Logistics lightly-thrown customer identification and application auditing sorting method and device, equipment and storage medium

Abstract

The invention relates to the technical field of logistics data processing, in particular to a logistics light throwing customer identification and application auditing sorting method and device, equipment and a storage medium. The method comprises the following steps: firstly, obtaining a submission application, combining a client code and a website code in the submission application, performing matching verification with all historical submission records, judging whether the submission is repeated or not, and outputting a first judgment result; if not, acquiring customer data, judging whether the customer is a lightly-thrown customer according to the customer data, and outputting a second judgment result; if the client is a lightly-thrown client, extracting historical delivery data, industry seasonal characteristics and a logistics fade and peak season coefficient of the client to predict the delivery amount of the client in a period of time in the future, and taking the delivery amount as a reference delivery amount; and finally, according to the reference document quantity and the customer level, the product attribute and the website priority in the submission application, grading, auditing and sorting the priority of the submission application. Priority services are effectively carried out on the light-throwing clients, and the service accuracy of the core clients is improved.

Description

Technical Field

[0001] This invention relates to the field of logistics data processing technology, and in particular to a method, apparatus, equipment, and storage medium for identifying and sorting customer applications for light-duty logistics. Background Technology

[0002] Currently, logistics companies primarily manage footwear and apparel (lightweight and disposable) customers through branch reporting, which leads to difficulties in accurately identifying and prioritizing applications from these customers, making it hard to meet the demand for refined services.

[0003] Currently, the system's customer attribute definition relies solely on basic field matching, resulting in insufficient accuracy. This leads to frequent misreporting of non-footwear and apparel customers, directly distorting the screening of lightweight customers. Furthermore, the review process uses a fixed workflow sequence without prioritizing based on factors such as order volume and value. This results in inefficient processing of applications from high-volume, high-value lightweight customers, causing significant review delays. Additionally, the order volume reference only uses historical data from T-1 day, failing to predict future shipping trends for lightweight customers and leading to a lack of foresight in service resource allocation.

[0004] The aforementioned problems combined result in low efficiency in managing light-weight and bulky goods customers and insufficient accuracy in service matching. This makes it impossible to prioritize responses to core light-weight and bulky goods customers, and also makes it difficult to meet the refined service demands of logistics companies for these customers, thus hindering the improvement of service quality and customer experience. Summary of the Invention

[0005] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a method, apparatus, device and storage medium for identifying and sorting applications of customers who handle light and bulky goods in logistics, in order to solve the technical problems of frequent misreporting of applications by customers who handle light and bulky goods in logistics and low processing efficiency of applications from important customers who handle light and bulky goods.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: The first aspect of this invention provides a method for identifying and sorting logistics light-shipment customers, including the following steps: obtaining a submission application; using the customer code and branch code in the submission application as a combination, matching and verifying it with all historical submission records to determine whether it is a duplicate submission, and outputting a first judgment result; if the first judgment result is a non-duplicate submission, obtaining customer information; determining whether the customer is a light-shipment customer based on the customer information, and outputting a second judgment result; if the second judgment result is a light-shipment customer, extracting the customer's historical shipping data, industry seasonal characteristics, and logistics peak and off-peak season coefficients to predict the customer's shipping volume in the future period as a reference volume; scoring the priority of the submission application based on the reference volume and the customer level, product attributes, and branch priority in the submission application; and sorting the review tasks of all bills of lading applications based on the scoring results.

[0007] Optionally, in a first implementation of the first aspect of the present invention, the step of obtaining the submission application, using the customer code and branch code in the submission application as a combination, and matching and verifying it with all historical submission records to determine whether it is a duplicate submission, and outputting a first judgment result, specifically includes: obtaining the submission application, using the customer code and branch code in the submission application as a combination, and matching it with the customer code and branch code combination in the historical submission records to obtain a first matching result; if the first matching result is non-duplicate, extracting all associated store numbers based on the customer code, combining each store number with the customer code and branch code as a matching target, and matching it again with the historical submission records to obtain a second matching result; if the second matching result is non-duplicate, using a semantic matching algorithm, matching the customer name and store name in the submission application with the historical submission records to obtain the first judgment result.

