High-Dimensional Multivariate Visual Analysis Based On Data Of Latin American Countries

National research involves complex data,and it is of great significance to design a visualization system for national research.Based on the World Bank's 1441 data from 17 countries in Latin America and the Caribbean from 1960 to 2016,this paper studies the visualization methods of high-dimensional multivariate data,and designs a national visual analysis model and visual analysis system.The main work of this article is mainly reflected in the following three aspects:1.Construct multi-level features of the attributes of Latin American country data.Too many attributes of high-dimensional multivariate data will cause the burden of manual screening tasks.Therefore,the text classification is used to hierarchically construct the attributes of the data set,which is convenient for users to quickly filter the data subset and provide an overview of the entire dimension of the data.2.Construct a dimensional space exploration model for Latin American countries.Aiming at the problem that high-dimensional data contains important information and will be disturbed by irrelevant information,the exploration of the data dimension space is completed from the perspective of attributes and data objects.First of all,three methods(feature selection,subspace exploration,and subspace clustering algorithm)of high-dimensional multivariate data processing based on subspace clustering are used to analyze high-dimensional multivariate data from the perspective of attributes.After using the feature selection method to select a suitable feature subset in the original dimensional space,the subspace clustering algorithm is used to cluster the feature subsets,and then the subspace exploration method is combined with visual analysis to assist users in discovering new interest subgroups.space.Then,in the reconstructed local subspace,from the perspective of the data object,the data is reduced by dimension-based fast clustering to observe the distribution characteristics of the data.3.Designed and implemented a high-dimensional multivariate visual analysis system based on data from Latin American countries.Various visual means such as polar coordinates and visual metaphors are used to design views such as individual details and projections of important attributes.The system realizes the function of screening and exploring the data of Latin American countries from the whole to the individual,from the macro to the characteristic events.It can also compare and show the time-varying evolution of multiple attributes in multiple countries.Finally,the effectiveness and applicability of the system were verified through four case studies.In addition,Latin American researchers were invited to conduct a preliminary evaluation of the system functions,visualization methods,and interactive technologies.The evaluation showed that the design of the visual analysis system was novel and visualized.It is very effective and easy to interact with,which is of great help to the research of Latin American countries.