[Purpose/Significance] The development of the Internet and digital economy has brought new opportunities for the development of digital marketing. At the same time, digital marketing is the implementer and forerunner of digital economy. Digital marketing is an important way to promote the digital transformation of enterprises. Therefore, the research of this paper has certain significance for the development of digital economy and digital marketing as well as the digital transformation of enterprises. [Method/Process] In line with the development direction and policy of digital marketing, this paper analyzes the policy direction of digital marketing and reviews and summarizes the enlightenment of the policy. Combined with the development process of the Internet technology, it visually reviews the evolution of digital marketing practice, providing reference for digital marketing research and guidance for the practice of digital marketing. Following the path of user-preference-evaluation-promotion, the research emphases of digital marketing were analyzed and reviewed. [Results/Conclusions] In terms of policy direction, the policy enlightenment of digital marketing can be summarized as follows: (1) Digital marketing should serve the national strategy and accelerate the construction and development of digital economy. (2) Data security belongs to the national strategy, and privacy security is the basis of digital marketing. (3) Connectivity builds a new ecology and new way of digital marketing, and the whole-area marketing reshapes the brand growth. (4) It is suggested to vigorously develop digital marketing to promote the rise of domestic products. In terms of practical development, the practical development of digital marketing is divided into four stages: one-way marketing, interactive marketing, precision marketing and smart marketing. The practical development of digital marketing can be summarized as follows: (1) The application scope of digital marketing is more extensive, involving all aspects of market behavior, such as medical industry, supply chain, and agricultural products. (2) The current focus of digital marketing has changed from the original dissemination of customer acquisition to user operation, the whole process of interaction, user experience, etc. (3) Data can drive marketing. The core of digital marketing is the analysis and mining of data. The current digital marketing practice focuses more on the collection and analysis of user data and the generation of marketing strategies and behaviors driven by data. In terms of research focus, it can be divided into the research of digital community user group, digital community user preference, consumption potential evaluation and development, and digital marketing effectiveness evaluation.

[Purpose/Significance] Consumption is an important aspect of stimulating domestic demand. In the current development stage, the spatial and geographical distribution characteristics of our country indicate that domestic demand has great potential. However, the evaluation of consumption potential in the digital economy is a complex systematic project. Current research does not fully consider the correlation between evaluation indicators, and there are some problems such as low efficiency and too much reliance on the subjective experience of experts, which affect the scientific and reasonable evaluation results of consumption potential. [Method/Process] First of all, starting from each major work link that affects the consumption potential under the background of the digital economy, the preliminary evaluation index was determined through expert consultation and literature research and screening, and the primary evaluation index set of consumption potential under the background of the digital economy wasconstructed. Based on the consumption potential assessment data of Guangxi Zhongyan Industry Co., Ltd. from January to December 2021, the correlation and association rules among the assessment indicators were analyzed by using Apriori association rule mining algorithm, to streamline the assessment indicators and build an assessment index system of consumption potential in the context of the digital economy. The index system includes 11 secondary indicators and 19 tertiary indicators, such as user value, marketing communication effect, marketing ability, user participation, user influence, and channel effectiveness. Combining the influence relationship and influence degree between the quantitative evaluation indexes of DEMATEL and ANP, the weights of the evaluation indexes were calculated by the subjective and objective combination weighting method, to construct a multi-modal evaluation model of consumption potential in the digital economy. [Results/Conclusions] Taking six real dragon cigarette product marketing activities of Guangxi China Tobacco Industry Co., Ltd. from March to September 2021 as examples, the comprehensive scores of marketing activities in March, April, May, and June are 0.53, 0.52, 0.51 and 0.48, respectively, and the overall assessment result of consumption potential is "medium" level, among which the effect of marketing activities in August and September is better than that in March and June, and the comprehensive scores are 0.57 and 0.63, which are "medium to upper" level. On the whole, the impact of market influence ability, user value and marketing communication effect on the release of consumption potential decreases in turn. The analysis results are in good agreement with the actual situation, which verifies the feasibility and applicability of the multi-modal assessment model of consumption potential. In addition, due to the limited space, this paper only takes the tobacco industry as an example. In our future work, the specific index system can be adjusted according to the actual application needs.

