Bo Zhang

For a deep understanding of Beijing-Tianjin-Hebei collaborative innovation, we detected and visualized the communities of innovation in BTH Urban Agglomeration based on the patent cooperation network. China Patent Database was connected with Business Registration Database and the Tianyan Check to achieve the geographical information of organizational innovators. Spinglass algorithm was applied and ultimately 12 communities of innovation were detected. Based on the different structure characteristics, we further clustered the 12 communities into four typical structures that are hierarchical, single-center, polycentric, and flat structures. The hierarchical structure is usually large in scale and the cooperative intensity is relatively high. Single-center structure has a center with a high proportion of centrality and the cooperative intensity is relatively low. Polycentric structure has multiple centers with similar proportions of centrality. Flat structure is usually small in scale and has no obvious network center. In the patent cooperative network of BTH Urban Agglomeration, universities and state-owned enterprises occupied the centers and acted important roles to connect other organizations. Some communities of innovation showed significant industry characteristics, mainly involving six industry fields that are electric power, construction, petroleum, metallurgy and materials, municipal transportation, and railway. From the geographical perspective, some communities manifested local attributes and some demonstrated cooperation between regions. Beijing was the center of the Beijing-Tianjin-Hebei patent cooperation network. Compared with the pair of Beijing-Tianjin and the pair of Beijing-Hebei, Tianjin and Hebei were not closely connected. In the future, Beijing-Tianjin-Hebei collaborative innovation should strengthen cooperation between Tianjin and Hebei.

The application of machine learning touches all activities of human behavior such as computer network and routing packets in LAN. In the field of our research here, emphasis was placed on extracting weights that would affect the speed of the network's response and finding the best path, such as the number of nodes in the path and the congestion on each path, in addition to the cache used for each node. Therefore, the use of these elements in building the neural network is worthy, as is the exploitation of the feed forwarding and the backpropagation in the neural network in order to reach the best prediction for the best path. The goal of the proposed neural network is to minimize the network time delay within the optimization of the packet paths being addressed in this study. The shortest path is considered as the key issue in routing algorithm that can be carried out with real time of path computations. Exploiting the gaps in previous studies, which are represented in the lack of training of the system and the inaccurate prediction as a result of not taking into consideration the hidden layers' feedback, leads to great performance. This study aims to suggest an efficient algorithm that could help in selecting the shortest path to improve the existing methods using weights derived from packet ID and to change neural network iteration simultaneously. In this study, the design of the efficient neural network of appropriate output is discussed in detail including the principles of the network. The findings of the study revealed that exploiting the power of computational system to demonstrate computer simulation is really effective. It is also shown that the system achieved good results when training the neural network system to get 2.4% time delay with 5 nodes in local LAN. Besides, the results showed that the major features of the proposed model will be able to run in real time and are also adaptive to change with path topology.

The detection of moving targets is to detect the change area in a sequence of images and extract the moving targets from the background image. It is the basis. Whether the moving targets can be correctly detected and segmented has a huge impact on the subsequent work. Aiming at the problem of high failure rate in the detection of sports targets under complex backgrounds, this paper proposes a research on the design of an intelligent background differential model for training target monitoring. This paper proposes a background difference method based on RGB colour separation. The colour image is separated into independent RGB three-channel images, and the corresponding channels are subjected to the background difference operation to obtain the foreground image of each channel. In order to retain the difference of each channel, the information of the foreground images of the three channels is fused to obtain a complete foreground image. The feature of the edge detection is not affected by light; the foreground image is corrected. From the experimental results, the ordinary background difference method uses grey value processing, and some parts of the target with different colours but similar grey levels to the background cannot be extracted. However, the method in this paper can better solve the defect of misdetection. At the same time, compared with traditional methods, it also has a higher detection efficiency.