To deal with the problems of time-space imbalance between supply and demand of shared bicycle stops, obvious tidal effect and low level of optimization selection, by combining region coding with classification learning method, an optimal selection method of shared bicycle stops is proposed based on GeoHash and KNN. Firstly, based on the analysis of the existing shared bicycle parking problem, the spatial distribution of parking points is obtained by using the region coding segmentation of GeoHash algorithm. Then, under the double constraints of the number of stops and the bicycles that can be accommodated, the KNN clustering algorithm is used for secondary division to complete the optimal selection of shared bicycle stops. Finally, a comprehensive evaluation is carried out based on the shared bicycle data of Siming District and Huli District in Xiamen. The result shows that this method is reasonable, which can provide a useful reference for alleviating the tide phenomenon of shared bicycle stops.

Drawing on the theory of elaboration likelihood model and service quality, a real data set from an online healthcare community is used and an empirical study of patients' medical choice behavior and the moderating effect of disease risk under different paths provided. The results show that, overall, both factors from central and peripheral route affect patients' behavioral intention of online medical choice. Patients with low disease risks are mainly influenced by factors from the peripheral route including doctors' service reputation and disclosure of service qualifications. Whereas, patients with high disease risks are mainly influenced by factors from the central route including service quality of treatment and reviews quality. The results provide a theoretical basis for improving online medical services, and provide management inspiration for medical service providers.

As the recommender system of co-purchase network is more and more widely used, there are certain limitations in studying the economic significance of co-purchase link only based on the endogenous structure variable of network. Therefore, the exogenous variable of network word-of-mouth is added to make a more comprehensive analysis. The exponential random graph model in the social network method is used to build the model, mainly focusing on the four factors of product sales volume, product indegree, poor evaluation rate and the number of comments, to explore its influence on the formation of co-purchase links in the co-purchase network. Among them, sales volume, product indegree and comments have a positive correlation with the formation of co-purchase links, but the rate of poor reviews will significantly reduce the probability of co-purchase. The co-purchase network constructed by the exponential random graph model provides a reference for the optimization design of online e-commerce platform management network word-of-mouth and recommendation system.

In order to study the optimal strategy and coordination of three-layer e-commerce supply chain based on F2F with the market demand affected by online retail price, service level of e-platform and freshness, a Stackelberg Game model consisting of a fresh product manufacturers, TPL(Third Party Logistics service provider) and e-platform is constructed to analyze the optimal decisions and maximum profits of all parties. Then the rebate and revenue sharing contract is designed to promote the coordination of the supply chain. The results show that when the rebate and revenue sharing ratio meet certain conditions, all parties in the supply chain can achieve a win-win situation. Finally, a numerical example is given to verify the effectiveness of the contract and to analyze the influence of parameters such as price elasticity coefficient, elasticity coefficient of freshness, elasticity coefficient of service level, product preservation cost coefficient, service level cost coefficient on contract acceptance. The research results can provide a reference for the decision-making and coordination of the three-level supply chain of fresh e-commerce under the direct selling mode of household agricultural products.

The research on the evolution of public opinions in public emergencies based on stance detection and topic mining can help the government and stakeholders quickly grasp the evolution law of online public opinions in public emergencies, which is of great significance. The public opinion life cycle of specific public emergencies is divided, and a new stance detection model and a topic sentence mining method are proposed. For each life cycle stage, on the basis of identifying the stance information of mass Internet users, the efficient public opinion information is screened out, and then the topics of efficient public opinion information are mined, so as to deeply analyze the evolution law of the topic information of public emergencies. Taking the microblog data of "A Hangzhou Woman Missing Case" as an example, firstly, the experimental results of this method and various methods are compared to verify the effectiveness of this method.Then, the evolution of public opinions is analyzed based on the experimental results of this method, which proves that this method can quickly focus on the key points in the actual public opinions of public emergencies, and better analyze the evolution law and characteristics of public opinions of public emergencies. This method can effectively and accurately identify and analyze the content of public opinions, and provide a new perspective for the study of the evolution of network public opinions.

