Cross-border e-commerce supply chain data analysis method and system based on artificial intelligence

Abstract

The invention relates to the technical field of e-commerce management, in particular to a cross-border e-commerce supply chain data analysis method and system based on artificial intelligence, and the method comprises the steps: based on commodity transaction data of a cross-border e-commerce transaction platform in a specific sales cycle, responding to a condition that a first transaction growth rate of a commodity in the commodity transaction data exceeds a preset growth amplitude threshold value; determining first potential purchase popularity distribution of the commodity in the target market region based on the first market region distribution data; based on the first potential purchase popularity distribution, obtaining historical supply chain operation data of the commodities in the corresponding target market region, and performing association rule analysis on the historical supply chain operation data to obtain an association operation feature mode of the commodities in each link of the supply chain. According to the invention, by dynamically adjusting the inventory management and control strategy, the inventory distribution can be optimized in real time according to the market demand change and the commodity transaction data, so that the phenomena of inventory overstock and stockout are reduced, and the overall inventory turnover rate is improved.

Description

Technical Field

[0001] This invention relates to the field of e-commerce management technology, specifically to a method and system for cross-border e-commerce supply chain data analysis based on artificial intelligence. Background Technology

[0002] In recent years, consumers' shopping habits and preferences have changed significantly. More and more people are choosing to shop online and expect fast and personalized services. Businesses need to understand consumer needs through accurate data analysis and adjust product supply and marketing strategies in a timely manner to improve customer satisfaction and loyalty.

[0003] Currently, traditional methods often rely on fixed inventory levels and periodic checks, which cannot reflect changes in market demand in real time. This static management approach may lead to excess or shortage of inventory, increasing capital occupation and storage costs. Moreover, the decision-making process is based on historical data and seasonal forecasts, which usually suffers from information lag and cannot capture market changes in a timely manner. This makes companies slow to respond to emergencies or fluctuations in market demand, resulting in lost sales opportunities.

[0004] Furthermore, traditional methods often lack integrated analysis of data from different sources, resulting in an inability to fully understand consumer behavior and market trends. Data silos at various stages can hinder overall business optimization. Moreover, due to the failure to accurately identify market demand, traditional methods often lead to unreasonable resource allocation, wasting human and material resources. For example, they may over-order certain products while neglecting other more popular products. Summary of the Invention

[0005] To achieve the above objectives, the present invention provides the following technical solution: a cross-border e-commerce supply chain data analysis method based on artificial intelligence, comprising: Based on the commodity transaction data of the cross-border e-commerce platform within a specific sales cycle, in response to the fact that the first transaction growth rate of a commodity in the commodity transaction data exceeds a preset growth rate threshold, the first potential purchase popularity distribution of the commodity in the target market region is determined based on the first market area distribution data. Based on the distribution of the first potential purchase popularity, historical supply chain operation data of the product in the corresponding target market region is obtained, and association rule analysis is performed on the historical supply chain operation data to obtain the association operation feature pattern of the product in each link of the supply chain. Based on the distribution of the first potential purchase popularity and the associated operational characteristics, the first inventory control strategy for the product within a specific sales cycle is adjusted to generate a second inventory control strategy; based on the second inventory control strategy, the cross-border e-commerce transaction platform is controlled to dynamically allocate the inventory of the product. Based on the commodity transaction data and inventory allocation data of the cross-border e-commerce platform, the distribution of the second potential purchase popularity of the commodity in the next sales cycle is predicted. Based on the second potential purchase popularity distribution, the operational strategies for the products in each link of the cross-border e-commerce supply chain in the next sales cycle are re-planned to generate a third operational strategy; based on the third operational strategy, the cross-border e-commerce trading platform is controlled to optimize and adjust the supply chain operation of the products in the next sales cycle.

[0006] Preferably, the first market area distribution data includes the sales proportion of the product in different market segments, customer evaluation feedback information, and logistics and transportation timeliness data, and the first potential purchase popularity distribution is used to characterize the intensity of potential purchase tendency of the product in different sub-regions of the target market area. Based on the first market area distribution data, determine the first potential purchase popularity distribution of the product in the target market region, including: Based on the regional distribution data of the primary market, the trend of purchasing popularity changes in each sub-market is determined along the market segmentation dimension; Based on the trends in purchasing activity across different market segments, peak purchasing activity ranges are determined. Based on the regional characteristics of the segmented market corresponding to each peak purchase popularity interval, the target market region is segmented to obtain each sub-market region within the target market region; Based on the sales share of the product in each sub-market region, customer feedback information, and logistics and transportation timeliness data, the distribution of the first potential purchase popularity is determined in each sub-market region.

[0007] Preferably, association rule analysis is performed on historical supply chain operation data to obtain the correlation operation characteristic patterns of the product in each link of the supply chain, including: Determine sample data points for the product at each stage of the supply chain during the sales cycle; Based on the historical operational information of sample data points of the product in each link of the supply chain during the sales cycle, and the market dynamic information during a specific sales cycle, the operational efficiency change ratio between two adjacent sample data points is determined. Based on the operational efficiency change ratio between two adjacent sample data points and the overall operating cost data of the product during the sales cycle, the correlation operation characteristic pattern of the product in each link of the supply chain during the sales cycle is determined.

[0008] Preferably, based on the first potential purchase popularity distribution and related operational characteristic patterns, the first inventory control strategy for the product within a specific sales cycle is adjusted to generate a second inventory control strategy, including: Prioritize the supply chain links corresponding to sub-regions with high potential purchase interest in the associated operation characteristic pattern, and use the priority ranking results to determine the inventory allocation ratio in the first inventory control strategy. Based on the current total inventory of the goods and the updated inventory allocation ratio, the inventory layout is replanned to form a second inventory control strategy.

[0009] Preferably, based on commodity transaction data and inventory allocation data from cross-border e-commerce platforms, the distribution of the second potential purchase popularity of the commodity in the next sales cycle is predicted, including: Based on transaction volume changes, customer purchase frequency, and inventory turnover speed information from cross-border e-commerce transaction data, the trends of transaction volume, customer purchase frequency, and inventory turnover speed of the product in the next sales cycle are predicted. The distribution of the second potential purchase interest is determined based on the predicted trends in transaction volume, customer purchase frequency, and inventory turnover.

[0010] Preferably, after optimizing and adjusting the supply chain operation of the goods on the cross-border e-commerce platform in the next sales cycle based on the third operational strategy, the method further includes: Obtain the first logistics path data of the goods; extract multiple continuous logistics nodes based on the first logistics path data, and establish multiple dynamic analysis periods for the multiple continuous logistics nodes according to preset time intervals, specifically including: Based on the first logistics path data, obtain multiple logistics stage paths with multiple consecutive logistics nodes as key nodes; Obtain the logistics area characteristics and initial inventory levels corresponding to multiple logistics stage paths; The characteristics of the logistics area and the initial inventory level are input into the logistics trend prediction model to simulate the logistics trend and generate logistics fluctuation cycles corresponding to multiple logistics stage paths. The logistics trend prediction model is trained based on historical logistics data of the same commodity type and the same logistics area. The logistics fluctuation cycle is adjusted according to a preset elastic constraint to generate a preset time interval corresponding to the logistics stage path; Configure the preset logistics fluctuation monitoring duration; The preset logistics fluctuation monitoring duration is used as the duration of multiple dynamic analysis periods, and the preset time interval is used as the time distance from multiple dynamic analysis periods to multiple continuous logistics nodes. Multiple dynamic analysis periods are configured in multiple logistics stage paths. Extract dynamic analysis periods from multiple dynamic analysis periods and obtain the corresponding logistics nodes; According to the first logistics path data, the goods are controlled to flow to the logistics nodes. When the goods are detected to enter the dynamic analysis period, logistics trend analysis is performed to generate logistics trend analysis results. Based on the logistics trend analysis results, the inventory management parameters are optimized for the first logistics stage between the dynamic analysis period and the logistics node, generating inventory optimization results.