[0008] Optionally, in a second implementation of the first aspect of the present invention, the step of obtaining customer information, determining whether a customer is a "lightweight" customer based on the customer information, and outputting a second judgment result if the first judgment result is a non-duplicate declaration, specifically includes: obtaining customer information, extracting customer features from the customer information, wherein the customer features include basic information, product category tags, historical transaction records, and store attributes; classifying each customer feature to obtain multiple conversion features, dynamically setting weights for each conversion feature; performing weighted summation calculation on each conversion feature to obtain a comprehensive judgment value, comparing the comprehensive judgment value with a preset threshold, and outputting a second judgment result.

[0009] Optionally, in the third implementation of the first aspect of the present invention, the step of classifying and processing each customer characteristic to obtain multiple conversion features and dynamically setting weights for each conversion feature specifically includes: matching product category tags with a preset core category list and calculating the tag matching degree; calculating the light-weight attribute correlation degree based on the transaction proportion of light-weight products in historical transaction records; statistically analyzing the historical cooperation frequency with the corresponding category outlets in the basic information as a cooperation correlation feature; dynamically setting weights for the tag matching degree, light-weight attribute correlation degree, and historical cooperation frequency respectively, and adding additional basic weights for the features of stores whose store attributes are specific category specialty stores.

[0010] Optionally, in the fourth implementation of the first aspect of the present invention, if the second judgment result is a customer with light-weight goods, extracting the customer's historical shipping data, industry seasonal characteristics, and logistics peak-off-peak coefficients to predict the customer's shipping volume in the future period as a reference quantity specifically includes: extracting the customer's historical shipping data, preprocessing the historical shipping data to form a continuous historical shipping sequence; extracting industry seasonal characteristics and logistics peak-off-peak coefficients, converting industry seasonal characteristics into monthly coefficients according to the seasonal sales patterns of the customer's light-weight goods, and setting weekly coefficients for logistics peak-off-peak coefficients according to the fluctuations in the overall regional logistics business volume; using the historical shipping sequence as basic trend data, and the monthly coefficients of industry seasonal characteristics and logistics peak-off-peak coefficients as adjustment factors, inputting them into the prediction model to obtain the predicted result of the average daily shipping volume in the future period, wherein the predicted result of the average daily shipping volume is the reference quantity.

[0011] Optionally, in the fifth implementation of the first aspect of the present invention, the step of extracting industry seasonal characteristics and logistics peak-off-peak coefficients, converting industry seasonal characteristics into monthly coefficients based on the seasonal sales patterns of customers' lightweight and disposable goods, and setting weekly coefficients for logistics peak-off-peak coefficients based on fluctuations in the overall regional logistics volume, specifically includes: extracting industry seasonal characteristics, calculating monthly coefficients for each sub-category of lightweight and disposable goods, obtaining climate data, and adjusting the monthly coefficients using the climate data; extracting logistics peak-off-peak coefficients, and setting weekly coefficients for logistics peak-off-peak coefficients based on regional consumption habits, natural weekly fluctuations, and event-driven factors; obtaining customer inventory data, and adjusting the weekly coefficients for logistics peak-off-peak coefficients based on the customer inventory data.

[0012] Optionally, in the sixth implementation of the first aspect of the present invention, the step of scoring the priority of the application based on the number of reference documents and the customer level, product attributes, and branch priority in the application, and sorting the review tasks of all bills of lading applications according to the scoring results, specifically includes: quantifying the number of reference documents, customer level, product attributes, and branch priority according to a preset grading standard to obtain multiple scores for different indicators; comprehensively calculating the priority score of the application based on the preset indicator weights and corresponding scores; and sorting the review tasks of all applications from high to low priority scores, with review tasks with the same score sorted according to the order of submission time to form the final review task sequence.

