2023, 12(1):4-16. DOI: 10.12146/j.issn.2095-3135.20220826001
Abstract:Under the “Internet ＋” environment, the integration of government big data with public services and social sensor data has transformed the urban management model from single to comprehensive, the urban service system from isolation to sharing, and the urban decision-making model from mechanical to intelligent, which is the development trend of government service. In order to solve the problems of insufficient integration of government affairs data, public service data and social sensing data, insufficient sharing, low utilization efficiency and lack of technical platforms, the overall framework of the “Internet ＋” government affairs big data intelligent service platform is proposed. This paper designs the platform construction theory, system architecture, technical architecture, core technology and service application, etc., and realizes the transparent access of the “Internet ＋” government big data intelligent service platform, city information unit, transparent intelligent agent, transparent management and microservices, etc. Exploratory applications have been carried out on platforms such as Digital Guangdong and Dandong Social Security in Liaoning, showing that the platform can effectively support the efficient sharing, management and utilization of government affairs data, public service data and social sensor data, and improve department service effectiveness and capability. The overall design scheme of the platform proposed can provide valuable reference for provincial and municipal governments to carry out the construction of “Internet ＋” government big data intelligent service.
2023, 12(1):17-25. DOI: 10.12146/j.issn.2095-3135.20220711001
Abstract:Government data sharing is an important foundation for promoting the construction of digital government and realizing the modernization of government’s governance system and capacity. In the past, the research on government data sharing focused on the establishment of data sharing management mechanism and the construction of technical support capacity of the platform, while the key to achieving efficient allocation of data value lies in the quality of supplied data and the guarantee of the sharing services. This paper summarizes the current operation mode of government data sharing service, and points out that the mode of “physical convergence and unified service” needs to pay attention to the possible security risks of high concentration of data, and there are some problems in the data sharing operation under the decentralized service mode, such as the demand of each department cannot be met quickly, the efficiency of data sharing is low, and the quality of data sharing service is poor. On this basis, a “joint operation” mode is proposed, which is jointly constructed by the data providers and the platform operators. This mode requires the data provider to be responsible for ensuring the quality of the data source, the platform operator to be responsible for the labeling and service capability of the data service, and both parties to jointly optimize the authorization mechanism and ensure the quality of shared services.
2023, 12(1):26-41. DOI: 10.12146/j.issn.2095-3135.20220730001
Abstract:Internet＋government affairs big data has the characteristics of cross-domain, multi-protocol and difficult convergence. In the process of big data collection and aggregation, there is a need for a variety of protocol conversions, which require a gateway to realize a unified protocol adaptation conversion and provide the data support for multi-source heterogeneous data aggregation and data fusion. Traditional protocol conversion methods are usually designed for specific protocol conversion requirements, and their scalability is poor, so they are not suitable for the situation where there are multiple protocol conversion requirements. By studying and analyzing the structure of protocol messages and the characteristics of protocol conversion, this paper proposes a method to construct the knowledge graph of the protocol conversion. By constructing the schema layer and instance layer of the graph, a knowledge graph of protocol conversion containing the structure of protocol messages and the mapping relationship between message fields is established. On this basis, a protocol conversion method based on the knowledge graph is proposed to realize the packet conversion between different protocols. The effectiveness of the proposed method is verified by an application example of protocol conversion and a comparison experiment with the existing protocol conversion methods.
2023, 12(1):42-55. DOI: 10.12146/j.issn.2095-3135.20220715004
Abstract:In order to recommend matters related to government services and improve user efficiency and government service level, a recommendation algorithm is proposed, that is, a collaborative filtering recommendation method for government services combined with user characteristics. Unlike traditional collaborative filtering which does not consider user attributes, this method combines user portrait technology with it. First, the method establishes a user portrait of government services, and then uses the singular value metric analysis method to integrate the user portrait and the user-based collaborative filtering algorithm, so that the feature attributes can participate in the similarity calculation, improve the similarity between users, and solve the problem of data sparsity. To make the results more practical, the method calculates the predicted government service score, and recommends the TOP-N with the highest score to the user. In the experimental part, the actual data of the government affairs service of a city’s enterprise legal person is used for verification. The results show that the algorithm can meet the personalized requirements of the government affairs service recommendations and improve the prediction accuracy.