[0011] Preferably, goods are controlled to flow to logistics nodes according to the first logistics path data. When goods are detected to enter the dynamic analysis period, logistics trend analysis is performed to generate logistics trend analysis results, including: Establish logistics trend analysis sample constraints, wherein the logistics trend analysis sample constraints include a first logistics sample constraint and a second logistics sample constraint. The first logistics sample constraint is that the logistics growth rate meets a first preset logistics growth rate, and the second logistics sample constraint is that the logistics growth rate meets a second preset logistics growth rate, wherein the first preset logistics growth rate is greater than the second preset logistics growth rate. When a product is detected to have entered a dynamic analysis period, the first real-time volume of the product is obtained. Based on the first real-time logistics volume, and combined with the constraints of the first logistics sample and the second logistics sample, data on the continuous logistics growth stage is generated. Logistics trend analysis data is generated from continuous logistics growth phase data, which includes two logistics growth control nodes. Perform logistics fluctuation deviation analysis on logistics trend analysis data to generate logistics trend analysis results.

[0012] Preferably, logistics trend analysis data is subjected to logistics fluctuation deviation analysis to generate logistics trend analysis results, including: Collect market consumption preference characteristics and logistics area characteristics of commodity logistics areas to model logistics trends and generate logistics trend simulation models; Input logistics trend analysis data into the logistics trend simulation model to simulate logistics trends and generate a standard time series of logistics volume changes. Perform logistics trend analysis on logistics trend analysis data and record the time sequence of actual logistics volume changes; After time and space alignment, the standard material flow rate change time series and the actual material flow rate change time series are compared to generate material flow rate deviations corresponding to multiple time and space nodes. Logistics trend analysis results are generated based on the deviations in material flow corresponding to multiple spatiotemporal nodes.

[0013] Preferably, based on the logistics trend analysis results, inventory management parameters are optimized for the first logistics stage between the dynamic analysis period and the logistics node, generating inventory optimization results, including: Determine the initial inventory level and replenishment time points in the first logistics stage; Based on the logistics trend analysis results, determine whether the deviation in the volume of goods meets the preset deviation threshold; If not, after correcting the initial inventory level based on the deviation in material flow, simulate inventory consumption, optimize replenishment timing based on the forecast results, and generate inventory optimization results.

[0014] An AI-based cross-border e-commerce supply chain data analysis system, applicable to the aforementioned AI-based cross-border e-commerce supply chain data analysis methods, including: The first analysis unit is used to determine the first potential purchase popularity distribution of the product in the target market region based on the product transaction data of the cross-border e-commerce transaction platform within a specific sales cycle, in response to the first transaction growth rate of a product in the product transaction data exceeding a preset growth rate threshold, and based on the first market area distribution data. The second analysis unit is used to obtain historical supply chain operation data of the product in the corresponding target market region based on the first potential purchase popularity distribution, and to perform association rule analysis on the historical supply chain operation data to obtain the association operation feature pattern of the product in each link of the supply chain. The third analysis unit is used to adjust the first inventory control strategy for the product within a specific sales cycle based on the first potential purchase popularity distribution and related operational characteristics, and generate a second inventory control strategy; based on the second inventory control strategy, it controls the cross-border e-commerce transaction platform to dynamically allocate the inventory of the product. The fourth analysis unit is used to predict the second potential purchase popularity distribution of the product in the next sales cycle based on the product transaction data and inventory allocation data of the cross-border e-commerce transaction platform. The fifth analysis unit is used to re-plan the operational strategies of the product in each link of the cross-border e-commerce supply chain in the next sales cycle based on the second potential purchase popularity distribution, and generate a third operational strategy; based on the third operational strategy, the cross-border e-commerce transaction platform is controlled to optimize and adjust the supply chain operation of the product in the next sales cycle.

[0015] Compared with the prior art, the beneficial effects of the present invention are: (1) By dynamically adjusting the inventory management strategy, the present invention can optimize the inventory allocation in real time according to changes in market demand and commodity transaction data, thereby reducing inventory backlog and stockouts and improving the overall inventory turnover rate; Moreover, by analyzing the correlation between purchase popularity distribution and historical supply chain operation data, it can more effectively identify potential demand and consumption trends in the target market, provide data support for enterprise decision-making, and enhance market sensitivity. (2) This invention generates targeted operational strategies by conducting in-depth analysis of commodity transaction data in different sales cycles, which helps enterprises respond quickly to market changes, optimize resource allocation in all links of the supply chain, and improve operational efficiency. Moreover, with the help of logistics trend prediction models and dynamic analysis time periods, enterprises can better monitor the logistics process, identify potential fluctuation risks, and carry out timely logistics scheduling and inventory management, thereby reducing logistics costs and improving customer satisfaction. Attached Figure Description

[0016] Figure 1 This is a schematic flowchart of the overall method in one embodiment of the present invention; Figure 2 This is a schematic diagram of the overall system architecture in one embodiment of the present invention.

[0017] In the diagram: 1. First analysis unit; 2. Second analysis unit; 3. Third analysis unit; 4. Fourth analysis unit; 5. Fifth analysis unit. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1, please refer to Figure 1 This invention provides a technical solution: a cross-border e-commerce supply chain data analysis method based on artificial intelligence, comprising: S1. Based on the commodity transaction data of the cross-border e-commerce transaction platform within a specific sales cycle, in response to the fact that the first transaction growth rate of a commodity in the commodity transaction data exceeds the preset growth rate threshold, based on the first market area distribution data, determine the first potential purchase popularity distribution of the commodity in the target market region. S2. Based on the distribution of the first potential purchase popularity, obtain the historical supply chain operation data of the goods in the corresponding target market region, analyze the association rules of the historical supply chain operation data, and obtain the association operation characteristic pattern of the goods in each link of the supply chain. S3. Based on the distribution of the first potential purchase popularity and the associated operational characteristics, adjust the first inventory control strategy for the product within a specific sales cycle to generate a second inventory control strategy; based on the second inventory control strategy, control the cross-border e-commerce transaction platform to dynamically allocate the product's inventory. S4. Based on the commodity transaction data and inventory allocation data of the cross-border e-commerce transaction platform, predict the distribution of the second potential purchase popularity of the commodity in the next sales cycle. S5. Based on the distribution of the second potential purchase popularity, re-plan the operation strategy of the product in each link of the cross-border e-commerce supply chain in the next sales cycle and generate the third operation strategy; based on the third operation strategy, control the cross-border e-commerce transaction platform to optimize and adjust the supply chain operation of the product in the next sales cycle.

