit 11 which calculates the value of an evaluation index representing a feature degree for each of combinations of a first attribute group and a second attribute group in terms of data on the second attribute group including one or more attributes and conditioned with the first attribute group including one or more attributes among high-dimensional data to be visualized; and a visualization processing unit 12 which generates image information for presenting combinations of the first attribute group and the second attribute group, which are determined to have large evaluation index values based on a predetermined criterion.