Inventory dynamic balancing method based on sales law prediction and procurement optimization system
By analyzing historical data to predict the number of new customers and personalize displays, the problems of inaccurate inventory forecasting and unmet customer needs were solved, achieving inventory balance and maximizing sales.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NANJING DONGPIN TECHNOLOGY CO LTD
- Filing Date
- 2026-04-20
- Publication Date
- 2026-07-24
Smart Images
Figure CN122072881B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the fields of inventory balancing and procurement optimization technology, specifically to a dynamic inventory balancing method and procurement optimization system based on sales pattern forecasting. Background Technology
[0002] Addressing the inefficiency caused by inaccurate sales forecasting in retail inventory management, and automating inventory demand calculation through dynamic analysis of sales patterns and providing intelligent procurement suggestions are crucial steps in achieving inventory balance and maximizing sales. Therefore, this application proposes a dynamic inventory balancing method and procurement optimization system based on sales pattern forecasting.
[0003] Existing technology, such as the invention patent with authorization announcement number CN110322203B, discloses a method for optimizing inventory analysis in the retail industry, including the following steps: Step S1, determining safety stock based on customer demand, order lead time, and customer service level; Step S2, calculating the reorder point based on the safety stock and the average daily demand within the order lead time; Step S3, using an autoregressive model to predict future sales based on historical sales data, thereby determining the order quantity; Step S4, for multi-product sales forecasting where multiple individual items have intertwined multi-dimensional relationships, improving forecast accuracy through multi-level forecast adjustments.
[0004] Existing technology, such as the invention patent with authorization announcement number CN116843419A, discloses a retail e-commerce system and management method. The retail e-commerce system includes: a product management device configured to manage product information; a supplier management device connected to the product management device, configured to manage supplier information; a warehouse management device connected to the supplier management device, configured to manage warehouse information; a procurement management device connected to the warehouse management device, configured to manage procurement information; and a store device connected to the product management device, configured to manage store information and system settings.
[0005] The above solution has the following technical problems: 1. The current technology only predicts the safety stock and order quantity based on the forecast of future sales volume, without taking into account the analysis of the number of new customers in the future. The influx of new customers will cause fluctuations in the forecast of future sales volume at the sales site. The current technology's neglect of this aspect will lead to inaccurate inventory forecasts.
[0006] 2. Current technology for managing retail goods does not take into account the personalized needs of each customer. Since different customers have different price acceptance and quality requirements, their purchasing needs for different goods vary. The current technology's neglect of this aspect leads to the inability to maximize the sales volume of goods at sales sites and is also detrimental to customer evaluation of sales sites. Summary of the Invention
[0007] The purpose of this application is to provide a dynamic inventory balancing method and procurement optimization system based on sales pattern prediction, which solves the problems existing in the background technology.
[0008] To solve the above-mentioned technical problems, this application adopts the following technical solution: In the first aspect, this application provides a method for dynamic inventory balancing based on sales pattern prediction, including: Step 1, obtaining commodity sales data, commodity inventory data, customer transaction data and basic information of sales sites for each historical period from the data storage center.
[0009] Step 2: Based on customer transaction data from each historical period, predict the number of new customers for the current period and determine whether to update inventory demand. Then, analyze the inventory demand for each product in the current period based on product sales data, product inventory data, and customer transaction data from each historical period.
[0010] Step 3: Analyze customer transaction data, product sales data, and product inventory data from each historical period to obtain the product display page for each historical customer in the current period. Then, analyze product sales data and product inventory data from each historical period to obtain the product display page for each new customer in the current period.
[0011] Step 4: After the current cycle ends, determine whether the inventory demand analysis and display page analysis for the current cycle are reasonable.
[0012] In its second aspect, this application provides a dynamic inventory balancing method and procurement optimization system based on sales pattern prediction, including: a data acquisition module for acquiring historical sales data, inventory data, customer transaction data, and basic information of sales sites from a data storage center.
[0013] Inventory Analysis Module: This module is used to predict the number of new customers in the current period based on customer transaction data from each historical period, and to determine whether to update inventory demand. It then analyzes the inventory demand of each product in the current period based on product sales data, product inventory data, and customer transaction data from each historical period.