[0020] It's important to note that on cross-border e-commerce platforms, merchants analyze product transaction data based on specific sales cycles (such as holiday promotional seasons). If a product's transaction growth rate exceeds a preset threshold (e.g., 30%) during this cycle, the system will combine this data with the product's distribution data across different market regions to identify the potential purchase popularity distribution of the product in the target market. For example, suppose a new smartphone's transaction growth rate reached 50% during Singles' Day. By analyzing the sales data, the system found that the phone was extremely popular in Southeast Asian markets, especially in Indonesia and Malaysia. After identifying potential purchase interest, the system will obtain historical supply chain operation data for the product in the target market and perform association rule analysis. This step aims to identify the operational characteristics and patterns of the product in each link of the supply chain (such as procurement, warehousing, and transportation) to understand which factors drive product sales. For example, continuing with smartphones, by analyzing historical data, it was found that when the inventory level of smartphones is below 500 units, sales volume drops significantly. In addition, the sales of headphones that come with smartphones also increase significantly at the same time as smartphone sales, indicating a strong correlation between the two. Based on the distribution of the first potential purchase heat and the associated operational characteristics, the system will adjust the inventory management strategy to ensure that market needs can be met during periods of high demand. After generating a new inventory strategy, the system will control the dynamic allocation of inventory to optimize supply chain efficiency. For example, in the case of the aforementioned smartphones, the platform decided to increase the inventory in the Southeast Asian market to 1,000 units and set up a priority delivery mechanism in the Indonesian warehouse to ensure a rapid response to high demand. Leveraging cross-border e-commerce transaction data and inventory allocation data, the system will predict the distribution of the second potential purchase activity in the next sales cycle. This prediction will be based on previous sales data and current inventory status. For example, after implementing a new inventory strategy, the system predicts that smartphone purchase activity will remain high in the next quarter (such as the Christmas season), with sales expected to increase by 20% in Indonesia. Based on the distribution of the second potential purchase popularity, the system will re-plan the supply chain operation strategy of the goods and generate new operation strategies to optimize resource allocation and improve efficiency. For example, in order to cope with the upcoming purchase peak, the platform decides to increase more marketing activities in the next sales cycle and cooperate with local logistics companies to optimize delivery time. At the same time, based on forecasts, it will increase the procurement of related accessories (such as phone cases, screen protectors, etc.) to provide consumers with a one-stop shopping experience when purchasing mobile phones.

[0021] In an optional embodiment, the first market area distribution data includes the sales proportion of the product in different market segments, customer evaluation feedback information, and logistics and transportation timeliness data, and the first potential purchase popularity distribution is used to characterize the intensity of potential purchase tendency of the product in different sub-regions of the target market area. Based on the distribution data of the first market region, determine the distribution of the first potential purchase popularity of the product in the target market region, including: Based on the regional distribution data of the primary market, the trend of purchasing popularity changes in each sub-market is determined along the market segmentation dimension; Based on the trends in purchasing activity across different market segments, peak purchasing activity ranges are determined. Based on the regional characteristics of the segmented market corresponding to each peak purchase popularity interval, the target market region is segmented to obtain each sub-market region within the target market region; Based on the sales share of products in each sub-market region, customer feedback information, and logistics and transportation timeliness data, the distribution of the first potential purchase popularity is determined in each sub-market region.

[0022] It should be noted that the first market regional distribution data refers to a comprehensive analysis of the sales performance, customer feedback, and logistics efficiency of products in different market segments; this data helps to identify consumer purchasing tendencies and preferences in different regions. Based on market segmentation, analyze the changing trends in purchasing popularity of products in different sub-market regions; this can be achieved by examining historical sales data, seasonal variations, and the impact of promotional activities; for example, suppose the sales data of a certain athletic shoe in different cities are as follows: Beijing: growth rate 10%, Shanghai: growth rate 20%, Guangzhou: growth rate 5%; through analysis, it can be seen that the change in purchasing popularity in Shanghai is the most significant. After understanding the changes in purchasing activity in various market regions, it is necessary to determine the peak purchasing activity in these regions. These peaks represent the high points of consumer willingness to purchase goods. For example, continuing with the example of athletic shoes, analysis shows that the peak purchasing activity in Shanghai is in the summer (June to August) because there are many summer promotions at this time; the peak purchasing activity in Beijing is in the spring (April to May) because consumers start to engage in outdoor activities. Based on the identified peak purchase activity, further segment the target market by region to gain a more detailed understanding of the characteristics of each sub-market. For example, for the Chinese market, the target market can be segmented into: first-tier cities (such as Beijing and Shanghai); second-tier cities (such as Hangzhou and Nanjing); and third-tier cities (such as Wenzhou and Nanchang). Through segmentation, different marketing strategies can be developed for the characteristics of each city. By combining the sales share of products in each sub-market region, customer feedback information, and logistics delivery time data, the distribution of the primary potential purchase interest in each sub-market region can be determined. For example, suppose the analysis yields the following conclusions: Beijing accounts for 30% of sales, with positive customer feedback and fast logistics delivery (1-2 days); Shanghai accounts for 40% of sales, with relatively good customer feedback but slightly slower logistics delivery (3 days); Guangzhou accounts for 20% of sales, with average customer feedback and average logistics delivery (2 days). Based on this information, we can conclude that: Beijing has a high primary potential purchase interest, making it suitable for large-scale inventory deployment; Shanghai, although having a high sales share, needs to improve logistics and enhance customer satisfaction; and Guangzhou needs to strengthen marketing and improve customer feedback.

[0023] In one optional embodiment, association rule analysis is performed on historical supply chain operation data to obtain the correlation operation characteristic patterns of goods at each stage of the supply chain, including: Determine sample data points for the product at each stage of the supply chain during the sales cycle; Based on historical operational information of sample data points of goods in each link of the supply chain during the sales cycle, as well as market dynamic information during a specific sales cycle, the operational efficiency change ratio between two adjacent sample data points is determined. Based on the operational efficiency change ratio between two adjacent sample data points and the overall operating cost data of the goods during the sales cycle, the correlation operation characteristics of the goods in each link of the supply chain during the sales cycle are determined.