2023, 12(1):56-67. DOI: 10.12146/j.issn.2095-3135.20220715002
Abstract:Methods such as BERT and the combination of neural network model have been gradually applied to the acquisition of disaster information. However, such methods have many problems, such as large number of parameters, inconsistent data sets and fine-tuning data sets, and local instability. In this paper, an information recognition model based on MacBERT and adversarial training is proposed. The model obtains the initial vector representation through MacBERT pre-training model, and then adds some perturbations to generate adversarial samples. Then input to the bi-directional long short-term memory and conditional random field in turn, which not only reduces the pre-training times and fine-tuning stage differences, but also improves the robustness of the model. The experimental results show that the information recognition model based on MacBERT and adversarial training are improved the accuracy rate and F1 value on the microblog dataset and the 1998 People’s Daily dataset, and the execution is excellent than other models, which indicates that the model has certain feasibility for urban waterlogging information recognition.
2023, 12(1):68-78. DOI: 10.12146/j.issn.2095-3135.20220901001
Abstract:The large-scale population mobility between cities reflects the correlation between cities to a certain extent. This paper selects the national population migration data during the Spring Festival from 2015 to 2018 in the Tencent migration database and uses Gephi software to conduct visualization analysis to obtain the strength of population connections between the cities as well as the evolutionary characteristics of the overall pattern of inter-connectivity strength. The DBSCAN algorithm is used to cluster the time series of cities across the country, and the commonalities and differences in urban mobility at different time nodes during the Spring Festival are discussed. And the distance attenuation effects of intra-provincial and inter-provincial population flow are analyzed from the improved gravity model. The results show that: (1) The national cities have formed a stable “diamond” structure as a whole, Beijing, Shanghai, Xi’an, and other cities have formed a single-axis spatial organization model, Guangdong and Shenzhen have formed a multi-axis spatial organization model, and the first-tier cities have strong centrality. In addition to cities with a “diamond” structure, cities in the northwest region and cities in the northeast region have strong advantages in terms of transit and acceptance capabilities. (2) The changing trend of net migration population during the Spring Festival was relatively stable, and a symmetrical inflow and outflow were formed before and after the Spring Festival. (3) The flow during the Spring Festival travel period conforms to the law of distance attenuation. With the increase in distance, the attractiveness of the city gradually decreases.
2023, 12(1):79-90. DOI: 10.12146/j.issn.2095-3135.20220909001
Abstract:The current method of dividing commuting groups takes less into account the time continuity characteristics of commuting trips. Based on the one-week subway card swipe data in Shanghai, this paper constructs a work-life recognition model for commuters, defines a commuting trip time similarity calculation method, and then extracts the features to classify commuter groups hierarchically, and uses the hot spot analysis model to perform spatial analysis and visual expression for spatial analysis and visualization, and explores the spatiotemporal regularity of commuters and the spatial distribution characteristics of work-housing organization characteristics in Shanghai. The results show that: (1) The employment single center model is obvious, and the employment hotspots of different clusters are distributed in the city center, and the settlements are characterized by the spatial organization of “hot in the west and cold in the east”. (2) The mainstream commute hours are 7:00—8:30 and 17:00—19:00, with nearly half of the commuters commuting during the morning rush hours and 90% leaving the work places before 19:30. (3) The travel time characteristics of the different commuting types are generally consistent with the distribution of their work and housing hotspots. The proposed research method reveals that the travel time law of commuters has a strong correlation with the spatial distribution of work-housing hotspots, which provides reference information for urban operation management and urban planning.
2023, 12(1):91-104. DOI: 10.12146/j.issn.2095-3135.20220712001
Abstract:The traditional person re-identification methods are difficult to independently cope with the complex and diverse recognition tasks in the security scenario of smart city in practice. In order to meet the needs of multi-level person re-identification, the deep integration of person re-identification and multi-level urban information units is proposed. Existing models and attentions for person re-identification tasks only focus on learning the robust features while neglecting the difference between features of pairs. Diff attention module is proposed to guide the network to learn a more discriminative attention map based on the difference of feature vectors. Taking the diff attention module, diff attention framework which matches many backbone models is developed. Two training strategies: joint training and separate training are proposed. Compared with other person re-identification methods, these framework and strategies have achieved excellent performance on Market-1501, CUHK03, and MSMT17 datasets.