[0013] The second aspect of this invention provides a device for identifying and sorting logistics light-shipment customers, comprising: a first judgment module, used to acquire a submission application, and using the customer code and branch code in the submission application as a combination to match and verify with all historical submission records to determine whether it is a duplicate submission, and outputting a first judgment result; a second judgment module, used to acquire customer information if the first judgment result is a non-duplicate submission, and determine whether the customer is a light-shipment customer based on the customer information, and outputting a second judgment result; a prediction module, used to extract the customer's historical shipping data, industry seasonal characteristics, and logistics peak and off-peak season coefficients if the second judgment result is a light-shipment customer, to predict the customer's shipping volume in the future period as a reference volume; and a sorting module, used to score the priority of the submission application based on the reference volume and the customer level, product attributes, and branch priority in the submission application, and sort the review tasks of all bills of lading applications based on the scoring results.

[0014] Optionally, in a first implementation of the second aspect of the present invention, the first judgment module includes: a first matching unit, configured to obtain a submission application, and match the customer code and outlet code in the submission application with the customer code and outlet code combination in historical submission records to obtain a first matching result; a second matching unit, configured to, if the first matching result is non-repeating, extract all associated store numbers based on the customer code, combine each store number with the customer code and outlet code to form a combination, and use it as a matching target, and match it again with historical submission records to obtain a second matching result; and a third matching unit, configured to, if the second matching result is non-repeating, use a semantic matching algorithm to match the customer name and store name in the submission application with historical submission records to obtain a first judgment result.

[0015] Optionally, in a second implementation of the second aspect of the present invention, the second judgment module includes: an extraction submodule, used to obtain customer information if the first judgment result is a non-duplicate declaration, and extract customer features from the customer information, wherein the customer features include basic information, product category tags, historical transaction records and store attributes; a processing submodule, used to classify and process each customer feature to obtain multiple conversion features, and dynamically set weights for each conversion feature; and a comparison submodule, used to perform weighted summation calculation on each conversion feature to obtain a comprehensive judgment value, compare the comprehensive judgment value with a preset threshold, and output a second judgment result.

[0016] Optionally, in a third implementation of the second aspect of the present invention, the processing submodule includes: a first calculation unit, used to match product category tags with a preset core category list and calculate the tag matching degree; a second calculation unit, used to calculate the light-weight attribute correlation degree based on the transaction proportion of light-weight products in historical transaction records; a statistics unit, used to count the historical cooperation frequency with the corresponding category outlet in the basic information as a cooperation correlation feature; and a weighting unit, used to dynamically set weights for the tag matching degree, the light-weight attribute correlation degree, and the historical cooperation frequency, and to add additional basic weights for the features of stores whose store attributes are specific category specialty stores.

[0017] Optionally, in the fourth implementation of the second aspect of the present invention, the prediction module includes: a preprocessing submodule, used to extract historical shipment data from customers, preprocess the historical shipment data to form a continuous historical shipment sequence; a conversion submodule, used to extract industry seasonal characteristics and logistics peak and off-peak season coefficients, convert industry seasonal characteristics into monthly coefficients according to the seasonal sales patterns of customers' light and bulky goods, and set weekly coefficients for logistics peak and off-peak season coefficients according to the fluctuations in the overall regional logistics business volume; and a prediction submodule, used to input the historical shipment sequence as basic trend data, the monthly coefficients of industry seasonal characteristics and logistics peak and off-peak season coefficients as adjustment factors into the prediction model to obtain the daily average shipment volume prediction result for a future period, wherein the daily average shipment volume prediction result is a reference shipment volume.