[0024] It's important to note that when analyzing the supply chain, you first need to select a specific sales cycle (such as a quarter, a month, etc.) and collect relevant data for each stage within that cycle (such as procurement, production, warehousing, transportation, sales, etc.). These data points will serve as the basis for subsequent analysis. For example, suppose we are analyzing the supply chain operation data of a certain brand of athletic shoes in the first quarter of 2023. Within this cycle, the following sample data points can be collected: Procurement: 1000 pairs of shoes were purchased on January 1st, and 1200 pairs of shoes were purchased on January 15th; Production: 500 pairs of shoes were produced on January 5th, and 800 pairs of shoes were produced on January 20th; Sales: 300 pairs of shoes were sold on January 10th, and 600 pairs of shoes were sold on February 5th. After collecting sample data points, the next step is to analyze the operational efficiency change ratio between adjacent data points. This can be achieved by comparing the performance indicators of each link at different time points, such as sales volume, inventory turnover rate, and production efficiency. For example, suppose we analyze the operational efficiency change in the procurement link: January 1st: Purchase 1000 pairs of shoes at a cost of 100,000 yuan; January 15th: Purchase 1200 pairs of shoes at a cost of 120,000 yuan. By calculating the change ratio: First data point (January 1st): Efficiency = 1000 pairs / 100,000 yuan = 100 pairs / 10,000 yuan; Second data point (January 15th): Efficiency = 1200 pairs / 120,000 yuan = 100 pairs / 10,000 yuan. For example, let's analyze the production process: On January 5th, 500 pairs of shoes were produced at a production cost of 50,000 yuan, and the efficiency was 500 pairs / 50,000 yuan = 100 pairs / 10,000 yuan; On January 20th, 800 pairs of shoes were produced at a production cost of 60,000 yuan, and the efficiency was 800 pairs / 60,000 yuan ≈ 133.33 pairs / 10,000 yuan. By combining the operational efficiency change ratios between adjacent sample data points with overall operating cost data within the sales cycle, we can identify the correlation patterns of product operations across the supply chain. This helps companies understand which links are performing well and which need improvement, thereby optimizing the efficiency of the entire supply chain. For example, in this case, assuming the overall operating cost of athletic shoes is 500,000 yuan, sales revenue is 800,000 yuan, and profit is 300,000 yuan during the first quarter sales cycle, we know from the previous analysis that: the efficiency change ratio in the procurement link is 1 (no change); the efficiency change ratio in the production link is approximately 1.333 (an increase of 33.3%). Combining this information, we can derive the following correlation patterns: Procurement link: Although there was no improvement, the increased procurement volume kept the supply stable, therefore, it is necessary to maintain good relationships with suppliers to ensure sufficient supply; Production link: Production efficiency has improved significantly, possibly due to the commissioning of new equipment or optimization of the production process. It is recommended to continue investing in this area to further increase capacity; Sales link: If sales revenue does not match production capacity, it is necessary to analyze changes in market demand and adjust production plans to avoid oversupply or shortage.

[0025] In an optional embodiment, based on the first potential purchase popularity distribution and associated operational characteristic patterns, the first inventory control strategy for goods within a specific sales cycle is adjusted to generate a second inventory control strategy, including: Prioritize the supply chain links corresponding to sub-regions with high potential purchase interest in the associated operation characteristic pattern, and use the priority ranking results to determine the inventory allocation ratio in the first inventory control strategy. Based on the current total inventory of goods and the updated inventory allocation ratio, the inventory layout is re-planned to form a second inventory control strategy.

[0026] It's important to note that identifying high-potential purchase areas, such as a specific city or sales channel (online or offline), involves analyzing historical data to pinpoint the corresponding supply chain segments, such as production, warehousing, and distribution. These segments are then prioritized, typically using metrics like past sales figures and inventory turnover. For example, suppose you're selling athletic shoes, and analysis shows high online sales in city A, accounting for 40% of total sales. Further analysis reveals high efficiency in production, enabling rapid order response, while distribution is also efficient in city A. Therefore, the priority might be: production; distribution; warehousing. Based on this priority, adjust the inventory allocation in the first inventory management strategy. For instance, assuming a total inventory of 10,000 pairs of shoes, you might allocate 70% (7,000 pairs) to production, 20% (2,000 pairs) to distribution, and 10% (1,000 pairs) to warehousing. Calculate the specific inventory quantity for each stage to ensure the overall inventory remains constant; optimize inventory storage locations to make inventory in high-potential-purchase-demand areas more easily accessible, thereby reducing delivery time and costs. Example: Continuing with the athletic shoe example, assume two main warehouses: Warehouse X (near city A, primarily serving online sales); Warehouse Y (near city B, primarily serving offline retail). Based on the new inventory allocation ratio: Production: 7,000 pairs; Distribution: 2,000 pairs; Warehousing: 1,000 pairs. Assume that all production inventory is placed in Warehouse X because it supports high-demand online sales. For the distribution stage, 2,000 pairs of shoes can be allocated to Warehouse X for rapid delivery to consumers in city A; for the warehousing stage, 1,000 pairs of shoes can be allocated to Warehouse Y to support offline sales demand in city B. Ultimately, the new inventory layout might be as follows: Warehouse X: 7,000 pairs (production inventory) + 2,000 pairs (distribution inventory) = 9,000 pairs; Warehouse Y: 1,000 pairs (warehousing inventory).

[0027] In an optional embodiment, based on commodity transaction data and inventory allocation data from a cross-border e-commerce platform, predicting the distribution of the second potential purchase popularity of commodities in the next sales cycle includes: Based on transaction volume changes, customer purchase frequency, and inventory turnover speed information from cross-border e-commerce transaction data, the trends of transaction volume, customer purchase frequency, and inventory turnover speed in the next sales cycle can be predicted. The distribution of the second potential purchase interest is determined based on the predicted trends in transaction volume, customer purchase frequency, and inventory turnover.

[0028] It's important to note that the following information is relevant: Transaction volume change information: This analyzes the transaction volume of goods over a past period to identify trends. For example, an increase or decrease in sales of a specific product within a specific timeframe can reflect changes in market demand. Customer purchase frequency information: This analyzes customer purchasing behavior, such as the average number of times each customer purchases the product within a given timeframe. This can reveal customer preferences and loyalty. Inventory turnover speed information: This analyzes the efficiency of inventory turnover, i.e., the time from when goods enter the warehouse to when they are sold. This helps merchants understand product liquidity and determine replenishment and promotional strategies. Through the analysis of the above data, the following predictive results can be obtained: Predicting transaction volume trends: Based on historical transaction volume data, predicting whether the transaction volume of a product will increase, decrease, or remain flat in the next sales cycle; Predicting customer purchase frequency trends: Based on customer purchasing behavior data, determining whether customer attention to the product is increasing, thus affecting future purchase frequency; Predicting inventory turnover speed trends: By assessing the current inventory turnover situation and potential changes in market demand, predicting whether the inventory turnover speed will accelerate or slow down. Based on the predicted trends in transaction volume, identify high-potential hot-selling products (i.e., products expected to have high sales volume); based on the predicted trends in customer purchase frequency, assess which products are likely to attract more repeat customers; based on the predicted trends in inventory turnover, identify which products have good market liquidity, indicating strong demand; finally, by comprehensively considering the above factors, form a complete second potential purchase popularity distribution list. Specific example: Suppose you operate a cross-border e-commerce platform selling sportswear. Recently, you've collected the following data: Transaction volume changes: Last month, 500 units of a certain sports T-shirt were sold, but in the past two weeks, sales have gradually increased to 700 units, showing an upward trend; Customer purchase frequency information: Customers on average purchase this type of product twice a month, but after the recent promotional activities, this frequency has increased to three times a month; Inventory turnover rate information: Currently, the inventory turnover rate for this sports T-shirt is 20 days, and it is expected to accelerate to 15 days in the upcoming promotional season; Forecasted transaction volume trend: Due to continued sales growth, it is projected that the sales volume of this sports T-shirt will reach 900 units in the next sales cycle; Predicting customer purchase frequency trends: As customer interest in this product increases, the purchase frequency is expected to rise to 4 times per month; Predicting inventory turnover trends: Inventory turnover is expected to accelerate to 15 days, indicating increased product liquidity; Based on the above predictions, the following conclusions can be drawn: Second highest potential purchase interest: Sports T-shirts (estimated sales of 900 units, increased customer purchase frequency, and faster inventory turnover); Medium potential purchase interest: For example, a certain type of sports shoe, although sales are currently stable, there is no obvious growth trend, and customer frequency remains at 1.5 times / month; Low potential purchase interest: Some non-seasonal products, such as winter coats, although historically good in sales, are currently less popular due to seasonal factors.

