Deep Neural Network Based Complex-Heterogeneous Cyberspace Cartographic Visualization

다국어 초록 (Multilingual Abstract)

It seems complicated to imagine our life without the Internet, but it is only a few decades old, and the Internet forms the most extensive cyberspace environment. Cyberspace is the non-physical environment created by joint computers inter-operating on a network. The complex heterogeneous cyberspace is a virtual computer world that encompasses a variety of computers, network devices, and systems that have been manufactured by different entities. As an essential spatial cognitive tool, the map has significantly contributed to human civilization for thousands of years. However, the cartographic elements of traditional maps are mostly geospatial entities or phenomena, and only some people apply them to draw abstract and virtual cyberspace resources, resulting in the development of cyberspace cartography lagging. Additionally, the process of data visualization involves the conversion of data into visual forms such as charts and graphs, with the aim of effectively conveying the data's importance. The cartographic visualization is to realize the visual expression of cyberspace, which is an essential basis for understanding cyberspace. Therefore, this paper introduces a deep neural network (DNN) to study complex-heterogeneous cyberspace cartographic visualization. At first, locally linear embedding is adopted to reduce the data dimensionality. Then, DNN is used to train the cartographic coordinates data obtained after data dimensionality reduction. Finally, several different data modeling methods are studied concerning temporal and spatial attributes to achieve complex-heterogeneous cyberspace cartographic visualization effectively. The simulation is powered by datasets sourced from cyberspace, which are accessible via data.world. The results of the simulation demonstrate that the suggested approach is superior in efficiency when compared to the baseline methods.