[0018] Optionally, in a fifth implementation of the second aspect of the present invention, the conversion submodule includes: a seasonal feature unit, used to extract industry seasonal features, calculate monthly coefficients according to the subcategories of lightweight and portable goods, obtain climate data, and adjust the monthly coefficients using the climate data; a first adjustment unit, used to extract logistics peak and off-peak season coefficients, and set the weekly coefficients of logistics peak and off-peak season coefficients in combination with regional consumption habits, natural weekly fluctuations, and event-driven factors; and a second adjustment unit, used to obtain customer inventory data and adjust the weekly coefficients of logistics peak and off-peak season coefficients according to the customer inventory data.

[0019] Optionally, in the sixth implementation of the second aspect of the present invention, the sorting module includes: a quantification unit, used to quantify the reference quantity, customer level, product attributes and outlet priority according to a preset grading standard, and obtain multiple scores for different indicators; a scoring calculation unit, used to comprehensively calculate the priority score of the submitted application according to the preset indicator weights and corresponding scores; and a sorting unit, used to sort all the review tasks of the submitted applications from high to low according to the priority scores, and sort the review tasks with the same score according to the order of submission time, to form the final review task sequence.

[0020] A third aspect of the present invention provides a logistics lightweight customer identification and application review and sorting device, including a memory and at least one processor, wherein the memory stores computer-readable instructions; the at least one processor invokes the computer-readable instructions in the memory to execute the various steps of the logistics lightweight customer identification and application review and sorting method described above.

[0021] A fourth aspect of the present invention provides a computer-readable storage medium storing computer-readable instructions, wherein, when executed by a processor, the computer-readable instructions implement the various steps of the logistics lightweight customer identification and application review sorting method described above.

[0022] Beneficial Effects: This invention provides a method for identifying and sorting logistics light-shipment customers' applications. First, it obtains the submitted application, combining the customer code and branch code from the application with all historical submission records to determine if it is a duplicate application. A first judgment result is output, eliminating duplicate applications and reducing unnecessary repetitive work. If the first judgment result indicates no duplicate application, customer information is further obtained. Based on this information, it is determined whether the customer is a light-shipment customer, and a second judgment result is output. Precise screening avoids non-light-shipment customers consuming resources, thereby improving processing efficiency. If the second judgment result indicates a light-shipment customer, the customer's historical shipping data, industry seasonal characteristics, and logistics peak / off-peak season coefficients are used to predict the customer's future shipping volume as a reference quantity, a crucial basis for determining priority for review. Finally, based on the reference quantity and key factors such as customer level, product attributes, and branch priority in the application, the priority of the submitted applications is scored. The review tasks of all bills of lading applications are then sorted according to the scoring results, thereby identifying core light-shipment customers and prioritizing their applications for refined service. Attached Figure Description

[0023] Figure 1 This is a first flowchart of a logistics lightweight customer identification and application review sorting method provided in an embodiment of the present invention; Figure 2 This is a second flowchart of the logistics lightweight customer identification and application review sorting method provided in an embodiment of the present invention; Figure 3 This is a third flowchart of the logistics lightweight customer identification and application review sorting method provided in an embodiment of the present invention; Figure 4 This is a fourth flowchart of the logistics lightweight customer identification and application review sorting method provided in an embodiment of the present invention; Figure 5 This is the fifth flowchart of the logistics lightweight customer identification and application review sorting method provided in the embodiments of the present invention; Figure 6 The sixth flowchart of the logistics lightweight customer identification and application review sorting method provided in the embodiments of the present invention; Figure 7 The seventh flowchart of the logistics lightweight customer identification and application review sorting method provided in the embodiments of the present invention; Figure 8 A schematic diagram of a logistics lightweight customer identification and application review sorting device provided in an embodiment of the present invention; Figure 9 Another structural schematic diagram of the logistics lightweight customer identification and application review sorting device provided in an embodiment of the present invention; Figure 10 This is a schematic diagram of the structure of a logistics lightweight customer identification and application review sorting device provided in an embodiment of the present invention. Detailed Implementation