[0014] Retail Analytics Module: This module analyzes customer transaction data, product sales data, and product inventory data from each historical period to obtain the product display page for each historical customer in the current period. It also analyzes product sales data and product inventory data from each historical period to obtain the product display page for each new customer in the current period.
[0015] Feedback module: After the current cycle ends, it determines whether the inventory demand analysis and display page analysis for the current cycle are reasonable.
[0016] The beneficial effects of this application are as follows: 1. The inventory dynamic balancing method and procurement optimization system based on sales pattern prediction provided in this application predicts the number of new customers in the current period by analyzing customer transaction data from each historical period. It then analyzes the inventory demand of each product in the current period by combining historical sales data, inventory data, and customer transaction data. Furthermore, it obtains the product display pages for each historical customer and each new customer in the current period based on the same historical data. Finally, it judges whether the inventory demand analysis and display page analysis for the current period are reasonable. This application reduces inventory risk by combining multi-dimensional dynamic data for inventory prediction, and simultaneously creates personalized product display pages for each customer, which helps strengthen customer loyalty and maximize retail volume.
[0017] 2. This application forecasts the number of new customers in the current period, and then analyzes the inventory demand for the current period based on the number of new customers and sales data from each historical period. This generates automated procurement suggestions, achieving inventory balance and maximizing sales. The method of forecasting inventory demand by combining multi-dimensional data avoids the limitations of traditional inventory forecasting, reduces operational risks, and optimizes the allocation of commodity resources.
[0018] 3. This application analyzes the price and quality of goods purchased by each historical customer to obtain the weighting factors of each customer for product price and product quality. This analysis then yields the product display pages for each historical user. Since different customers have varying acceptance levels for product prices and requirements for product quality, this personalized product display page analysis can accurately match the individual shopping needs of each historical customer, improve customer conversion efficiency, strengthen customer loyalty to the sales site, reduce customer churn due to unmet shopping needs, and quickly establish trust with customers, thereby maximizing retail sales. Attached Figure Description
[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0020] Figure 1 This is a flowchart illustrating the implementation steps of the method described in this application.
[0021] Figure 2 This is an illustration of a product display page.
[0022] Figure 3 This is a schematic diagram of the system structure connection of this application. Detailed Implementation
[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.
[0024] Reference Figure 1 As shown, this application provides a method for dynamic inventory balancing based on sales pattern forecasting in the first aspect, including the following steps: Step 1: Obtain commodity sales data, commodity inventory data, customer transaction data and basic information of sales sites for each historical period from the data storage center.
[0025] It should be noted that the specific length of a historical cycle is determined by the relevant staff and is not subject to any specific restrictions here.
[0026] In a specific example, the product sales data includes the transaction amount, transaction quantity, cancellation amount, and cancellation quantity for each type of product.
[0027] The commodity inventory data includes the inventory demand, sales volume, and final inventory of each type of commodity.
[0028] The customer transaction data includes historical customer transaction data and new customer transaction data. The historical customer transaction data includes the number of times a customer visits a sales site, the number of times a product is traded, the number of times a product is returned, the transaction amount for each type of product, the amount of a product returned, and the number of times a product is visited.
[0029] New customer transaction data includes the total number of customer visits to sales sites, the number of product transactions, the number of product cancellations, the transaction amount for each type of product, the cancellation amount for each type of product, the number of visits for each type of product, and the number of new customers.
[0030] The basic information of the sales sites includes information on the sources of new customers, operational information of the sales sites, and information on the external environment.
[0031] It should be noted that the inventory demand refers to the initial inventory level for each period.
[0032] It should be noted that customer source information refers to the user's source channel, such as search engines, social media, advertising, external partnership links, and direct access; website operation information includes website marketing activities, feature adjustments, and product iteration information, among which website marketing activities include the timing, intensity, and coverage of activities such as coupon distribution, live streaming sales, and member days; feature adjustments include registration process optimization, new user benefits, and page loading speed improvements; product iteration information includes new product launches, category expansion, and other actions that may attract new users; external environment information includes e-commerce promotions and holidays.
[0033] Step 2: Based on customer transaction data from each historical period, predict the number of new customers for the current period and determine whether to update inventory demand. Then, analyze the inventory demand for each product in the current period based on product sales data, product inventory data, and customer transaction data from each historical period.