The combination of online and offline teaching mode is a trend of teaching in the future. The previous researches show that learners’ learning behavior can directly affect the learning results, so the research focuses on studying the influence level of learning behavior on academic performance. For the current common evaluation models, there are still some problems such as low credibility and weak interpretability. In this study, the Bayesian network (BN) based on evidential reasoning can effectively solve this problem. Compared with the commonly used machine model and deep learning model, it shows a lower error rate and a stronger interpretability.

The reconfigurable intelligent surface (RIS) is an effective technique that can improve wireless signal propagation environment and enhance communication security by adjusting the reflected signal through its controlled reflecting elements. The RIS-assisted secure communication in a non-orthogonal multiple access system is considered, where multiple users simultaneously send confidential information to a single-antenna base station in a non-orthogonal multiple access way in the presence of a single-antenna eavesdropper. Our objective is to maximize the total security rate of the system by optimizing the transmit power of the users and the phase shifts of the reflecting elements of the RIS while satisfying the constraints of transmit power and the quality-of-service for each user, as well as the constant modulus constraint of each reflecting element of the RIS. Although the considered optimization problem is non-convex and difficult to solve, an efficient algorithm based on the alternate optimization and semidefinite relaxation techniques is proposed to obtain a near-optimal solution. Simulation results validate that as compared with two benchmark schemes, the proposed algorithm significantly improves the total security rate of the system. It provides a reference for the secure communication research of reconfigurable intelligent surface assisted non orthogonal multiple access system.

Palmprint recognition, as a promising biometric recognition technology, has the advantages of high recognition accuracy and stable characteristics. Conventional hand-craft based palmprint recognition methods extract the principal lines and minutiae points depending on prior knowledge, which suffer from various limitations, such as the difficulties of scalability, feature extraction and information learning. To address this, a learning-based palmprint recognition method is proposed. Specifically, the Pixel Different Feature (PDV) feature is first extracted from palmprint image. Then, the binary representation and projection function based on PDV features is simultaneously learned via a cosine similarity preserving model. To alleviate the effects of noise such as palmprint position shifting and illumination changing, the binary representation is finally converted into a histogram descriptor. The experimental results on three widely used palmprint databases demonstrate that the proposed method can better excavate the intrinsic structure of palmprint images and significantly improve the performance of palmprint recognition.

The development of digital power grid puts forward higher requirements for distribution network planning. A distribution network reliability evaluation method based on graph-model integration technology is proposed, which can improve the efficiency of power network planners while improving the visualization and precision of distribution network reliability evaluation. This method is based on graph-mode integration technology, and a scalable vector graphics (SVG) map is proposed as a visual interactive object to map the planning content of the simulation on the feeder’s SVG map to the common information model (CIM). Based on the CIM, the distribution network topological model is built, and the feeder and load point reliability index solved using the fault diffusion traversal algorithm. Finally, by simulating the feeder planning transformation of a certain area, the reliability index changes before and after the feeder planning are given, showing that this method is effective and practicable. This method can effectively improve the efficiency of reliability evaluation of various distribution network planning projects by planners and provide important support for distribution network analysis, distribution network planning scheme evaluation, benefit evaluation of distribution self-healing project construction and so on.

There is not only white noise in GPS position time series, but also a lot of flicker noise. If only the covariance matrix of pure white noise is used to estimate the velocity, the velocity accuracy will be seriously overestimated. Spectrum analysis and maximum likelihood estimation methods are used to analyze the 20-year GPS coordinate time series of 264 GPS continuous stations of the Chinese Continental Crustal Movement Observation Network, and the velocity accuracy of each station calculated. The results show that there are different noise characteristics in E, N and U directions. The spectral exponents estimated by spectral analysis are higher than those estimated by maximum likelihood, but they are all around −1, which indicates that the main noise model is white noise plus flicker noise. The velocity accuracy calculated by different noise models is quite different. In the east, the north, and the vertical directions, the velocity accuracy calculated by the optimal noise model is, respectively, (11.5±2.1), (12.9±2.9), and (14.8±3.7) times that of the white noise model only. The velocity between maximum likelihood estimation method and FOGMEx method is not very different. In the east, the north, and the vertical directions, the velocity accuracy calculated by the maximum likelihood estimation method is (2.8±1.5), (1.5±0.7), and (3.5±2.8) times that of the FOGMEx method, respectively.

