Abstract
With the popularity of the Internet and the rapid development of e-commerce, online shopping has gradually become an indispensable part of people’s lives. Among them, the rise of cross-border e-commerce has become a focus of attention. The operation traces left by visitors during shopping on the e-commerce platform are stored in the database of the system, and the platform holds such a large amount of valuable data resources. How to unearth valuable content from these resources and apply them becomes very important. This article mainly introduces the research on the visitor information analysis system of the cross-border e-commerce platform based on mobile edge computing. This article first establishes the mobile edge computing framework based on the advantages of the mobile edge computing method and uses it to visit visitors in the visitor information analysis system. In the data filtering, secondly, the requirements of the visitor information analysis system of the cross-border e-commerce platform are analyzed to provide a design basis for the design of the visitor information system. Finally, the visitor information analysis based on the mobile edge algorithm is designed through the demand analysis of the system that has also been tested for visitor information analysis. The test pass rate is as high as 98%, and the accuracy rate of visitor information analysis reaches 80%.
1. Introduction
1.1. Research Background and Significance
As the Internet enters the era of full-scale explosion, the number of e-commerce platforms is also showing a continuous growth trend, and cross-border e-commerce platforms are also gradually increasing. The marketing and promotion of e-commerce platforms and the effects of visitor experience have become an urgent need for most e-commerce managers to understand index [1]. The era when traffic was the only measurement has passed. E-commerce platform data analysis technology has suddenly emerged and has become a brand-new industry. Cross-border e-commerce platform data has received more and more attention and has become a decision support recognized by company managers [2]. Among them, in the processing of data, the advantages of mobile edge computing have brought good news to data processing. Therefore, in the current era of big data, obtaining platform visitor analysis data and conclusions has become an urgent need for Internet practitioners. Based on the basic data of the cross-border e-commerce platform to analyze the behavior of platform visitors and then make product adjustments, the potential value of the e-commerce platform can be continuously tapped, thereby driving the healthy and effective development of the e-commerce platform [3, 4]. However, nowadays, there is a process that focuses on user tasks and always targets users who can generate platform benefits. Obviously, data analysis based on traffic data can no longer meet the needs of e-commerce platforms and cannot be refined to the platform for solutions. For specific operations, it is necessary to design an effective visitor information analysis system and use the basic data of the visit to mine the potential operation rules of visitors and provide more understandable result processing and analysis [5]. Through the analysis of visitor information on the cross-border e-commerce platform based on mobile edge computing, it is possible to understand the process experience of visitors to the product, obtain data related to visitor behavior, and assist designers, product managers, and visual design based on the conclusions and data of the above research. Engineers and others conduct product design and redesign, product innovation, and interface visual guidance. This is of great significance for improving product strategies, guiding cross-border e-commerce interface improvements, and enhancing visitor experience [6]. At the same time, the transaction rules between visitors and cross-border e-commerce platforms are hot and cutting-edge issues in the research of visitor behavior analysis technology. They are mainly reflected in how to distinguish from big data mining, establish a visual and storable information analysis model, and study their relationship with cross-border e-commerce. The difference is between the initial design of the business platform and the expected task flow. Therefore, the research on the visitor information analysis system of the cross-border e-commerce platform based on mobile edge computing will have an important influence on the development trend of the visitor behavior analysis system in the future [7].