[Purpose/Significance] Multi-modal feature analysis and service efficiency improvement of digital community consumers will help to provide a new vision for the construction of digital intelligent online communities and provide new impetus for relevant departments to make decisions. In addition, although the current research on digital consumption includes the relevant content of user analysis, it mainly aims at the formulation of detailed operation plans, and lacks the analysis of service efficiency improvement of digital communities. On the other hand, the research on user value orientation for online service quality optimization is mostly based on profile technology, which only considers the difference characteristics of a single target user, and lacks the horizontal comparison and difference attribution research of multi-modal features among groups. Based on this, this paper, from the perspective of value discovery, achieves clustering by aggregating user profiles, analyzes the multi-modal characteristics of consumer groups in digital communities, and proposes a service efficiency improvement scheme. [Method/process] First, this paper analyzed the target consumers in the digital community and established a cluster indicator system. Then, users were grouped, and the multi-modal information profile of the target group was restored based on group characteristics and inter-group interaction characteristics. Finally, it proposed the path to improve the efficiency of digital community services. In terms of technical implementation, the data related to consumer activities were extracted from the digital community, integrated, cleaned, and distributed to the storage bucket. The clustering indicator system was built through feature mining and existing indicators, and the indicators were mapped to aims, and DBSCAN clustering was carried out on the basis of using AP to realize the image. After grouping and naming, the characteristics analysis, interaction analysis, and drift and penetration phenomenon analysis were carried out according to the characteristics of various groups. We extracted various parameters of the design of digital community consumption activities, and built a decision variable function to find the optimal behavior equilibrium conditions of the digital product supplier, consumer and digital community. Based on this, we built an efficiency improvement tree, and proposed community service efficiency improvement strategies at the initial, middle and later stages of consumption activities. [Results/Conclusions] The empirical analysis results show that the model in this paper can first generate reasonable and effective clustering results, and then realize the classification of group characteristics and the analysis of inter-group infiltration and drift. The clustering results show six types of consumer groups: focus, center, special, sleeping, loss and general groups. Most groups will have user penetration, and only general user groups will have inter-group drift. The service efficiency improvement model shows that the most valued group is the center and key group. The inadequacy of this study is that the applicability of the model to multi-source heterogeneous data needs to be tested and there is still room for improvement in clustering granularity.

[Purpose/Significance] Digital communities have become a way for enterprises to manage users efficiently. The existing research on digital community rarely considers the importance of user behavior information and user's customer life cycle value to the mining of user preferences in digital community. This research aims to give full play to the digital community's characteristics such as intuitive, convenient, interesting, and interactive properties so that the research results can benefit every user in their use of the digital community and every enterprise in their user management. [Method/Process] Aiming at the user groups in digital community, this paper proposes a preference mining model ClV-Preference mining (CLV-PM) based on Customer Lifetime Value (CLV). First, in order to reflect the real preferences of users, the three indicators of the RFM model are used to quantify user behavior information, and the group characteristics of users are mined through K-mean ++ algorithm to generate user value category labels. Second, in order to consider the timeliness and difference of users and enhance the model's cognition of preferences, this paper uses the entropy weight method to solve the indicator weights of each activity, obtains user CLV to construct user-project scoring matrix, and uses the collaborative filtering algorithm to predict user preferences. Finally, based on the user value category, user historical preference and user forecast preference, the user preference profile of target users in digital community is generated, and feasible suggestions are put forward for the operation and maintenance of target users according to the user preference profile. [Results/Conclusions] The user dataset of the "Wechat community" management platform can be divided into four user value categories: important value users, low value users, returned users and important retention users. Target users 16254 are important value users, and the operation strategy of "retention and maintenance" is adopted. The historical preferences are happy hop, sec-kill and other activities; the prediction preference is flying chess battle, guessing code map and other activities; the target user preference sketch provides the basis for the operation and maintenance of users in the digital community. In terms of data source, the CLV-PM model proposed in this paper directly reflects user preferences based on user behavior information and reduces the problem of score distortion. To provide a new perspective for the research of user behavior in digital community, the construction of user-project scoring matrix based on user CLV fully considers the user value of digital community and provides a new direction for the extension and application of CLV. However, due to limited research space, this paper did not conduct model evaluation research on the proposed model, which can be further discussed in subsequent studies.