[0024] This invention provides a method, apparatus, device, and storage medium for identifying and sorting logistics customers prone to lightweight applications. First, after obtaining the submitted application, the invention uses a unique combination of "customer code and branch code" and performs a full-domain matching verification with historical application records to generate a duplicate application determination result. This directly excludes duplicate applications within the system, avoiding invalid review work and improving overall processing efficiency. If the application is determined to be non-duplicate, further customer information is retrieved for lightweight application attribute verification. By effectively screening applications from non-lightweight customers, irrelevant resource consumption is avoided, ensuring that review resources are tilted towards target customers. For applications determined to be from lightweight customers, a predictive model is used to accurately calculate the customer's shipment volume in the future period based on the customer's historical shipping data, seasonal characteristics of the footwear and apparel industry, and the peak and off-peak season coefficients of logistics, forming a core reference for priority determination. Finally, combining the predicted shipment volume with key indicators such as customer level, product attributes, and branch priority in the submitted application, a comprehensive scoring system is established. All applications are reviewed and sorted according to the scoring results, accurately identifying core lightweight customers and achieving priority review and refined service adaptation for their applications.

[0025] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 The first embodiment of the logistics lightweight customer identification and application review sorting method in this invention includes: S101. Obtain the submission application, use the customer code and branch code in the submission application as a combination, match and verify with all historical submission records, determine whether it is a duplicate submission, and output the first judgment result; S102. If the first judgment result is that the declaration is not repeated, obtain the customer information, determine whether the customer is a customer who throws things away carelessly based on the customer information, and output the second judgment result. S103. If the second judgment result is a customer who throws goods lightly, extract the customer's historical shipping data, industry seasonal characteristics, and logistics peak and off-peak season coefficients to predict the customer's shipping volume in the future period as a reference quantity. S104. Based on the reference volume and the customer level, product attributes, and branch priority in the application, score the priority of the application, and sort the review tasks of all bill of lading applications according to the score results.

[0026] In existing technologies, due to the lack of standardized management of application submissions at some branches, duplicate applications frequently occur, leading to wasted review resources, duplicate subsequent planning, and management chaos. In this embodiment, by performing pre-emptive matching and verification of application submissions to determine if they are duplicates, the adverse effects of duplicate applications can be effectively avoided. Specifically, within a specific period, a branch typically only submits one application for the same customer. Therefore, by matching the customer code and branch code with all historical submission records within the period, it can be determined whether a duplicate application is valid. For applications deemed duplicates, the system will issue a corresponding prompt and provide relevant information about the duplicate application.

[0027] In addition to handling duplicate submissions, the system also deals with applications from customers who are not eligible for price adjustments. When initiating applications, operators often employ a trial-and-error strategy due to uncertainty about whether a customer qualifies for price adjustments. This results in a large number of ineligible applications mixed in with the system's list, severely consuming review resources. By comprehensively determining whether a customer qualifies for price adjustments based on their information, the system addresses the applicant's inability to accurately identify eligible customers and eliminates the need for significant manpower on the review side to filter out ineligible applications. The system also provides corresponding prompts for ineligible applications.

[0028] After passing the two initial checks mentioned above, the system extracts the customer's historical shipping data, industry seasonal characteristics, and logistics peak / off-peak season coefficients. Based on these key indicators, it predicts the customer's shipping volume for a future period as a reference quantity. The higher the customer's future shipping volume, the greater their importance and the higher their review priority. To more accurately predict the customer's shipping volume, this embodiment uses historical shipping data as the base data, combined with current industry seasonal characteristics and logistics peak / off-peak season coefficients as adjustment factors. The prediction uses a time-series forecasting algorithm to simulate the customer's shipping volume for a future period.