1.2. Related Content
With the evolution of latency-sensitive applications, latency constraints have become an obstacle to running complex applications on mobile devices. Partial computing offload is expected to enable these applications to be executed on low-latency mobile user devices. However, most of the existing research focuses on cloud computing or mobile edge computing (MEC) to reduce the task burden. Ning et al. designed an iterative heuristic MEC resource allocation algorithm to dynamically make offloading decisions [8]. They considered the two comprehensively, and it is an early work to study the cooperation between cloud computing and MEC in the Internet of Things. They started with the problem of single-user computing offloading that does not limit MEC resources solved by branch and bound algorithm. Later, considering the resource competition between mobile users, the problem of multiuser computing shunting was expressed as a mixed-integer linear programming problem, which was very difficult. Due to the computational complexity of the proposed problem, simulation results show that their algorithm is superior to the existing scheme in terms of execution delay and offloading efficiency, but the algorithm has errors. E-commerce is one of the most active and important areas in the economy. The latest trend in this market is cross-border trade. It includes selling products to customers in other countries. However, this is accompanied by some problems, such as high cost and long delivery time, language barriers, various legal and tax conditions, etc. Kawa proposed the original concept of integrators in cross-border e-commerce [9]. His research is based on the author’s experience in the field of e-commerce. The analysis of cross-border trade issues is mainly based on the report of the European Commission. To improve the competitiveness of micro- and small electronics companies, especially on the international stage. In addition, the application of his concept can increase the attractiveness of cross-border trade, which will help increase the total amount of e-commerce by increasing the level of sales. But correspondingly, there are high risks. Land managers rely on access to data to make policy and management decisions. However, access to data is usually expensive and difficult to obtain and provides limited information depth [10, 11]. Sessions confirmed a new method for measuring leisure visits [12], providing opportunities for land managers around the world to extend current data collection methods by using easily accessible and cheap crowdsourced online data to track and learn about visits. He evaluated the effectiveness of using online photos from crowd sources to infer information about the habits and preferences of leisure tourists. By comparing empirical data from the National Park Service with photo data in 38 national parks from the online platform Flickr, using multiple regression analysis, he found that the number of photos released by the park each month can reliably indicate the number of visitors to the park in a given month [9, 13]. Through other statistical tests, he also found that the location of the photographer’s residence voluntarily provided in his online profile can accurately show the residence of the park visitors. In addition, his method also enables people to conduct future research on how access rates vary with visits, management or infrastructure, weather events, or ecosystem health and promotes valuation studies such as travel cost studies [14, 15]. However, the scope of application of this method is relatively small.
1.3. Main Content and Innovation
The main content of this article is to study the visitor information analysis system of cross-border e-commerce platform based on mobile edge computing and conduct visitor information analysis system through the advantages of mobile edge computing and the demand analysis of the visitor information analysis system of cross-border e-commerce platform design, analyze information from the number of visitors, order records, etc. And the system has been tested, and data was analyzed. The innovation of this article is to fully design the technology and advantages of mobile edge computing into the visitor information analysis system of the cross-border e-commerce platform to achieve better visitor information filtering and processing.
2. The Advantages of Mobile Edge Computing and Similarity Calculation
2.1. Advantages of Mobile Edge Computing
Mobile edge computing is computing that extends to the edge of the network. It is similar to cloud computing in that it provides storage, computing, and network services for end users [16, 17]. The main characteristics of mobile edge computing are as follows.
As shown in Table 1, mobile edge computing has nearby physical locations, low service node performance, support for multiple communication protocols, low latency, improved link capacity, and improved efficiency. The proximity of the physical location can make the data transmission speed of the data network fast, the performance of the service node is low, it can process a small range of data, and the low latency can allow visitors to access without waiting, improve the link capacity, and reduce the pressure on the backhaul link [18, 19]. Improved performance can also reduce long-distance transmission, which are the advantages of mobile edge transmission. These advantages and effects of mobile edge computing bring information filtering to the information processing in the visitor information analysis system.
2.2. Calculation of Visitor Behavior Similarity
Identify the tag data information and behavior sequence information in the user’s log according to the user’s unique identifier, and then construct the user’s access behavior [20, 21], calculate the user’s tag data similarity and behavior sequence similarity, and combine the two together. Evaluate the similarity of two users [22, 23]. Similarity calculation can categorize the behavior information of visitors, and then perform category processing and analysis. For a set of data labels, the similarity calculation is relatively simple, and the calculation formula iswhere W represents the weight of the label. After calculating the similarity of a label in the label set [23, 24], all the similarities are further integrated to calculate the comprehensive label similarity. The calculation formula is
Among them, SIM is the comprehensive similarity expression of two visitors, c is the centralized sorting of visitor tags, and the other is the weighting factor [25, 26]. The distance formula used to calculate the tags in the two visitor tag sets is shown in the following formula:
After calculating the similarity of the tag features of visitors, it is necessary to calculate the similarity between the visitor information models [27, 28], so when calculating the model similarity, the matrix similarity calculation method is used. The calculation formula is as follows; first, set the conditions aswhere tr is the sum of information elements. From the premise, the matrix norm can be derived as follows:
Therefore, the matrix similarity r can be defined as
The value range of r in formula (8) is (−1, 1). When the angle is zero, r = 1, the similarity between the two matrices is the best [29].