[0029] In an optional embodiment, after controlling the cross-border e-commerce platform to optimize and adjust the supply chain operations of goods in the next sales cycle based on a third operating strategy, the method further includes: Obtain the first logistics path data of the goods; extract multiple continuous logistics nodes based on the first logistics path data, and establish multiple dynamic analysis periods for the multiple continuous logistics nodes according to preset time intervals, specifically including: Based on the first logistics path data, obtain multiple logistics stage paths with multiple consecutive logistics nodes as key nodes; Obtain the logistics area characteristics and initial inventory levels corresponding to multiple logistics stage paths; The characteristics of the logistics area and the initial inventory level are input into the logistics trend prediction model to simulate the logistics trend and generate logistics fluctuation cycles corresponding to multiple logistics stage paths. The logistics trend prediction model is trained based on historical logistics data of the same commodity type and the same logistics area. The logistics fluctuation cycle is adjusted according to a preset elastic constraint to generate a preset time interval corresponding to the logistics stage path; Configure the preset logistics fluctuation monitoring duration; The preset logistics fluctuation monitoring duration is used as the duration of multiple dynamic analysis periods, and the preset time interval is used as the time distance from multiple dynamic analysis periods to multiple continuous logistics nodes. Multiple dynamic analysis periods are configured in multiple logistics stage paths. Extract dynamic analysis periods from multiple dynamic analysis periods and obtain the corresponding logistics nodes; According to the first logistics path data, the goods are controlled to flow to the logistics nodes. When the goods are detected to enter the dynamic analysis period, logistics trend analysis is performed to generate logistics trend analysis results. Based on the logistics trend analysis results, the inventory management parameters are optimized for the first logistics stage between the dynamic analysis period and the logistics node, generating inventory optimization results.

[0030] It should be noted that this involves obtaining logistics route data for a specific product (such as an electronic product); this data includes every node in the entire transportation process of the product from the supplier to the warehouse and then to the consumer (such as the origin, various transit warehouses and the destination). Based on the acquired logistics route data, several important logistics nodes are extracted; for example: node A: supplier shipment; node B: transit warehouse 1; node C: transit warehouse 2; node D: end customer; then, a dynamic analysis period is established for each node, which can be divided based on a preset time interval (e.g., monitoring and analysis every 48 hours); For each logistics node, determine its corresponding logistics stage path, that is, the transportation process from one node to the next; at the same time, analyze the characteristics of each region (such as traffic conditions, weather, etc.) and the initial inventory level (i.e. the quantity of goods at each node). The collected logistics area characteristics and initial inventory levels are input into a logistics trend prediction model trained on historical data. This model simulates future logistics fluctuation cycles based on historical logistics data of the same product type and the same logistics area. For example, historical data shows that the round-trip turnover rate may increase in a specific period, so it can be predicted that the logistics demand for the product will fluctuate in the next sales cycle. Based on the predicted logistics fluctuation cycle, certain flexible constraints (such as maximum response time, minimum inventory level, etc.) are set, and these fluctuation cycles are adjusted to generate new preset time intervals (i.e. logistics processing capacity in each time interval). Set the monitoring duration (e.g., 24 hours) for each dynamic analysis period so that logistics data can be collected and analyzed within this period; in this way, each dynamic analysis period can be configured according to the set time interval (preset time interval); Select the first dynamic analysis period (e.g., within 48 hours from node A to node B) to obtain the information of the first corresponding logistics node; at this time, the system will start logistics trend analysis to detect any abnormalities (such as delays, stockouts, etc.). Based on the logistics trend analysis results, for specific dynamic analysis periods and logistics nodes, the inventory management parameters are optimized, such as adjusting inventory replenishment strategies and optimizing delivery routes. Finally, corresponding inventory optimization results are generated, such as reducing stockout rates or improving logistics efficiency. Specific example: Suppose you are managing the cross-border e-commerce sales of a smartphone, focusing on its logistics and delivery process; obtain logistics route data: from supplier (node ​​A) to transit warehouse 1 (node ​​B), then to transit warehouse 2 (node ​​C), and finally delivered to the customer (node ​​D); extract multiple consecutive logistics nodes: node A (shipping), node B (transit warehouse 1), node C (transit warehouse 2), node D (customer); obtain regional characteristics and initial inventory: the initial inventory of node B is 1000 phones, and the regional characteristics show that the traffic conditions in the area are good, but holidays may cause delays; input the prediction model: using data from the past three months, the model predicts the upcoming... During peak shopping seasons, demand for this product is expected to increase by 50%. Adjusting the logistics fluctuation cycle: Based on forecasts, the logistics fluctuation cycle during peak periods is set to 72 hours, with flexible constraints to handle unforeseen circumstances. Configuring monitoring duration: The monitoring duration is set to 24 hours, with dynamic analysis occurring every 72 hours. Extracting dynamic analysis periods: Logistics data analysis is performed within the first dynamic analysis period (e.g., day 1 to day 3). Optimizing inventory management: If the transportation time from node A to node B is found to be longer than expected during the dynamic analysis period, measures are taken to replenish inventory in advance, ensuring that the inventory at node B meets customer demand, ultimately optimizing inventory management results, such as reducing stockouts.

[0031] In an optional embodiment, goods are controlled to flow to logistics nodes according to first logistics path data. When goods are detected to enter a dynamic analysis period, logistics trend analysis is performed to generate logistics trend analysis results, including: Establish sample constraints for logistics trend analysis, which include a first logistics sample constraint and a second logistics sample constraint. The first logistics sample constraint is that the logistics growth rate meets a first preset logistics growth rate, and the second logistics sample constraint is that the logistics growth rate meets a second preset logistics growth rate, wherein the first preset logistics growth rate is greater than the second preset logistics growth rate. When a product is detected to have entered a dynamic analysis period, the first real-time volume of the product is obtained. Based on the first real-time logistics volume, and combined with the constraints of the first logistics sample and the second logistics sample, data on the continuous logistics growth stage is generated. Logistics trend analysis data is generated from continuous logistics growth phase data, which includes two logistics growth control nodes. Perform logistics fluctuation deviation analysis on logistics trend analysis data to generate logistics trend analysis results.