[Purpose/Significance] Digital live streaming is an important way for enterprises to carry out online activities. Incorporating clear set qualitative comparative analysis into the research on improving the effectiveness of enterprise digital live streaming activities is an innovation in the application of configuration analysis methods and a new breakthrough in the field of improving activity effectiveness. By identifying elements or combinations of elements that can create greater value for enterprise digital live streaming activities, and adopting various ways to enhance these element combinations, the goal of improving the effectiveness of enterprise digital live streaming activities can be achieved, which has innovative guiding significance for enterprise development. [Method/Process] Based on the consideration of sample homogeneity, the background data of the digital live streaming activity of Guangxi Zhongyan Industry Co., Ltd. was used as the data source. Four influencing factors related to the digital live streaming activity were selected as conditional variables using literature induction and problem oriented methods, and some conditional variables were assigned values using linear discriminant dimensionality reduction. At the same time, the activity effect was used as the outcome variable, configuration analysis was conducted on digital live streaming activities in the tobacco industry through clear set qualitative comparative analysis, and combined with SOR model and the 4I theory, a configuration path for improving the effectiveness was generated. [Conclusions/Results] Research has found that a single factor does not constitute a necessary condition for improving the effectiveness of digital live streaming activities. There are two configuration paths: game assisted and topic supported, which can improve the effectiveness of digital live streaming activities. The mechanism of action can be summarized as follows: rich topic types are conducive to meeting users' curiosity psychology, topic popularity and game quantity can ensure users' participation in live streaming, and user interaction work is the basic guarantee for improving user retention. Meanwhile, these two paths can be applied to different forms of digital live streaming, and enterprises should choose according to their own needs and refine the paths in practice. In addition, although this article explores the key paths to improve the effectiveness of digital live streaming activities, due to limitations in research samples and industry fields, it has not fully revealed the various factors that affect the effectiveness of digital live streaming, and there is a certain degree of subjectivity in the criteria for assigning variables. Therefore, further revisions and improvements are needed in future research.

[Purpose/Significance] China has accelerated the construction of new infrastructure such as 5G networks and data centers, and included "smart library" in the 14th Five Year Plan and the Vision Outline for 2035, bringing significant development opportunities to the construction of smart libraries. "No one can create a smart library except smart librarians". Smart librarians are the source of vitality and power for the development of smart libraries. The construction of smart libraries puts forward higher requirements for the quality and ability of smart librarians. Developing the core elements of smart librarians is the focus of smart library construction, and also the primary task of smart library construction. [Method/Process] The research focus of the academic community on smart librarians is mostly focused on the research on the capacity building and capacity evaluation system of smart librarians. According to the connotation of the concept of core literacy, combined with the characteristics of smart libraries and the working characteristics of smart librarians, this paper puts forward four core librarian literacy elements of smart libraries: professional literacy, information service literacy, cooperation and communication literacy, and self-development literacy. Among them, professional literacy is the basis for the effective operation of smart libraries; information service literacy is the guarantee of the service quality of smart libraries; the quality of cooperation and communication is the condition of open sharing and cooperation of smart libraries; self development literacy is the guarantee for the sustainable development of smart libraries. On this basis, the development path of the core literacy of smart librarians is given in a targeted way. The paper contributes to the literature by providing the first case study on the cultivation of smart librarians. [Results/Conclusions] Developing the core literacy of smart librarians is a dynamic, complex and systematic project, involving the coordinated development of pre-service education, post-service training, self-learning and other aspects. By adopting multiple training methods, creating a "meaning construction" environment, building a community of practice, establishing a progressive mechanism and other training measures, we will be able to improve the core quality of smart librarians, provide human resources guarantee for the sustainable development of smart libraries, explore a new idea and new way for the training of smart librarians, and has certain reference value and reference function for the training of smart librarians and team building. With the transformation and development of the smart library to the meta-universe library, new requirements will be put forward for the core quality of the smart librarians, and the strategy on the cultivation of the core quality of the smart librarians will also need to be changed. The process of cultivating the core quality of the smart librarians is a process of sustainable development.