[0034] In a specific example, the process of predicting the number of new customers in the current period based on customer transaction data of each historical period is as follows: S11, obtain the number of new customers in each historical period based on customer transaction data, and then obtain the source information of each new customer, the operation information of the sales site and the external environment information of each historical period based on the basic information of the sales site.
[0035] S12. After preprocessing the number of new customers, the source information of new customers, the operation information of sales sites and the external environment information of each historical period, the data are divided into training set and validation set according to a preset ratio. Then, the preset time series model is trained and validated to obtain the new customer prediction model of the sales site.
[0036] It should be noted that preprocessing includes outlier handling, data completion, and time series splitting. Outlier handling, data completion, and time series splitting are all existing technologies, so they will not be described in detail here.
[0037] It should be noted that the time series model is the LightGBM model, which is an existing prediction model that can be directly trained and used, so it will not be described in detail here.
[0038] It should also be noted that the preset ratio is 8:2.
[0039] S13. Input the new customer data from the previous historical period, the data from each new customer source, the sales site operation data and the external environment data, as well as the sales site operation data and the external environment data for the current period into the new customer prediction model, and then predict the number of new customers for the current period.
[0040] In a specific instance, the determination of whether to update the inventory demand is as follows: if the number of new customers in the current period is 0 and the final inventory of each type of product in the previous historical period is not 0, then the inventory demand will not be updated in the current period.
[0041] Conversely, if the number of new customers in the current period is not zero, the sales volume of each type of product in the previous historical period is compared with the corresponding inventory demand. If the sales volume of a certain type of product in the previous historical period is less than half of the corresponding inventory demand, then that type of product is recorded as a second-tier best-selling product; otherwise, it is recorded as a first-tier best-selling product. Based on this, the first-tier best-selling products and the second-tier best-selling products are obtained, and the inventory demand for each first-tier best-selling product and the second-tier best-selling products is updated.
[0042] In a specific example, the analysis of the inventory demand of each commodity in the current period based on the commodity sales data, commodity inventory data and customer transaction data of each historical period is as follows: S21, compare the number of new customers in the current period with the number of new customers in each historical period, and record the data with the largest ratio as the first-level inventory increment. At the same time, record the ratio of the number of new customers in the current period to the number of new customers in the previous historical period as the second-level inventory increment.
[0043] S22. Products whose final inventory in the previous historical period is less than the highest sales volume in the corresponding historical periods are classified as first-level demand purchase products; otherwise, they are classified as second-level demand purchase products.
[0044] S23. Use the inventory demand of each type of commodity at the sales station in the previous historical period as the base inventory of each type of commodity in the current period; for each first-tier best-selling commodity, increase the base inventory according to the first-tier inventory increment to obtain the inventory demand of each first-tier best-selling commodity in the current period; for each second-tier best-selling commodity, when a second-tier best-selling commodity is a first-tier demand purchase commodity, increase the base inventory of the second-tier best-selling commodity according to the first-tier inventory increment to obtain the inventory demand of the second-tier best-selling commodity in the current period; when a second-tier best-selling commodity is a second-tier demand purchase commodity, increase the base inventory of the second-tier best-selling commodity according to the second-tier inventory increment to obtain the inventory demand of the second-tier best-selling commodity in the current period; thereby obtaining the inventory demand of each first-tier best-selling commodity and each second-tier best-selling commodity.
[0045] Reference Figure 2 As shown, step three involves analyzing customer transaction data, product sales data, and product inventory data from each historical period to obtain the product display page for each historical customer in the current period, and then analyzing product sales data and product inventory data from each historical period to obtain the product display page for each new customer in the current period.
[0046] In a specific example, the process of analyzing customer transaction data, product sales data, and product inventory data from each historical period to obtain the product display page for each historical customer in the current period is as follows: S31, First, based on the customer transaction data from each historical period, the weighting factor of the corresponding product price for each historical customer is obtained. Weighting factors for product quality , where j represents the historical customer number, and j is a positive integer.
[0047] S32. Then, based on the sales data, obtain the transaction amount and number of returned orders for each type of product, and record them as follows: and Where i represents the product category number and f represents the product number, both i and f are positive integers, calculated according to the formula: The attractiveness feature value of product number f in category i to historical customer number j was calculated. ,in and Let represent the average transaction amount and the average number of order cancellations for the i-th type of goods, respectively.
