Three-dimensional data visualization method based on PCA-Radviz

Abstract

The invention provides a three-dimensional data visualization method based on PCA-Radviz. The method comprises the following steps: 1) carrying out standardization processing on original data; 2) performing principal component analysis (PCA) on the processed data to obtain a plane formed by taking a two-dimensional vector with the maximum contribution degree as an x axis and a y axis; 3) respectively solving cosine distances between the original dimension of the data and the x axis and the y axis to obtain a projection angle and a vector length of the original dimension on the plane; 4) designing a reasonable dimension anchor point layout for the Radviz circumference by utilizing the angle obtained in the previous step; 5) projecting the data sample into the Radviz unit circle in a dot mode; and 6) expanding the z-axis, and pulling the two-dimensional plane point into the three-dimensional space. Data from different types of data samples are displayed based on PCA-Radviz,, so that dataclustering information as much as possible is provided for a user, a better data visualization effect is achieved, and the user can make a decision more quickly, visually and accurately.

Description

technical field [0001] The invention belongs to the technical field of computer information processing, and relates to a three-dimensional data visualization method based on PCA-Radviz. Background technique [0002] With the advent of the era of big data, data mining has become a popular research direction of data analysis. The analysis methods it contains are fully automated, efficient and economical, but due to the lack of a correction process of human-computer interaction, their accuracy is usually lower than the results of manual analysis. Data visualization makes up for this shortcoming and improves human cognitive ability. Radviz (Radial Coordinate Visualization) is a visualization algorithm based on the spring model. It provides a special way of thinking to observe the relationship between data, and can study high-dimensional data from a global perspective. [0003] The main idea of ​​the classic Radviz algorithm is to use a nonlinear method to map points with high...

AI Technical Summary

Problems solved by technology

This algorithm has the advantages of low computational complexity and strong visualization effect, but it has certain limitations: 1. It will weaken or even eliminate some clustering information between data in the original n-dimensional space; 2. The nonlinear mapping of the algorithm leads to the final The result of the mapping is a many-to-one relationship

These two limitations weaken the final visualization effect of the algorithm to a certain extent, making it impossible to achieve the optimal

Images