Visualization Method And System Development Of Hierarchical High-dimensional Data

With the 21 st century entering the era of big data,data explosion is a major challenge in the current information science field.In the era of big data,the complexity of data to be processed is getting higher and higher.Hierarchical high-dimensional data is a mixed feature data set that can be seen everywhere in daily production and life,and information visualization is an effective method to help visual mining.At present,most of the visualization methods are only applicable to data with a single data feature,and there are few visual studies on hierarchical high-dimensional data with multiple data characteristics.How to visualize high-dimensional data and quickly and accurately mine useful information to provide decision-making assistance to decision-makers is a topic worth studying.The main work and results of this thesis are as follows:1.For the visualization of hierarchical high-dimensional data,firstly,a VPM(Venn Parallel Coordinates Mixing)hybrid visual layout method based on circular nested graph and parallel coordinates is proposed.The circular nested graph based on D3 layout algorithm is adopted to represent hierarchy.According to the idea of parallel coordinates,the circumference of the leaf nodes is divided into different attribute axes.Visualization of high-dimensional attributes in hierarchical structures is achieved through user interaction such as data preprocessing,layout design,attribute mapping,attribute point join,scaling,and highlighting.Then,based on the visual analysis task of the teacher's performance,the teacher evaluation model is established,and the visualization components such as the word cloud view,the sorting view,the performance composition ratio view,the time series view,and the cluster view are designed to visually analyze the data from multiple views.Finally,through experimental verification,the user is allowed to analyze the teacher development model from interactive exploration.2.In order to solve the problem that two-dimensional visualization easily overflows the screen,a three-dimensional visual layout based on tree structure and adaptive hierarchy is proposed.Combining LOD technology with interactive methods such as character roaming and camera follow-up,the correspondence between model and data is established.The tree structure represents hierarchical information,and high-dimensional attribute information is presented through three-dimensional visual interaction.Firstly,the modeling of leaves and branches is carried out in Maya,and the 3D scenes are built in Unity3 D.LOD technology is used to draw different levels of the scene from different viewpoint positions,which can speed up the scene rendering and improve the visualization efficiency.At different LOD levels,the high-dimensional attribute information in the teacher performance data is displayed through the interactive method of the tag,and the user is allowed to observe the data from different views in the form of three-dimensional roaming,and the improved K-means clustering algorithm is used for college classification.Finally,the results of the college's performance for the next year are predicted through existing data sets,and compared with the actual data to prove the accuracy of the forecast.3.Design and implement a two-dimensional visual analysis system based on B/S mode and a three-dimensional visualization scene.Front page of the 2D visualization system exposes the various visualization components described above for user interaction.The 3D system uses Unity3 D as the development platform and uses the LOD algorithm to speed up the real-time generation of the scene.We propose two visual layout methods for displaying hierarchical high-dimensional data.The two-dimensional visual layout solves the problem that the previous attribute dimension cannot exceed four dimensions.Compared to two-dimension,the three-dimensional visual layout can visualize more data and support user-aware roaming.This article uses the teacher performance data for experimental demonstration,which can help system users-college administrators to well manage teachers and performance evaluation.