[0048] S33. Finally, retrieve the inventory quantities of each type of product at each historical user's visit time point in the current period from the data center, and record them as follows: After standardizing the inventory quantity of each type of commodity and the attractiveness characteristics of each type of commodity to each historical customer, the calculation formula is as follows: The display characteristic value of the historical customer with ID j corresponding to product ID f in category i was calculated. Then, the display feature values of each type of product corresponding to each historical customer are sorted from largest to smallest to obtain the display page of each type of product corresponding to each historical customer.
[0049] In a specific example, the weighting factor for the commodity price corresponding to each historical customer is obtained by analyzing customer transaction data from each historical period. Weighting factors for product quality The specific analysis process is as follows: S41. First, based on the customer transaction data of each historical period, obtain the transaction amount, number of transactions, and number of order cancellations for each historical customer for each type of product. Then, calculate the average transaction amount for each historical customer for each type of product based on the transaction amount and number of transactions for each historical customer for each type of product. Then, through the calculation formula: The price utility characteristic value of the historical user with ID j for the i-th type of goods was calculated. ,in This represents the maximum transaction amount for the i-th type of goods; simultaneously, the cancellation rate for each historical customer's category and type of goods is calculated based on the number of transactions and cancellations for each historical customer. The cancellation rate is the quotient of the number of cancellations and the number of transactions, and is calculated using the following formula: The quality utility characteristic value of the historical user with ID j for the i-th type of goods was calculated. .
[0050] S42. Combining the price utility characteristic values and therapeutic utility characteristic values of various historical users for different types of goods, according to the calculation formula: The overall utility characteristic value of the historical user with ID j for the i-th type of goods was calculated. .
[0051] S43. Based on customer transaction data for each historical period, obtain the transaction volume of each historical customer for each type of product in the previous historical period, and then calculate the order volume ratio of each historical customer for each type of product in the previous historical period. Simultaneously, the comprehensive utility characteristic values of each historical user for various types of goods in the previous historical period were calculated. This allows for the calculation of the ratio of the overall utility characteristics of each historical customer for various types of goods in the previous historical period. Through the Solve to obtain the price weighting factor and quality weighting factor for the corresponding products of each historical user. and .
[0052] It should be noted that, .
[0053] It should be noted that, for example, a sales site has three categories of goods: A, B, and C. The highest-priced items in categories A, B, and C are 50 yuan, 200 yuan, and 155 yuan, respectively. A historical customer with the order number j placed a total of 2000 orders on the sales site in the previous historical period, and the average transaction amount for categories A, B, and C was 5 yuan, 100 yuan, and 155 yuan, respectively. Therefore, the price utility eigenvalues for categories A, B, and C are 0.9, 0.5, and 0, respectively. The order cancellation rates for categories A, B, and C are 15%, 8%, and 3%, respectively. Therefore, the quality utility eigenvalues for categories A, B, and C are 0.85, 0.92, and 0.97, respectively. The order volume proportions for categories A, B, and C are 50%, 30%, and 20%, respectively. Then, a comprehensive analysis can yield the following results: ,and Therefore, it can be calculated that , The examples above are merely illustrative and not the only possible ones.
[0054] In a specific example, the product display page for each new customer in the current period is obtained by analyzing the product sales data and product inventory data of each historical period. The specific analysis process is as follows: When each new customer visits the sales site, the product inventory data and product sales data at each time point of the visit are obtained from the data center. Based on the product inventory data, the inventory quantity of each type of product on the sales website at each visit time point is obtained. Based on the product sales data, the transaction amount and number of returned orders for each type of product in the current period are obtained. After standardizing the inventory quantity, transaction amount, and number of returned orders for each type of product, they are recorded as follows: , and Substitute into the calculation formula: The calculation yields the product of category i with number f for product number f. New customer display characteristics ,in For new customers' numbers, The value is a positive integer. The display feature value of each type of product for each new customer is sorted from largest to smallest to obtain the product display page for each new customer.
[0055] Step 4: After the current cycle ends, determine whether the inventory demand analysis and display page analysis for the current cycle are reasonable.