3. Design of Visitor Information System Based on Mobile Edge Computing
3.1. Mobile Edge Computing Architecture
Edge computing refers to the use of an open platform that integrates network, computing, storage, and application core capabilities on the side close to the source of things or data to provide services nearby, and its applications are initiated on the edge side, resulting in faster network service response [30]. The combination of edge computing and cloud computing makes the entire intelligent system not only clear in mind, but also clear in ears and eyes and dexterous in hands and feet. The mobile edge architecture is shown in the figure below.
As shown in Figure 1, the computer and server are first entered into the platform through the network, and then the data is transmitted to the Internet through each network core. This data transmission has a filtering function, which can eliminate impractical data, thereby reducing the system operating pressure. The mobile edge computing framework is able to process user task requests by itself. This not only enables the platform to directly process user requests at the network edge and respond to the results but also actively adjusts itself according to user needs at the network edge, executes active task caching from remote data centers or content distribution networks, and saves some storage space. The services that may be needed by the surrounding users are pulled to the edge of the network, so that when the user requests it, the result of the request is immediately fed back to the user. Through this framework, the burden of the mobile terminal can be reduced to the greatest extent, and at the same time, it is possible to achieve a reasonable and full use of edge resource devices. This kind of framework is very dependent on the status of the network. If the network condition is not good, this model has certain limitations, such as network delays and reduced mobile terminal experience.
3.2. Visitor Relevance Screening
With the help of mobile edge computing, the preference information of all visitors and the positive correlation parameters of the service information city can be selected, and they can be analyzed and researched. The visitor information screening is to analyze the relevance of visitor preferences and e-commerce platforms. We can obtain the user data set Y from the data collection unit and give the correlation calculation formula of the relationship number to calculate the correlation between the e-commerce platform and user preferences:where s is a feature in the data set.
I is used to refer to visitor preferences and the status of services. Next, the research problem in the given parameters can be expressed by the following formula:
Among them,
Formula (3) is responsible for calculating the probability of whether to request the application or service according to the given data set s, and the state with the largest value obtained by multiplying the prior probability by the conditional probability is determined as the classification state of the data. Formula (4) is responsible for calculating the prior probability, that is, whether the application or service has been requested in the respective probabilities. Formula (5) is responsible for calculating the conditional probability, that is, when the application or service is requested or not. The i-th feature data represents its probability in the j-th feature data. The preclassification results of the data set S: calculate the correlation between the platform in R and each application or service in I according to formula (1); sort the data showing positive correlation results according to the result of formula (1); according to formula (1), formulas (5) and (4) train a preallocation model on the sensor data set showing positive correlation results; and then use the allocation model to preclassify s according to formula (2).
3.3. Design of Visitor Information Data Collection
Visitor information data collection is a process of collecting relevant data purposefully according to needs. It is a prerequisite basis for information data analysis, and it is indispensable in all information data analysis systems. With the increase in the number of visitors, the challenge of collecting visitor information and data has become particularly prominent, including the variety of information data sources, the large amount of data, and how to ensure the quality of the data. The goal of data collection design: collect as much data as possible to reduce the rate of data loss; reduce the intrusion to the business system as much as possible, and reduce the impact on the business system. The data collection method is mainly that the PC side sends data to the server through the JS SDK and the mobile side through the Java SDK of the background system, monitored by the Flume component and then uploads the information data to the file system to complete the information data collection task. The following figure shows the design of visitor information data collection.
As shown in Figure 2, the visitor information data collection design is as above. It enters the entry server through different SDKs through the PC and the program backend and then collects information through service 1 and server N. At the same time, Flume1 and Flume2 monitor the server at the same time. Prevent the loss of data, and then transfer the complete data to the file system. Two servers and two monitoring components are set up. The purpose of the design is to prevent data loss caused by excessive visitor data. Among them, the Flume component is designed to reduce the loss of visitor information and data and solve a major problem in information data collection.
3.4. Design of Visitor Information Analysis System
Combined with the current development of cross-border e-commerce platforms and the application of data warehouse, data mining, and data analysis in banks, the design method adopted in the software part of this design is a structured development method. In the development process of the management information system, it is necessary to select the appropriate development method according to the functional requirements of the system, implementation technology, and application occasions. Figure 3 shows the software development method architecture diagram.