[Purpose/Significance] We used to process soybean leaf data by looking at them and process data manually, but this method is very inefficient. In order to improve the classification accuracy and efficiency of soybean leaf images, further for storage and management of these images, we used the deep learning technique to make an in-depth study of text data and image data of soybean leaves for the image recognition and classification. The application of deep learning in agricultural data management mainly focuses on the image recognition and classification of plants and plant phenotypes in large-scale data, detection and classification of agricultural diseases and pests, detection and classification of crops and weeds, and prediction of crop yield. Through case analysis, our sample data demonstrated the application process of deep learning technology. [Method/Process] This paper systematically described the whole process of classification and recognition of agricultural data by using the deep learning technique. Through recognition and disease monitoring of plant leaves, the leaf morphology of soybean plants in the soybean experimental field of Heilongjiang Academy of Agricultural Sciences was taken as an example. We analyzed the image features of soybean leaf morphology, and carried out the classification and recognition research of soybean leaf morphology based on deep learning. Deep learning techniques have replaced shallow classifiers that use manual feature training and can identify soybean leaves with a high degree of accuracy as long as sufficient data are available for training. We adopted DenseNet model, which is suitable for common network model. The advantages of this model are that it has the best performance and the least storage requirements. First,we selected support vector machine (SVM) and random forest (RF) in traditional machine learning methods to identify soybean leaf morphology. Second, AlexNet and ResNet were selected to identify soybean leaf morphology. Finally, the recognition accuracy of different methods were compared with the DenseNet network adopted in this paper. [Results/Conclusions] Through the training of DenseNet model, the recognition accuracy of 94% was achieved, which successfully solved the problems of long time, low efficiency and low recognition accuracy of traditional methods in processing image classification of soybean leaves, and could meet the actual needs of agricultural image data classification. Future research efforts will strive to collect a wide range of large and diverse data sets to facilitate soybean leaf recognition, and focus on developing reliable background removal techniques and incorporating other forms of data to improve the accuracy and reliability of soybean leaf recognition systems.

[Purpose/Significance] Evidence-informed decision-making is a means to bridge the gap between research and policy and evidence synthesis has become an important tool for evidence-based decision-making in many fields. However, evidence synthesis is resource-intensive, especially when it comes to scientific knowledge on complex issues. The efficiency of evidence synthesis currently cannot meet the needs of decision makers. Crowdsourcing is seen as a potential way to improve the productivity of evidence synthesis. At present, the research and practice on the applications of crowdsourcing in evidence synthesis is still in its infancy. This study takes the application of crowdsourcing in the Cochrane Crowd citizen science project as an example to summarize the practical applications of crowdsourcing in evidence synthesis. The comprehensive analysis of the application mechanism of crowdsourcing in Cochrane Crowd project will provide certain reference and inspiration for the use of crowdsourcing in evidence synthesis, so as to improve the production efficiency of evidence synthesis and provide timely and powerful scientific information for evidence-based decision-making. [Method/Process] The application mechanism of crowdsourcing in the Cochrane Crowd citizen science project was analyzed from five dimensions: crowdsourcer, volunteers, crowdsourcing task, Cochrane Crowd platform and effectiveness evaluation, using literature research, network investigation, case analysis and other methods. Cochrane Crowd provides an easy-to-use interface for contributors to engage volunteers to participate and design , in addition to task-focused learning activities, diverse ways of accessing tasks, interactive online training modules and feedback mechanisms to improve the likelihood of volunteers' performing tasks correctly. At the same time, an agreement algorithm is provided at the platform level to aggregate the crowd classification results, which further improves the possibility of correct classification of records. In addition, the platform has used the records identified by the crowd to build a machine-learning model called as RCT classifier which can predict how likely a new citation is to be described an RCT to reduce the manual burden. [Results/Conclusions] Crowdsourcing is an effective method to improve the efficiency of evidence synthesis and shorten the production cycle. With comprehensive participant training and appropriate quality control mechanisms, it is possible to produce high quality crowdsourcing results that meet the "gold standard" of evidence synthesis. In order to motivate volunteers to participate and promote continued engagement, participants are suggested to be provided with clear goals, clear tasks, and timely feedback or rewards. Interest and activity in introducing crowdsourcing into evidence synthesis is growing rapidly, and new tools and platforms to facilitate crowdsourcing also need to be further developed as researchers from different disciplines use crowdsourcing in the evidence synthesis projects. In the future, the application of crowdsourcing in evidence synthesis in different fields and in different stages of evidence synthesis should be further studied.