[0032] It should be noted that, based on the first logistics path data, the flow of specific goods (such as a certain electronic product) to the first logistics node (such as transit warehouse 1) is controlled; when the system detects that the goods have entered the preset first dynamic analysis period (such as within 48 hours from the first monitoring time point), logistics trend analysis will begin. Before conducting logistics trend analysis, it is necessary to establish logistics sample constraints to ensure the accuracy and effectiveness of the analysis. Sample constraints are divided into the following two categories: First logistics sample constraint: setting the logistics growth rate to reach or exceed a higher preset logistics growth rate (e.g., 10%); Second logistics sample constraint: setting the logistics growth rate to reach or exceed a lower preset logistics growth rate (e.g., 5%). Note: The first preset logistics growth rate is greater than the second preset logistics growth rate, that is, a higher growth rate is required to meet the first sample constraint. Once a product enters the dynamic analysis period, the system will obtain the first real-time flow of that product; for example, at the beginning of the dynamic analysis period, it may be found that 100 mobile phones are flowing from node A to node B. Using the first real-time logistics volume (100 mobile phones) as the base, and based on the constraints of the first and second logistics samples, continuous logistics growth stage data is generated by analyzing past logistics data. For example, if the logistics volume growth rate is 12% in the next 24 hours (satisfying the first sample constraint), then the first stage data is 112 mobile phones; if the logistics volume growth rate is 6% in the next 24 hours (satisfying the second sample constraint), then the second stage data is 118.72 mobile phones (112 mobile phones + 6%). Based on continuous logistics growth phase data, logistics trend analysis data is generated; this data may include two logistics growth control nodes, representing the logistics volume at different time stages; for example: the first control node (after 24 hours): 112 vehicles; the second control node (after 48 hours): 118.72 vehicles; The generated logistics trend analysis data is subjected to fluctuation deviation analysis. This process aims to identify the difference between the actual and expected logistics volume and to derive the logistics trend analysis results. By comparing the actual logistics volume with the preset growth rate, the effectiveness of logistics operations is evaluated. For example, if the actual observed logistics volume after 48 hours is 115 units, while the expected volume is 118.72 units, the deviation is -3.72 units, indicating that the expected volume has not been met, and further analysis of the reasons and adjustment of strategies are required. Specific example: Suppose a smartwatch is sold on a cross-border e-commerce platform. The process is as follows: Product flow control: The smartwatch is shipped from the factory (node ​​A) to transit warehouse 1 (node ​​B), entering the dynamic analysis period; Sample constraint settings: The first preset logistics growth rate is 10% (requiring a minimum logistics volume of 110 units); The second preset logistics growth rate is 5% (requiring a minimum logistics volume of 105 units); Obtain real-time logistics volume: When entering the dynamic analysis period, the real-time logistics volume is 100 units; Generate continuous logistics growth stage data: First stage: After 24 hours, the logistics volume increased by 12%, becoming 112 units (satisfying the first sample constraint); Second stage: In the next 24 hours, the logistics volume increased by 6%, becoming 118.72 units (satisfying the second sample constraint); Logistics trend analysis data generation: First control node (after 24 hours): 112 units; Second control node (after 48 hours): 118.72 units; Logistics fluctuation deviation analysis: After 48 hours, the actual logistics volume is 115 units, with a deviation of -3.72 units; Analyze possible causes, such as transportation delays or inaccurate demand forecasts, and formulate corresponding improvement measures.

[0033] In an optional embodiment, logistics trend analysis data is subjected to logistics fluctuation deviation analysis to generate logistics trend analysis results, including: Collect market consumption preference characteristics and logistics area characteristics of commodity logistics areas to model logistics trends and generate logistics trend simulation models; Input logistics trend analysis data into the logistics trend simulation model to simulate logistics trends and generate a standard time series of logistics volume changes. Perform logistics trend analysis on logistics trend analysis data and record the time sequence of actual logistics volume changes; After time and space alignment, the standard material flow rate change time series and the actual material flow rate change time series are compared to generate material flow rate deviations corresponding to multiple time and space nodes. Logistics trend analysis results are generated based on the deviations in material flow corresponding to multiple spatiotemporal nodes.

[0034] It should be noted that before modeling logistics trends, it is necessary to collect market consumer preference characteristics (such as consumer age, gender, and purchasing habits) and logistics area characteristics (such as geographical location, population density, and transportation conditions) for relevant products (e.g., a smartwatch). This data will help to better understand consumer demand and logistics conditions, thereby building a more accurate logistics trend model. Based on the collected market consumption preferences and logistics regional characteristics, a logistics trend simulation model is constructed. This model can simulate logistics flow under different conditions, such as changes in logistics demand during promotional periods, holidays, or ordinary sales periods. This model will provide a foundation for subsequent logistics trend simulations. Input the previously obtained logistics trend analysis data (e.g., the volume of 112 and 118.72 units mentioned above) into the logistics trend simulation model and run the simulation to generate a standard time series of changes in volume; this time series will represent how the volume of goods should change under normal circumstances, such as a standard forecast (e.g., 120 units, 130 units, etc.) for the next 48 hours. Within the same time frame, the actual volume of goods is recorded in time sequence; for example, the actual volume of goods may be 115, 125, and 135 units in 48 hours; these data will be compared with the standard volume of goods change time sequence. For effective comparison, the standard material flow change time series and the actual material flow change time series need to be aligned in time and space; for example, the material flow at each point in time within 48 hours can be mapped to form multiple spatiotemporal nodes (such as the material flow per hour), and then the deviation of each spatiotemporal node can be calculated separately. By comparing the standard and actual logistics volume at each spatiotemporal node, the logistics volume deviation for each node is calculated. For example, in the first hour, the standard logistics volume is 120 units, the actual volume is 115 units, and the deviation is -5 units; in the second hour, the standard logistics volume is 130 units, the actual volume is 125 units, and the deviation is -5 units; in the third hour, the standard logistics volume is 140 units, the actual volume is 135 units, and the deviation is -5 units. If this process continues, a complete deviation dataset will eventually be generated. Then, the deviations of all spatiotemporal nodes are summarized to form logistics trend analysis results, so that decision-makers can understand the performance and potential problems of logistics operations. Specific example: Continuing with the smartwatch example, the process is as follows: Market and logistics characteristic collection: Data collection reveals that demand for smartwatches from young consumers (18-30 years old) increases significantly during holidays, and city A has relatively high transportation convenience; Generating a logistics trend simulation model: A simulation model based on the above consumption and regional characteristics is established to simulate changes in logistics demand during holidays; Inputting analysis data for simulation: Logistics data of 112 and 118.72 units are input into the model to generate standard predictions: During the holidays, the logistics volume should be 120, 130, and 140 units respectively; Recording actual logistics volume changes: During the holidays... Within the 48-hour period, the actual logistics volume was 115, 125, and 135 units. Comparing the actual and standard logistics volumes: Hour 1: Standard 120 units vs. Actual 115 units (deviation -5 units); Hour 2: Standard 130 units vs. Actual 125 units (deviation -5 units); Hour 3: Standard 140 units vs. Actual 135 units (deviation -5 units). The resulting logistics trend analysis showed an overall deviation of -5 units, indicating that although the logistics volume was increasing, it did not reach the expected level. Possible reasons included inaccurate demand forecasting, insufficient inventory, or inefficient transportation, providing a basis for subsequent decision-making.