To solve the problem of poor monitoring results generated by time-series Interferometric Synthetic Aperture Radar (TS-InSAR) technology in low coherence areas, the traditional TS-InSAR method is improved from the perspective of polarization data. Taking Zhuhai Gaolan Port Economic Zone as the research subject, the article examines 47 dual-polarized Sentinel-1A satellite images with vertical-transmit-and-vertical-receive mode (VV) and vertical-transmit-and-horizontal-receive mode (VH) to experiment and obtain results and do the validation analysis. Results show that the monitoring quality of the time series InSAR technology can be effectively improved by applying dual-polarization data; and that the subsidence in this area is caused by geological conditions and reclamation projects. This research can provide a theoretical basis for the monitoring and prevention of regional subsidence.

Shadow play is a traditional folk-art form in China, and it has an important cultural heritage value. However, it is deeply affected by the impact of the new modern entertainment methods. On the road of modernization, shadow play presents a situation of weak performance on market, few successors, facing shrinkage or even extinction. Therefore, the modernization and innovation of the shadow play are important topics. A human-shadow play experience device is proposed based on Kinect. The shadow play can be controlled by human movements, thus increasing its interactivity and fun. Specifically, the novel device can capture human movements via Kinect and send commands to the steering engines installed in the shadow joints, thus controlling the movements of the shadow. Finally, the movement angles of steering engines and those of human left arm are examined, and the motion accuracy of the steering engine calculated. Results demonstrate that the accuracy of the device ranges from 96.6% to 99.8%, indicating the device can achieve high accuracy and sensitivity. The device is highly entertaining and is conducive to people's better understanding and inheritance of shadow play culture.

In recent years, China's economy and urbanization have developed rapidly, and the development of cities often comes at the expense of the environment. Regional air pollution dominated by PM_2.5 has become the most pressing and prominent environmental problem in China. According to relevant studies, PM_2.5 concentrations vary greatly in different seasons. Based on the PM_2.5 monthly mean concentration data of Guangzhou from 2015 to 2019, combined with atmospheric pollutants and meteorological factors, with the seasonal index introduced, an improved multiple linear regression and multi-layer perceptual combination prediction model for PM_2.5 concentration is established, to analyze the variation law and future development trend of PM_2.5 concentration in Guangzhou. The results show that the combined prediction model with the improved seasonal index is used to predict and analyze PM_2.5, and the fitting results are good. Compared with the multi-layer perceptron prediction model by using different evaluation indexes, the RMSE(Root Mean Square Error), MAPE(Mean Absolute Percentage Error) and MAE(Mean Absolute Error) of the combined model are reduced by 23.1%, 31% and 24.2% . Compared with the multiple linear regression model, the reduction is 35.3%, 41.3% and 41% . The accuracy of the model is better than the traditional multiple linear regression model and multi-layer perceptron model, which can better predict environmental PM_2.5 concentration and provide reference for optimizing the environment.

Economic link is an important form of interaction among cities, especially in urban agglomeration areas, and it is increasing with the development level of urban agglomerations. The present study takes the Pearl River Delta urban agglomeration as an empirical case, based on the improved gravity model method, GIS spatial analysis and QAP analysis, combines with social and economic data, analyzes the evolution characteristics of urban economic links in 2000~2019, and discusses the influencing factors. The main results are as follows: While the economic links of the Pearl River Delta urban agglomeration kept increasing constantly, Guangzhou has the largest increase in economic links with other cities, and Shenzhen has the largest increase; the direction of economic links variations in stability, and the relative position of cities in economic links are changing. Economic distance difference, economic development and urban population growth, industrial division and industrial links, and policy promotion are the main influencing factors of economic links. Among them, the difference in economic distance is the basic factor, and highly negatively correlated with the total amount of economic links; the speed of economic development and the urban population growth rate are the direct factors, and its growth is synchronized with the strengthening of economic links; industrial division and industrial links of labor are indirect factors, which have a positive correlation with changes in economic links; policy factors are important promoters and play an important role in influencing economic links. This study can help to clarify the evolution characteristics and influencing factors of the economic links in the Pearl River Delta urban agglomeration, and may have a certain reference for optimizing the urban links and promoting the improvement of urbanization quality.