[0029] To better assess customer importance, when prioritizing submitted applications, the system incorporates customer level, product attributes, and branch priority into the scoring criteria, in addition to the aforementioned key reference volume. Specifically, customer level reflects the customer's long-term importance, while reference volume only reflects importance within the current period. Product attributes are closely related to the characteristics of lightweight parcels, and branch priority is related to the logistics company's branch support strategies. By comprehensively considering multiple factors, the final priority score more objectively reflects the importance of the bill of lading application. Ranking all bill of lading applications based on this score ensures a refined service approach for lightweight parcel customers.

[0030] Please see Figure 2 The second embodiment of the logistics lightweight customer identification and application review sorting method in this invention includes: S201. Obtain the submission application, and match the customer code and branch code combination in the submission application with the customer code and branch code combination in the historical submission record to obtain the first matching result; S202. If the first matching result is non-repeating, extract all associated store numbers based on the customer code, combine each store number with the customer code and the outlet code to form a combination, and use it as a matching target. Then match it again with the historical reporting records to obtain the second matching result. S203. If the second matching result is non-repeating, the semantic matching algorithm is used to match the customer name and store name in the application with the historical application records to obtain the first judgment result.

[0031] In this embodiment, the following method is used to more accurately filter out duplicate applications: The first step involves extracting the customer code and corresponding branch code from the submitted application. This is then precisely matched against the "customer code - branch code" combination in all historical submission records stored in the system. If a completely identical combination exists, the submission is automatically flagged as a duplicate. For example, if a footwear and apparel brand customer (code C001) submits a customer submission through a branch (code W002), and three days later, a branch employee submits another application with the same components, the first step directly compares the "C001 + W002" combination. Finding a perfect match with historical records, the application is immediately flagged as a duplicate and rejected. This eliminates the need for further verification, preventing auditors from repeatedly checking the same information and significantly reducing processing time.

[0032] If the first step does not identify a duplicate, a store code association check is performed. The store code in the submission request is extracted, and all recorded store codes under that customer code are associated with it. It is then checked whether the current store code has already been associated with the customer code and outlet code in a submission record. If a consistent "customer code-outlet code-store code" combination exists, it is considered a duplicate submission. For example, if customer C001 has two stores (store codes S003 and S004), the combinations "C001+W005+S003" and "C001+W005+S004" are used for verification. If a submission is found to be completed in either combination, it is considered a duplicate submission from the same customer and the same outlet. This combined matching check effectively improves computational efficiency, eliminating the need to traverse both outlet and store codes.

[0033] If the first two steps do not identify a duplicate, a semantic matching check is performed. The customer name and store name in the submitted application are converted into standardized text and compared with the customer name and store name in the historical records. By calculating word overlap and semantic relevance, if the similarity reaches a preset threshold, it is determined to be a semantically duplicate submission. This avoids missed duplicate submissions due to name changes or differences in expression, ensuring thorough screening for duplicates. For example, if a duplicate application is not detected in the first two steps due to an error in the previous or subsequent encoding input, the subsequent semantic matching can determine the similarity between the two submissions based on the customer name, store name, and the specific content of the submitted application.

[0034] Please see Figure 3 The third embodiment of the logistics lightweight customer identification and application review sorting method in this invention includes: S301. If the first determination result is non-duplicate declaration, obtain customer information and extract customer characteristics from the customer information. The customer characteristics include basic information, product category tags, historical transaction records and store attributes. S302. Classify each customer characteristic to obtain multiple conversion characteristics, and dynamically set weights for each conversion characteristic; S303. Perform weighted summation calculation on each transformation feature to obtain a comprehensive judgment value, compare the comprehensive judgment value with a preset threshold, and output the second judgment result.

[0035] In this embodiment, to better determine customer type, customer data is collected, including basic information corresponding to the customer code, product category tags, historical transaction records for the past 6 months, and store attributes. The collected features are then categorized to obtain multiple transformation features to better quantify each feature. Dynamic weights are then assigned to each feature, and the weighted features are aggregated into a comprehensive judgment value. The dynamic weight design can respond to factors such as industry trends and changes in customer operations.