[0035] In an optional embodiment, inventory management parameters are optimized for the first logistics stage between the dynamic analysis period and the logistics node based on the logistics trend analysis results, generating inventory optimization results, including: Determine the initial inventory level and replenishment time points in the first logistics stage; Based on the logistics trend analysis results, determine whether the deviation in the volume of goods meets the preset deviation threshold; If not, after correcting the initial inventory level based on the deviation in material flow, simulate inventory consumption, optimize replenishment timing based on the forecast results, and generate inventory optimization results.

[0036] It's important to note that in the first logistics phase, it's necessary to determine the initial inventory level (e.g., the number of smartwatches currently in the warehouse) and the replenishment timeline (i.e., when to replenish inventory to avoid stockouts or overstocking). This data is typically derived from historical sales records and market forecasts. For example, suppose in the first logistics phase, the initial inventory level is 300 smartwatches, and the expected replenishment timeline is the 5th hour (i.e., when inventory drops to a certain level, replenishment will begin). Based on the logistics trend analysis results, compare the deviation between the actual volume of goods and the standard volume of goods. If the deviation exceeds the preset threshold (e.g., ±5 units), it indicates that measures need to be taken to adjust the inventory. For example, continuing with the previous example, in the first hour, the actual sales volume was 115 units, while the standard sales volume was 120 units, with a deviation of -5 units. This deviation is within the allowable threshold range, so there is no need to adjust the inventory immediately. However, if the deviation is found to have expanded to -10 units in subsequent periods, it exceeds the preset threshold, and inventory adjustment needs to be considered. If the deviation in logistics volume does not meet the preset deviation threshold, the initial inventory level needs to be adjusted. This means that the inventory level will be adjusted based on the latest sales data and trend analysis results to better match actual demand. For example, suppose the analysis shows that due to increased market demand, actual sales may be 20 more units than expected in the next few hours, so the initial inventory level is adjusted from 300 units to 320 units. Next, a predictive model is used to simulate the inventory consumption in the next few hours, such as 25 units sold per hour. Based on the simulation results of the above inventory consumption, the replenishment time point is reassessed; the goal is to ensure timely replenishment before the inventory reaches the minimum limit to avoid stockouts; for example, assuming that the simulation predicts that the inventory will drop to 290 units in the 6th hour (320 units - 25 units / hour × 2 hours), so replenishment is required in the 5th hour; if it is found in the replenishment cycle that the replenishment time needs to be brought forward, the replenishment time point can be adjusted to the 4th hour.

[0037] Through the above steps, the final inventory optimization results will include a new initial inventory level, adjusted replenishment timing, and inventory management strategies derived from the newly predicted inventory consumption rate and deviation analysis.

[0038] Example 2, please refer to Figure 2 This invention provides a technical solution: an artificial intelligence-based cross-border e-commerce supply chain data analysis system, applicable to the aforementioned artificial intelligence-based cross-border e-commerce supply chain data analysis method, comprising: The first analysis unit 1 is used to determine the first potential purchase popularity distribution of a product in the target market region based on the product transaction data of the cross-border e-commerce transaction platform within a specific sales cycle, in response to the first transaction growth rate of a product in the product transaction data exceeding a preset growth rate threshold, and based on the first market area distribution data. The second analysis unit 2 is used to obtain historical supply chain operation data of goods in the corresponding target market region based on the first potential purchase popularity distribution, and to perform association rule analysis on the historical supply chain operation data to obtain the association operation feature pattern of goods in each link of the supply chain. The third analysis unit 3 is used to adjust the first inventory control strategy for goods within a specific sales cycle based on the distribution of the first potential purchase popularity and the associated operational characteristics, and generate a second inventory control strategy; based on the second inventory control strategy, the cross-border e-commerce transaction platform is controlled to dynamically allocate the inventory of goods. The fourth analysis unit 4 is used to predict the distribution of the second potential purchase popularity of goods in the next sales cycle based on the commodity transaction data and inventory allocation data of the cross-border e-commerce transaction platform. The fifth analysis unit 5 is used to re-plan the operational strategies of the goods in each link of the cross-border e-commerce supply chain in the next sales cycle based on the distribution of the second potential purchase popularity, and generate the third operational strategy; based on the third operational strategy, the cross-border e-commerce transaction platform is controlled to optimize and adjust the supply chain operation of the goods in the next sales cycle.

[0039] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A data analysis method for cross-border e-commerce supply chain based on artificial intelligence, characterized in that: include: Based on the commodity transaction data of the cross-border e-commerce platform within a specific sales cycle, in response to the fact that the first transaction growth rate of a commodity in the commodity transaction data exceeds a preset growth rate threshold, the first potential purchase popularity distribution of the commodity in the target market region is determined based on the first market area distribution data. Based on the distribution of the first potential purchase popularity, historical supply chain operation data of the product in the corresponding target market region is obtained, and association rule analysis is performed on the historical supply chain operation data to obtain the association operation feature pattern of the product in each link of the supply chain. Based on the distribution of the first potential purchase popularity and the associated operational characteristics, the first inventory control strategy for the product within a specific sales cycle is adjusted to generate a second inventory control strategy. Based on the second inventory management strategy, the cross-border e-commerce transaction platform is controlled to dynamically allocate the inventory of the goods. Based on the commodity transaction data and inventory allocation data of the cross-border e-commerce platform, the distribution of the second potential purchase popularity of the commodity in the next sales cycle is predicted. Based on the second potential purchase popularity distribution, the operational strategies for the products in each link of the cross-border e-commerce supply chain in the next sales cycle are re-planned to generate a third operational strategy; based on the third operational strategy, the cross-border e-commerce trading platform is controlled to optimize and adjust the supply chain operation of the products in the next sales cycle.

2. The cross-border e-commerce supply chain data analysis method based on artificial intelligence according to claim 1, characterized in that, The first market area distribution data includes the sales proportion of the product in different market segments, customer evaluation feedback information, and logistics and transportation timeliness data, and the first potential purchase popularity distribution is used to characterize the intensity of potential purchase tendency of the product in different sub-regions of the target market area; Based on the first market area distribution data, determine the first potential purchase popularity distribution of the product in the target market region, including: Based on the regional distribution data of the primary market, the trend of purchasing popularity changes in each sub-market is determined along the market segmentation dimension; Based on the trends in purchasing activity across different market segments, peak purchasing activity ranges are determined. Based on the regional characteristics of the segmented market corresponding to each peak purchase popularity interval, the target market region is segmented to obtain each sub-market region within the target market region; Based on the sales share of the product in each sub-market region, customer feedback information, and logistics and transportation timeliness data, the distribution of the first potential purchase popularity is determined in each sub-market region.

3. The cross-border e-commerce supply chain data analysis method based on artificial intelligence according to claim 2, characterized in that, By analyzing historical supply chain operation data using association rules, the correlation operation characteristics of the product at each stage of the supply chain are obtained, including: Determine sample data points for the product at each stage of the supply chain during the sales cycle; Based on the historical operational information of sample data points of the product in each link of the supply chain during the sales cycle, and the market dynamic information during a specific sales cycle, the operational efficiency change ratio between two adjacent sample data points is determined. Based on the operational efficiency change ratio between two adjacent sample data points and the overall operating cost data of the product during the sales cycle, the correlation operation characteristic pattern of the product in each link of the supply chain during the sales cycle is determined.

