|
A Visual Layout Algorithm for Showing Overlapping Community Structure of Networks
Collect
ZHANG Mingna, XIAO Jing, XU Xiaoke
Complex Systems and Complexity Science. 2023, 20 (4): 10-17.
DOI: 10.13306/j.1672-3813.2023.04.002
Visualization technology can be used to effectively analyze the community structure, help users to understand the network topology, and dig out the hidden information from it, but the traditional network visualization layout algorithm does not consider the overlapping characteristics of nodes in the overlapping communities. Aiming at above problems, firstly, this paper uses the existing overlapping community detection algorithm to divide the community, and then hard partition and determine the node position according to the weighted summation of the membership matrix. Finally, the overlapping nodes are accurately laid out to display the overlapping community structure. The visual results demonstrate that the algorithm can visually display the overlapping community structure, highlight the overlapping nodes and the communities to which the nodes belong. At the same time, the experiment uses indicators such as node deviation, edge length deviation, and point distribution variance to verify that the algorithm can reduce node deviation and point distribution variance. This study can meet the visualization needs of overlapping community structures in complex networks, and can provide certain help for understanding the structure and functions of complex communities.
References |
Related Articles |
Metrics
|
|
Complexity Analysis of Chinese Railway Express Freight Transportation Network
Collect
MA Liang, JIN Fucai, HU Chenhan
Complex Systems and Complexity Science. 2023, 20 (4): 26-32.
DOI: 10.13306/j.1672-3813.2023.04.004
In order to improve the accuracy of complex network model and the comprehensiveness of complexity analysis, the method of modeling undirected weighted complex network of CREFTN in L-space (UWCN-CREFTN), the comprehensive importance evaluation index of the nodes and network robustness analysis method considering node failure based on comprehensive importance index are proposed. Based on the timetable and wagon-flow data, the results of analyzing the topological characteristics of UWCN-CREFTN show that CREFTN is a small-world network which has high heterogeneity, strength-degree positive correlation, point-weighted equality, weighted non-homogeneous matching, strength-strength negative correlation at present. Based on network efficiency and node importance, the results of analyzing the robustness of UWCN-CREFTN contrastively under four situations at two aspects show that the random failure of nodes makes the network more robust than vandalism, maximum-comprehensive-importance vandalism make the network more vulnerable than maximum-strength vandalism and completely collapse faster than maximum-weighted-betweenness vandalism, which indicates that the proposed comprehensive importance evaluation index has the dual characteristics of node strength and node weighted betweenness.
References |
Related Articles |
Metrics
|
|
Controllability Analysis of InSCC Topology
Collect
XIAO Pengpeng, JI Zhijian, LIU Yungang, LIN Chong
Complex Systems and Complexity Science. 2023, 20 (4): 47-55.
DOI: 10.13306/j.1672-3813.2023.04.007
To study the controllability of a class of multi-agent systems, the concept of Input strongly connected component (InSCC) was first proposed, and analyzed using PBH criterion and graph theory. Firstly, the controllability of InSCC structure and the topology structure composed of InSCC and roadmap were analyzed, and a leader′s selection method was provided to achieve system controllability. Secondly, based on the InSCC structure, the influence of adding communication edges between different InSCC structures and in the road map on system controllability was studied. Research has found that for topologies containing InSCC, as well as those composed of InSCC and road maps, increasing communication edges in a certain way does not change the controllability of the system. A method for constructing controllable topologies was proposed. Finally, the necessary and sufficient conditions for the controllability of multi-agent systems containing InSCC structures under switching topology were given.
References |
Related Articles |
Metrics
|
|
Short Video User Behavior Prediction Algorithm Based on Multi-task and User Interest
Collect
GU Yiran, XU Zebin, YANG Haigen
Complex Systems and Complexity Science. 2023, 20 (4): 69-76.
DOI: 10.13306/j.1672-3813.2023.04.010
The user behavior of short video (such as viewing comments, likes, clicking on avatars, and forwarding) is predicted by considering the change of user interests. In this paper, the sorted user historical behavior sequence is introduced into word2vec as a corpus to train the word embedding model, learn the dynamic interests of users, and effectively capture the changes in user interests. The statistical features constructed by feature engineering and the user dynamic interest features constructed by the word embedding model are input into the multi task learning with multi gate mixture of experts (MMOE), and a new evaluation index W-uAUCis proposed to evaluate the prediction accuracy of the model. The experimental results show that compared with shared bottom, wide & deep and deepfm, the proposed MMOE model considering the change of user interest has the best prediction accuracy.
References |
Related Articles |
Metrics
|
|
Loop Closure Detection Based on Improved NetVLAD Image Feature Extraction
Collect
QIU Changbin, WANG Qingzhi, LIU Qipeng
Complex Systems and Complexity Science. 2023, 20 (4): 92-97.
DOI: 10.13306/j.1672-3813.2023.04.013
Traditional loop closure detection algorithms mostly utilize handcrafted features to represent images. The robustness of dealing environmental changes such as illumination and viewpoint changes is vulnerable, and the features extraction is time-consuming, which cannot meet the real-time requirements. To address these problems, we propose an improved algorithm which uses deep neural network to extract more robust image features. Specifically, the atrous spatial pyramid pooling (ASPP) module is introduced into the classic NetVLAD model to characterize the image. By the fusion of multi-scale features, the feature maps have fewer dimension and higher resolution, and thus, more accurate and compact image features are obtained. Experiments on public datasets show that the proposed algorithm has higher precision and recall rate. It can deal with the changes of illumination and viewpoint to a certain extent, and has less time cost in extracting image features.
References |
Related Articles |
Metrics
|
|