[0036] For example, when identifying customers who frequently throw away footwear and apparel, if the overall judgment value reaches a preset threshold, the customer is determined to be a customer who frequently throws away footwear and apparel; otherwise, they are determined to be a customer who does not frequently throw away footwear and apparel. By comprehensively considering multiple characteristics to determine whether someone is a customer who frequently throws away footwear and apparel, misjudgment based on a single characteristic can be avoided, making the judgment results more accurate and reducing false positives.

[0037] Please see Figure 4 The fourth embodiment of the logistics lightweight customer identification and application review sorting method in this invention includes: S401. Match the product category tags with the preset core category list and calculate the tag matching degree; S402. Calculate the correlation degree of the light-weight / light-weight attribute based on the transaction proportion of light-weight / light-weight goods in historical transaction records; S403. The historical frequency of cooperation with corresponding product category outlets in the basic statistical information is used as a cooperation association feature; S404. Dynamically set weights for tag matching degree, light-throw attribute correlation degree, and historical cooperation frequency respectively. For stores whose attributes are specific category specialty stores, additional basic weights for features are added.

[0038] In order to better determine whether a customer is a "light-disposal" customer by using characteristics such as basic customer information, product category tags, historical transaction records, and store attributes, this embodiment transforms and quantifies some of the above characteristics.

[0039] For example, when identifying customers who are considered "lightly disposable" in the footwear and apparel category, a core category list for such customers is pre-defined: dresses, knitwear, T-shirts, canvas shoes, socks, and light jackets (6 categories in total). The customer's product category tags are "T-shirts, jeans, suitcases, and backpacks." Among these, "T-shirts" and "canvas shoes" match the core list. The matching degree is calculated as: number of matching categories / total number of items in the core list. By quantifying the fit between product tags and "lightly disposable" categories, the crude logic of "any product with a footwear or apparel tag is considered lightly disposable" is avoided. Historical transaction records are processed as follows: if a customer has 120 orders in the past 6 months, and 90 of them match "lightly disposable" items, the "lightly disposable" attribute correlation is calculated as: number of lightly disposable orders / total number of transactions. The transaction percentage directly reflects the core nature of the customer's "lightly disposable" business; the higher the correlation, the stronger the credibility of identifying the customer as a "lightly disposable" customer. The historical frequency of cooperation with corresponding product category outlets in the basic statistical information reflects the depth of the bond between the customer and Qingpao's dedicated outlets. High-frequency cooperation indicates a high degree of compatibility between the customer's business model and Qingpao's services. In terms of weighting, dynamic weighting adapts to the industry's peak and off-peak season characteristics, allowing the judgment logic to align with business fluctuations, making it more flexible and accurate than fixed weighting. Furthermore, since franchised stores are a high-probability group of Qingpao customers, additional weighting can quickly identify core target customers and reduce misjudgments during the review process.

[0040] Please see Figure 5 The fifth embodiment of the logistics lightweight customer identification and application review sorting method in this invention includes: S501. Extract the customer's historical shipment data, preprocess the historical shipment data, and form a continuous historical shipment sequence; S502. Extract industry seasonal characteristics and logistics peak and off-peak season coefficients, convert industry seasonal characteristics into monthly coefficients based on the seasonal sales patterns of customers' lightweight and portable goods, and set weekly coefficients for logistics peak and off-peak season coefficients based on the fluctuations in the overall regional logistics business volume. S503. Using historical shipping sequences as basic trend data, and monthly coefficients of industry seasonal characteristics and logistics peak and off-peak season coefficients as adjustment factors, the data are input into the prediction model to obtain the daily average shipping volume prediction result for a future period of time. The daily average shipping volume prediction result is used as the reference quantity.