[0056] In a specific example, the process for determining whether the inventory demand analysis and display page analysis for the current period are reasonable is as follows: Obtain the product inventory data for the current period from the data storage center. If the final inventory of each type of product is greater than 0, it indicates that the inventory demand analysis for the sales site in the current period is reasonable; otherwise, it indicates that the inventory demand analysis for the sales site in the current period is unreasonable.
[0057] Simultaneously, customer transaction data for the current period is obtained from the data center. Based on this data, the number of transactions and cancellations for each type of product are calculated, and the cancellation rate for each type of product in the current period is calculated. Similarly, the cancellation rate for each type of product in the previous historical period is calculated. The cancellation rate for each type of product in the current period is compared with the cancellation rate for each type of product in the corresponding previous historical period. If the total number of products with a cancellation rate lower than that in the current period is greater than half of the total number of products in the sales site, then the analysis of the product display page for the current period is considered reasonable. Otherwise, the analysis of the product display page for the sales site in the current period is considered unreasonable.
[0058] It should be noted that the number of order cancellations and the order cancellation rate mentioned in this application refer to the number of order cancellations and the order cancellation rate caused by product quality issues.
[0059] Reference Figure 3 As shown, in the second aspect, this application provides a procurement optimization system based on sales pattern prediction, including the following modules: Data acquisition module: used to acquire commodity sales data, commodity inventory data, customer transaction data and basic information of sales sites for each historical period from the data storage center.
[0060] Inventory Analysis Module: This module is used to predict the number of new customers in the current period based on customer transaction data from each historical period, and to determine whether to update inventory demand. It then analyzes the inventory demand of each product in the current period based on product sales data, product inventory data, and customer transaction data from each historical period.
[0061] Retail Analytics Module: This module analyzes customer transaction data, product sales data, and product inventory data from each historical period to obtain the product display page for each historical customer in the current period. It also analyzes product sales data and product inventory data from each historical period to obtain the product display page for each new customer in the current period.
[0062] Feedback module: After the current cycle ends, it determines whether the inventory demand analysis and display page analysis for the current cycle are reasonable.
[0063] This application provides a dynamic inventory balancing method and procurement optimization system based on sales pattern forecasting. It predicts the number of new customers in the current period using customer transaction data from various historical periods, and analyzes the inventory demand for each product in the current period by combining historical sales data, inventory data, and customer transaction data. Furthermore, it generates product display pages for each historical customer and each new customer based on the same historical data. Finally, it assesses the rationality of the current period's inventory demand and display page analysis. This application reduces inventory risk by combining multi-dimensional dynamic data for inventory forecasting, and by creating personalized product display pages for each customer, it helps strengthen customer loyalty and maximize retail volume.
[0064] The above content is merely an example and illustration of the concept of this application. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the inventive concept or exceed the scope defined in this application, they should all fall within the protection scope of this application.
Claims
1. A dynamic inventory balancing method based on sales pattern forecasting, characterized in that, include: Step 1: Obtain historical sales data, inventory data, customer transaction data, and basic information of sales sites from the data storage center for each historical period; Step 2: Based on customer transaction data from each historical period, predict the number of new customers for the current period and determine whether to update inventory demand. Then, analyze the inventory demand for each product in the current period based on product sales data, product inventory data, and customer transaction data from each historical period. The specific analysis process is as follows: S21. Compare the number of new customers in the current period with the number of new customers in each historical period, and record the data with the largest ratio as the first-level inventory increment. At the same time, record the ratio of the number of new customers in the current period to the number of new customers in the previous historical period as the second-level inventory increment. S22. Products whose final inventory of each type of commodity in the previous historical period is less than the highest sales volume in the corresponding historical periods are classified as first-level demand procurement commodities. Conversely, goods purchased under secondary demand are recorded as such. S23. Use the inventory demand of each type of commodity at the sales site in the previous historical period as the base inventory of each type of commodity in the current period. For each top-selling product, the base inventory is increased according to the increase in top-selling inventory to obtain the inventory demand for each top-selling product in the current period; for each second-selling product, when a second-selling product is a top-selling product, the base inventory of that second-selling product is increased according to the increase in top-selling inventory to obtain the inventory demand for that second-selling product in the current period. When a