The patio is an important element of traditional architecture in hot and humid regions to adapt to the climate, build the intention of human habitation, and organize the space. As a thermal buffer space connecting indoor and outdoor, the patio space of Lingnan traditional buildings has good climatic adaptability in terms of shading and ventilation, and its wind and thermal environmental characteristics are important for the restoration, conservation and optimization of traditional buildings. A typical Guangfu-style "three rooms and two corridors" traditional building in Chatang Village, Guangzhou City, is studied by using a combination of grid and local distribution methods, and a total of 32 measurement points are used to test the wind and thermal environment parameters in the interior of the patio, connecting corridors, compartments and halls under a typical summer weather background. The results show that the thermal pressure ventilation effect of patio space is the largest at 11:00 and 14:00. The fluctuation range of parameters in the vertical patio space decreases from high to low, and the wind and heat parameters in the height of 2.5 m and above are greatly affected by outdoor parameters. There is diagonal polarization in the vertical distribution of wind speed. The door opening and closing of open roadway has a great influence on the spatial wind field of the patio within 2.5 m, but has no obvious influence on the temperature field. Except for the hall, the wind-heat parameters of patio space are positively correlated with corridor space and wing space. The results provide basis for the study and preservation of traditional architecture in Guangfu.

Primary air chamber is one of the key components of circulating fluidized bed boiler, and the uniformity of primary air flow distribution is very important for the efficient operation of circulating fluidized bed boiler. The uneven distribution of primary air flow will have an adverse impact on boiler combustion, resulting in the reduction of boiler thermal efficiency. In order to improve the operation efficiency of circulating fluidized bed boiler, a numerical simulation method is used to study and design the appropriate air modification. Numerical simulation results show that after installing the deflector in the primary air chamber, the velocity nonuniformity at the cross section of the primary air chamber decreases from 81% to 24%. In order to verify the numerical simulation results, a small-scale geometric model of a 220 t/h circulating fluidized bed boiler is used to carry out visualization and velocity distribution uniformity experiments under five different experimental conditions. The experimental results confirm the correctness of the numerical simulation. In addition, the hot test is carried out on a 220 t/h circulating fluidized bed boiler. The results of hot test show that the thermal efficiency of the boiler increases from 85.71% to 88.34% after installing the deflector in the primary air chamber, which is equivalent to saving 5000 tons of standard coal every year. The economic benefit is remarkable, which proves the effectiveness of the deflector.

Polyimide (PI) is a polymer synthesized by thermal imination of dianhydride and diamine as raw materials. As a battery electrodes material, it boasts high theoretical capacity, high mechanical strength and easy recovery. However, its insulation limits the utilization rate of internal active sites, resulting in poor rate performance of the battery. A research review is conducted concerning the improvement of the structural stability through enlarging conjugate structure and the introduction of more carbonyl structure (C=O) to increase PI redox center site. In order to obtain higher battery capacity, PI, carbide and graphene combination, the hybridization of carbon nanotubes and the electrostatic spinning process are applied to improve electrochemical performance of battery electrode. The research of other materials based on PI is summarized as well. And the current research direction of PI is prospected.

A new type of tubular composite phase change material (t-CPCM) structure is designed to replace the traditional block-shaped composite phase change material (b-CPCM) structure, coupled with forced convection heat transfer and applied to battery thermal management. The simulation results show that, compared with the b-CPCM model, the t-CPCM model not only has a more uniform flow channel distribution, but also a larger convective heat transfer area. The theoretically calculated convective heat transfer thermal resistance is 0.8 K·W⁻¹, which is 20 times less than b-CPCM model. The experimental results show that the excellent heat dissipation performance of the t-CPCM battery module can effectively control the battery temperature. For example, the maximum temperature of the t-CPCM battery module is 46.9 ℃, and the temperature difference is only 0.8 ℃; as for the b-CPCM battery module, the highest temperature is as high as 51 ℃, and the temperature difference is 5 ℃. The designed t-CPCM has important application value in battery thermal management.