The structured development method in the system has the characteristics of data determinism and data consistency, so this method has been widely concerned and used by people. This system development method system adheres to the principle of visitor and uses system ideas and organizational modules to the following method to analyze and design the system as a whole. This system brings the system to ensure the accuracy of visitor information.
3.5. Design of Visitor Information Analysis System for Cross-Border E-Commerce Platform Based on Mobile Edge Computing
Cross-border e-commerce platforms have different user needs from local e-commerce platforms. It is necessary to select a suitable information analysis and information processing system for cross-border e-commerce platforms and to mine and process visitor information obtained from cross-border e-commerce platforms. Mining and analysis to provide visitor data for cross-border e-commerce operators, and then analyze better operational decisions, and realize the profitability of the platform.
3.5.1. System Frame Design
The system consists of four parts: data application layer, data analysis layer, data layer, and infrastructure layer. It uses cross-border e-commerce platform clusters as the infrastructure to process and analyze the purchase records and business data of the data layer. Finally, the final analysis result is displayed on the main interface. The following figure shows the overall architecture of the system.
As shown in Figure 4, the data application layer includes preferential policies and user recommendations, the data analysis layer includes visitor behavior data analysis and product evaluation analysis, the data layer includes purchase records and business data, and the infrastructure layer includes servers, cross-border e-commerce clusters, and cluster operation management. The system framework has a comprehensive analysis of visitor information and a design compatible with cross-border e-commerce platforms. Among them, the information data analysis function can be subdivided into the following aspects.
As shown in Figure 5, the specific functions of visitor information data analysis are embodied in visitor information data analysis, visitor product evaluation analysis, conversation analysis, and geographic information analysis. The analysis of this information is applicable to cross-border e-commerce platforms. From these four specific aspects, the information analysis of the cross-border e-commerce platform can analyze visitors from different aspects, and then achieve a comprehensive analysis of visitors’ tags.
4. Visitor Information System Requirements and System Testing
4.1. System Requirements
Requirement analysis is an important activity in the analysis system planning stage and an important link in the system life cycle. Requirement analysis is the key to system design and the basis for system design. The content of requirements analysis is to provide complete, clear, and specific requirements for the software to be developed and to determine which tasks the software must achieve. In the development process, combined with the needs of users, it is mainly divided into functional requirements and nonfunctional requirements.
4.1.1. Functional Requirements
From the perspective of cross-border e-commerce platforms, this article analyzes the specific tasks required from visitor information. The main functional requirements of the system are as follows.
As shown in Table 2, the functional requirements of the system have the abovementioned aspects. The most critical ones are mobile edge computing, visitor behavior data analysis, and product evaluation analysis. Mobile edge computing enables the system to perform non-real-time analysis and processing. No time and data are wasted; product evaluation analysis is to analyze the review information of products on the cross-border e-commerce platform, and analyze the sentiment of the visitor to understand the voice of the visitor. The analysis of visitor behavior data is carried out by obtaining the data of visitors’ visits to cross-border e-commerce platforms, from which the behavior habits of visitors can be discovered, and by combining them with marketing strategies, problems in marketing activities can also be found from them, and then provide a basis for improving marketing strategies. The analysis items involved in the analysis of visitor behavior data are shown in Table 3.
As shown in Table 4, in the analysis of visitor data, the above indicators are all analysis objects, and these indicators are all based on the cross-border e-commerce platform. Analyzing visitor information from different aspects can better understand visitors so that cross-border e-commerce platforms can better improve marketing strategies to meet the needs of visitors.