certain second-tier best-selling product is a second-tier demand product, the base inventory of the second-tier best-selling product is increased according to the second-tier inventory increment to obtain the inventory demand of the second-tier best-selling product in the current period; based on this, the inventory demand of each first-tier best-selling product and each second-tier best-selling product is obtained. Step 3: Based on the customer transaction data, product sales data, and product inventory data of each historical period, obtain the product display page for each historical customer in the current period. Then, based on the product sales data and product inventory data of each historical period, obtain the product display page for each new customer in the current period. The specific process is as follows: S31. First, based on the analysis of customer transaction data for each historical period, the weighting factors for the corresponding product prices of each historical customer are obtained. Weighting factors for product quality , where j represents the historical customer ID, and j is a positive integer; S32. Then, based on the sales data, obtain the transaction amount and number of returned orders for each type of product, and record them as follows: and Where i represents the product category number and f represents the product number, both i and f are positive integers, calculated according to the formula: The attractiveness feature value of product number f in category i to historical customer number j was calculated. ,in and Let these represent the average transaction amount and the average number of returned orders for the i-th type of goods, respectively. S33. Finally, retrieve the inventory quantities of each type of product at each historical user's visit time point in the current period from the data center, and record them as follows: After standardizing the inventory quantity of each type of commodity and the attractiveness characteristics of each type of commodity to each historical customer, the calculation formula is as follows: The display characteristic value of the historical customer with ID j corresponding to product ID f in category i was calculated. Then, sort the display feature values of each type of product and each historical customer from largest to smallest to obtain the display page of each type of product and each historical customer. Step 4: After the current cycle ends, determine whether the inventory demand analysis and display page analysis for the current cycle are reasonable.
2. The inventory dynamic balancing method based on sales pattern forecasting according to claim 1, characterized in that, The product sales data includes the transaction amount, transaction quantity, refund amount, and refund quantity for each type of product; The commodity inventory data includes the inventory demand, sales quantity, and final inventory of each type of commodity; The customer transaction data includes historical customer transaction data and new customer transaction data. The historical customer transaction data includes the number of times a customer visits a sales site, the number of times a product is traded, the number of times a product is returned, the transaction amount for each type of product, the amount of each type of product returned, and the number of times a product is visited. New customer transaction data includes the total number of customer visits to sales sites, the number of product transactions, the number of product returns, the transaction amount for each type of product, the return amount for each type of product, the number of visits for each type of product, and the number of new customers; The basic information of the sales sites includes information on the sources of new customers, operational information of the sales sites, and information on the external environment.
3. The inventory dynamic balancing method based on sales pattern forecasting according to claim 2, characterized in that, The specific process for predicting the number of new customers in the current period based on customer transaction data from previous periods is as follows: S11. Obtain the number of new customers for each historical period based on customer transaction data, and then obtain the source information of each new customer, sales site operation information and external environment information for each historical period based on the basic information of the sales site. S12. After preprocessing the number of new customers, the source information of new customers, the operation information of sales sites and the external environment information of each historical period, the data are divided into training set and validation set according to a preset ratio. Then, the preset time series model is trained and validated to obtain the new customer prediction model of the sales site. S13. Input the new customer data from the previous historical period, the data from each new customer source, the sales site operation data and the external environment data, as well as the sales site operation data and the external environment data for the current period into the new customer prediction model, and then predict the number of new customers for the current period.
4. The inventory dynamic balancing method based on sales pattern forecasting according to claim 3, characterized in that, The specific process for determining whether to update the inventory demand is as follows: If the number of new customers in the current period is 0, and the final inventory of each type of product in the previous historical period is not 0, then the inventory demand will not be updated in the current period. Conversely, if the number of new customers in the current period is not zero, the sales volume of each type of product in the previous historical period is compared with the corresponding inventory demand. If the sales volume of a certain type of product in the previous historical period is less than half of the corresponding inventory demand, then that type of product is recorded as a second-tier best-selling product; otherwise, it is recorded as a first-tier best-selling product. Based on this, the first-tier best-selling products and the second-tier best-selling products are obtained, and the inventory demand for each first-tier best-selling product and the second-tier best-selling products is updated.
