Research And Implementation Of High-Dimensional Data Visualization Based On Dimesnion Reduction Techniques

Data is an important form of human recording information,and visualization is a kind of technology to convey information by using some of the more intuitive ways such as graphic images.Visualization makes it more efficient for human to obtain knowledge and it is an important way for human to access information.With the advent of the information age and the explosive growth of data,the data structure is becoming more and more complex and the dimension of the data is very high.High-dimensional data visualization is to be a difficult and hot problem in today's field of visualization.In this paper,we focus on the visualization of high-dimensional data and help users to discover the relationship between data object and the relationship between data and dimensions.The main research work is as follows:(1)An approach of visual exploration high-dimensional data is introduced.Due to the curse of dimensionality and the limited view space,it is difficult for users to visualize and explore high-dimensional data.Most of the earlier works create latent dimensions by using statistical methods.However,these methods have no good interpretability with user knowledge.So here we introduce an approach which derives dimensions by integrating user limited knowledge to reorganize data,and use the scatter-pie plot matrix which is an extension of scatterplot matrix to visualize and explore the reorganized data.It enables users to discover the relationship between the known and unknown data,and the relationship between the unknown data with the derived dimensions.Experiments on a real data set validate the effectiveness of our method.(2)A high-dimensional data visualization tool is designed and implemented.With the comparative analysis of visualization tools,we find that there are few high-dimensional data visualization tools.The existing high-dimensional data visualization tools cannot provide users a friendly process of visual exploration and not easy to extend new high-dimensional data visualization methods.Therefore,with the urgent need of a friendly high-dimensional data visualization tool to help users explore and analyze high-dimensional data efficiently.In this paper,we design and implement a high-dimensional data visualization tool.It provides users a more friendly visual exploration process.Users can explore data by interactions and save the data exploration results so that they can share with others or archive the results.Moreover,users can extend new visualization methods based on this tool for specific applications.