4.1.2. Nonfunctional Requirements
The first consideration for the nonfunctional requirements of the system is the digital information era, where visitors can quickly obtain digital information and personal information privacy requirements. In this regard, it is necessary to consider the convenience and reliability of mobile Internet access to digital resources, and it is necessary to use tunnel links to achieve data isolation to protect users’ private information. The second thing to consider is the reliability of the network. As the files of digital resources are getting larger and larger, the files of information such as book information and video are constantly increasing, which puts forward higher requirements for the reliability and stability of the network. The nonfunctional requirements have the following points. The first is the reliability of the system. The improvement of system reliability is mainly considered from the following two aspects: one is to use relatively mature and well-known hardware products; the other is to focus on software design. Consider what methods can be used to make the collected data more accurate and the business process clearer. When processing data, consider using more reasonable algorithms to make the most accurate and scientific results. The second is the security of the system. The core of the security of the system is data security. In order to avoid data loss and damage, the first thing to consider is the stable and reliable backup and restore method of the system. Through the administrator user authority management, the important data is incrementally and fully backed up to avoid the abnormal loss of system data. The third is the scalability of the system. In order to improve the accuracy of personalized recommendations and the diversity of data sources, more systems will be dynamically added, such as user credit reporting systems, which are mainly reflected in the increase in business flexibility and the redundancy of databases (data information backup, network security management, standardized interface design and other platforms docking, etc.).
4.2. System Test
System testing is a key step in system research. For testing in this visitor information analysis system, this article analyzes the data presented by some visitor information analysis, while hardware testing is not involved.
4.2.1. Test of Visitor Information Collection Function
The visitor information collection function is mainly through the Flume component for information collection. Flume transmits information data from the source to the final destination in the form of events. Take Flume to monitor a certain port and output the data written in the port as a logger as an example to verify the correctness of its configuration and the reliability of real-time data transmission, as shown in the table below.
As shown in Table 3, after testing the configuration components, port information, and data transmission, it can be seen that all have passed the test, so the visitor information collection function can operate normally.
4.2.2. Visitor Product Recommendation Test
The visitor product recommendation test part is mainly to implement the mobile edge computing method on the cross-border e-commerce platform. The specific test steps are shown in the following table.
As shown in Table 5, the functions of visitor product recommendation are all working normally, but the correctness of the recommended results is still unknown. Therefore, the visitor data of the recommended products will be analyzed again to see if the recommended products are liked by visitors. Select the data recommended by 100 visitors, and divide the visitors between men and women to see whether the visitors have purchased the recommended products. The analysis results are as follows.
As shown in Figure 6, it can be seen that among the 100 visitors, 61 visitors completed the recommended purchase of the product, with a purchase rate of 61%, and 25 purchased but did not pay and hesitated to recommend the product. There are more hesitant female visitors than men, indicating that women like to compare products before purchasing. There are 14 people who do not buy, and those who do not buy are a minority. This shows that the system’s product recommendation analysis is accurate for visitors. Through the recommendation of visitors’ products, visitors’ purchasing power can be improved, which will bring more profits to cross-border e-commerce platforms.
4.2.3. Order Analysis Module Test
Through the analysis of the order that the visitor successfully paid for the purchase, the analysis is carried out from the dimension of the order price, and the data analysis is shown in the figure below.
As shown in Figure 7, as the price rises, the number of purchases of men has decreased, while the number of women has increased. Therefore, it can be seen that women are not sensitive to high prices, while men are more sensitive. The price of male and female products on the e-commerce platform can be adjusted through systematic data analysis to adjust the price of high and low prices and to increase the recommendation of low-priced products for men, so as to achieve better sales.
5. Discussion
This paper designs and studies the visitor information analysis system of the cross-border e-commerce platform based on mobile edge computing and improves the technical design of the visitor information analysis system through the key technologies of mobile edge computing, which reduces the burden on the system and allows visitors to shop online more flexibly and conveniently. Information analysis system achieves better visitor information filtering and processing. Although the research of this article still has insufficient in-depth understanding of mobile edge computing technology and insufficient system design, the visitor information analysis system of this cross-border e-commerce platform provides certain help to future visitor data analysis and will provide some help for future visitor data analysis. Information analysis provides a certain basis.
Data Availability
No data were used to support this study.
Conflicts of Interest
The authors declare no conflicts of interest.
Authors’ Contributions
All authors have seen the manuscript and approved to submit it.
Acknowledgments
This work was supported by the Jiangxi Provincial Department of Education Science and Technology Research Project “Jiangxi Province Cross-Border E-Commerce Industry Chain Cluster Research” (GJJ202505) and Nanchang Institute of Science and Technology introduced talents research start project “Cross-Border E-Commerce Industry Chain Cluster Research” (NGRCZX-20-11).