5. The inventory dynamic balancing method based on sales pattern forecasting according to claim 4, characterized in that, The weighting factors for the corresponding product prices of each historical customer are obtained by analyzing customer transaction data from each historical period. Weighting factors for product quality The specific analysis process is as follows: S41. First, based on the customer transaction data for each historical period, obtain the transaction amount, number of transactions, and number of order cancellations for each historical customer for each type of product. Then, calculate the average transaction amount for each historical customer for each type of product based on the transaction amount and number of transactions for each historical customer for each type of product. Then, through the calculation formula: The price utility characteristic value of the historical user with ID j for the i-th type of goods was calculated. ,in This represents the maximum transaction amount for the i-th type of goods; simultaneously, the cancellation rate for each historical customer's category and type of goods is calculated based on the number of transactions and cancellations for each historical customer. The cancellation rate is the quotient of the number of cancellations and the number of transactions, and is calculated using the following formula: The quality utility characteristic value of the historical user with ID j for the i-th type of goods was calculated. ; S42. Combining the price utility characteristic values and therapeutic utility characteristic values of various historical users for different types of goods, according to the calculation formula: The overall utility characteristic value of the historical user with ID j for the i-th type of goods was calculated. ; S43. Based on customer transaction data for each historical period, obtain the transaction volume of each historical customer for each type of product in the previous historical period, and then calculate the order volume ratio of each historical customer for each type of product in the previous historical period. Simultaneously, the comprehensive utility characteristic values of each historical user for various types of goods in the previous historical period were calculated. This allows for the calculation of the ratio of the overall utility characteristics of each historical customer for various types of goods in the previous historical period. Through the Solve to obtain the price weighting factor and quality weighting factor for the corresponding products of each historical user. and .
6. The inventory dynamic balancing method based on sales pattern forecasting according to claim 5, characterized in that, The product display page for each new customer in the current period is then obtained by analyzing the product sales data and product inventory data from each historical period. The specific analysis process is as follows: When new customers visit sales sites, product inventory and sales data for each visit point are retrieved from the data center. Based on the inventory data, the inventory quantity of each product category at each visit point is calculated. Based on the sales data, the transaction amount and number of returned orders for each product category in the current period are calculated. After standardizing the inventory quantity, transaction amount, and number of returned orders for each product category, they are recorded as follows: , and Substitute into the calculation formula: The calculation yields the product of category i with number f for product number f. New customer display characteristics ,in For new customers' numbers, The value is a positive integer. The display feature value of each type of product for each new customer is sorted from largest to smallest to obtain the product display page for each new customer.
7. The inventory dynamic balancing method based on sales pattern forecasting according to claim 6, characterized in that, The specific judgment process for determining whether the current cycle's inventory demand analysis and product display page analysis are reasonable is as follows: If the final inventory of each type of product is greater than 0, it indicates that the inventory demand analysis of the sales site in the current period is reasonable; otherwise, it indicates that the inventory demand analysis of the sales site in the current period is unreasonable. Simultaneously, customer transaction data for the current period is obtained from the data center. Based on this data, the number of transactions and cancellations for each type of product are calculated, and the cancellation rate for each type of product in the current period is calculated. Similarly, the cancellation rate for each type of product in the previous historical period is calculated. The cancellation rate for each type of product in the current period is compared with the cancellation rate for each type of product in the corresponding previous historical period. If the total number of products with a cancellation rate less than that of the previous period is greater than half of the total number of products in the sales site, then the analysis of the product display page for the current period is considered reasonable. Otherwise, the analysis of the product display page for the sales site in the current period is considered unreasonable.
8. A procurement optimization system based on sales pattern forecasting, used to execute the dynamic inventory balancing method based on sales pattern forecasting as described in any one of claims 1-7, characterized in that, include: Data acquisition module: used to acquire historical sales data, inventory data, customer transaction data and basic information of sales sites from the data storage center for each period; Inventory Analysis Module: Used to predict the number of new customers in the current period based on customer transaction data from each historical period, and to determine whether to update inventory demand. Then, based on product sales data, product inventory data and customer transaction data from each historical period, it analyzes the inventory demand of each product in the current period. Retail Analytics Module: This module analyzes customer transaction data, product sales data, and product inventory data from each historical period to obtain the product display page for each historical customer in the current period, and then analyzes product sales data and product inventory data from each historical period to obtain the product display page for each new customer in the current period. Feedback module: After the current cycle ends, it determines whether the inventory demand analysis and display page analysis for the current cycle are reasonable.