4. The cross-border e-commerce supply chain data analysis method based on artificial intelligence according to claim 3, characterized in that, Based on the distribution of first potential purchase interest and related operational characteristics, the first inventory control strategy for the product within a specific sales cycle is adjusted to generate a second inventory control strategy, including: Prioritize the supply chain links corresponding to sub-regions with high potential purchase interest in the associated operation characteristic pattern, and use the priority ranking results to determine the inventory allocation ratio in the first inventory control strategy. Based on the current total inventory of the goods and the updated inventory allocation ratio, the inventory layout is replanned to form a second inventory control strategy.

5. The cross-border e-commerce supply chain data analysis method based on artificial intelligence according to claim 4, characterized in that, Based on commodity transaction data and inventory allocation data from cross-border e-commerce platforms, the distribution of the second potential purchase popularity of the commodity in the next sales cycle is predicted, including: Based on transaction volume changes, customer purchase frequency, and inventory turnover speed information from cross-border e-commerce transaction data, the trends of transaction volume, customer purchase frequency, and inventory turnover speed of the product in the next sales cycle are predicted. The distribution of the second potential purchase interest is determined based on the predicted trends in transaction volume, customer purchase frequency, and inventory turnover.

6. The cross-border e-commerce supply chain data analysis method based on artificial intelligence according to claim 5, characterized in that, After optimizing and adjusting the supply chain operations of the goods on the cross-border e-commerce platform in the next sales cycle based on the third operational strategy, the method further includes: Obtain the first logistics path data of the goods; extract multiple continuous logistics nodes based on the first logistics path data, and establish multiple dynamic analysis periods for the multiple continuous logistics nodes according to preset time intervals, specifically including: Based on the first logistics path data, obtain multiple logistics stage paths with multiple consecutive logistics nodes as key nodes; Obtain the logistics area characteristics and initial inventory levels corresponding to multiple logistics stage paths; The characteristics of the logistics area and the initial inventory level are input into the logistics trend prediction model to simulate the logistics trend and generate logistics fluctuation cycles corresponding to multiple logistics stage paths. The logistics trend prediction model is trained based on historical logistics data of the same commodity type and the same logistics area. The logistics fluctuation cycle is adjusted according to a preset elastic constraint to generate a preset time interval corresponding to the logistics stage path; Configure the preset logistics fluctuation monitoring duration; The preset logistics fluctuation monitoring duration is used as the duration of multiple dynamic analysis periods, and the preset time interval is used as the time distance from multiple dynamic analysis periods to multiple continuous logistics nodes. Multiple dynamic analysis periods are configured in multiple logistics stage paths. Extract dynamic analysis periods from multiple dynamic analysis periods and obtain the corresponding logistics nodes; According to the first logistics path data, the goods are controlled to flow to the logistics nodes. When the goods are detected to enter the dynamic analysis period, logistics trend analysis is performed to generate logistics trend analysis results. Based on the logistics trend analysis results, the inventory management parameters are optimized for the first logistics stage between the dynamic analysis period and the logistics node, generating inventory optimization results.

7. The cross-border e-commerce supply chain data analysis method based on artificial intelligence according to claim 6, characterized in that, Based on the data from the first logistics path, the flow of goods to logistics nodes is controlled. When goods are detected entering a dynamic analysis period, logistics trend analysis is performed, and logistics trend analysis results are generated, including: Establish logistics trend analysis sample constraints, wherein the logistics trend analysis sample constraints include a first logistics sample constraint and a second logistics sample constraint. The first logistics sample constraint is that the logistics growth rate meets a first preset logistics growth rate, and the second logistics sample constraint is that the logistics growth rate meets a second preset logistics growth rate, wherein the first preset logistics growth rate is greater than the second preset logistics growth rate. When a product is detected to have entered a dynamic analysis period, the first real-time volume of the product is obtained. Based on the first real-time logistics volume, and combined with the constraints of the first logistics sample and the second logistics sample, data on the continuous logistics growth stage is generated. Logistics trend analysis data is generated from continuous logistics growth phase data, which includes two logistics growth control nodes. Perform logistics fluctuation deviation analysis on logistics trend analysis data to generate logistics trend analysis results.

8. The cross-border e-commerce supply chain data analysis method based on artificial intelligence according to claim 7, characterized in that, Perform logistics fluctuation deviation analysis on logistics trend analysis data to generate logistics trend analysis results, including: Collect market consumption preference characteristics and logistics area characteristics of commodity logistics areas to model logistics trends and generate logistics trend simulation models; Input logistics trend analysis data into the logistics trend simulation model to simulate logistics trends and generate a standard time series of logistics volume changes. Perform logistics trend analysis on logistics trend analysis data and record the time sequence of actual logistics volume changes; After time and space alignment, the standard material flow rate change time series and the actual material flow rate change time series are compared to generate material flow rate deviations corresponding to multiple time and space nodes. Logistics trend analysis results are generated based on the deviations in material flow corresponding to multiple spatiotemporal nodes.

9. The cross-border e-commerce supply chain data analysis method based on artificial intelligence according to claim 8, characterized in that, Based on the logistics trend analysis results, inventory management parameters are optimized for the first logistics stage between the dynamic analysis period and logistics nodes, generating inventory optimization results, including: Determine the initial inventory level and replenishment time points in the first logistics stage; Based on the logistics trend analysis results, determine whether the deviation in the volume of goods meets the preset deviation threshold; If not, after correcting the initial inventory level based on the deviation in material flow, simulate inventory consumption, optimize replenishment timing based on the forecast results, and generate inventory optimization results.

10. An AI-based cross-border e-commerce supply chain data analysis system, applicable to the AI-based cross-border e-commerce supply chain data analysis method described in any one of claims 1-9, characterized in that, include: The first analysis unit is used to determine the first potential purchase popularity distribution of the product in the target market region based on the product transaction data of the cross-border e-commerce transaction platform within a specific sales cycle, in response to the first transaction growth rate of a product in the product transaction data exceeding a preset growth rate threshold, and based on the first market area distribution data. The second analysis unit is used to obtain historical supply chain operation data of the product in the corresponding target market region based on the first potential purchase popularity distribution, and to perform association rule analysis on the historical supply chain operation data to obtain the association operation feature pattern of the product in each link of the supply chain. The third analysis unit is used to adjust the first inventory control strategy for the product within a specific sales cycle based on the first potential purchase popularity distribution and related operational characteristic patterns, and generate a second inventory control strategy. Based on the second inventory management strategy, the cross-border e-commerce transaction platform is controlled to dynamically allocate the inventory of the goods. The fourth analysis unit is used to predict the second potential purchase popularity distribution of the product in the next sales cycle based on the product transaction data and inventory allocation data of the cross-border e-commerce transaction platform. The fifth analysis unit is used to re-plan the operational strategies of the product in each link of the cross-border e-commerce supply chain in the next sales cycle based on the second potential purchase popularity distribution, and generate a third operational strategy; based on the third operational strategy, the cross-border e-commerce transaction platform is controlled to optimize and adjust the supply chain operation of the product in the next sales cycle.

Images