[0041] The number of reference documents is a key factor in determining the priority of a proposal application, and it is directly related to the workload. Therefore, the more accurate the estimation of the number of reference documents, the more precise the company's review service resources will be.

[0042] Specifically, when obtaining customers' historical shipping data, it is necessary to preprocess the historical shipping data first, such as removing outliers and filling in the missing dates with the average of the three days before and after, to form a continuous historical shipping sequence.

[0043] Then, the seasonal characteristics of the industry and the coefficients for peak and off-peak seasons in logistics are extracted. The seasonal characteristics of the industry are converted into monthly coefficients based on the seasonal sales patterns of footwear and apparel. For example, based on the sales patterns of clothing, summer (May-July) is the peak season for dresses and T-shirts, with a monthly coefficient of 1.3; winter (November-January) is the peak season for down jackets and sweaters, with a coefficient of 1.2; spring (February-April) and autumn (August-October) are off-peak seasons, with a coefficient of 1.0.

[0044] The logistics peak / off-peak season coefficient is set as a weekly coefficient based on the overall fluctuation of regional logistics business volume. The pre-processed historical shipment sequences are then integrated with the seasonal and peak / off-peak season coefficients to form the input data for the prediction model. The historical shipment sequences serve as the basic trend data, while the seasonal and peak / off-peak season coefficients act as adjustment factors. The model analyzes the changing trends of the input data using time-series forecasting methods, first identifying periodic fluctuations in historical data, then adjusting the trend direction by combining the seasonal and peak / off-peak season coefficients, ultimately generating a prediction of the average daily shipment volume for the next 7-30 days. This dual-coefficient calibration allows the model to better reflect real-world scenarios, ensuring that the prediction results do not deviate from industry patterns and logistics realities.

[0045] Please see Figure 6 The sixth embodiment of the logistics lightweight customer identification and application review sorting method in this invention includes: S601. Extract industry seasonal characteristics, calculate monthly coefficients for each sub-category of lightweight and portable goods, obtain climate data, and adjust the monthly coefficients using the climate data; S602. Extract the peak and off-peak season coefficient of logistics, and set the weekly coefficient of the peak and off-peak season coefficient of logistics by combining regional consumption habits, natural weekly fluctuations and event-driven factors; S603. Obtain customer inventory data and adjust the weekly coefficient of the logistics peak and off-peak season coefficient based on the customer inventory data.

[0046] In this embodiment, to further improve the accuracy of predictions, climate data is introduced to account for industry seasonal characteristics. This climate data includes regional temperature forecasts for a future period, such as weather forecasts for the next half to one month. Industry seasonal characteristics are derived from historical data; however, with significant climate change, the start dates of seasons are becoming increasingly unstable, sometimes even delayed or advanced by half a month compared to the average. Consumers often make purchases in advance based on weather forecasts. For example, if the weather forecast predicts a mild winter, it will reduce the sales volume of down jackets, causing the peak season for winter clothing sales to be delayed, thus affecting the monthly coefficient. Therefore, combining regional climate data can yield a more accurate monthly coefficient.

[0047] The seasonality coefficient for logistics is mainly influenced by regional consumption habits, natural weekly fluctuations, and event-driven factors. Regional consumption habits refer to differences in local consumer groups; for example, people in the north tend to buy more winter clothing before winter arrives, while those in the south do so less. Natural weekly fluctuations refer to the difference in shipment volume between weekdays and weekends. Event-driven factors refer to fluctuations in shipment volume caused by specific holidays; for example, shipment volume surges during shopping festivals and then drops sharply afterward. Since shopping festivals and traditional holidays are predictable, the weekly coefficients around specific holidays can be adjusted accordingly.

[0048] Furthermore, this embodiment adjusts the weekly coefficient of the logistics peak / off-peak season coefficient based on customer inventory data to avoid excessive deviation of the projected reference shipment volume from reality. All the above coefficients are based on industry and external data. If all coefficients are high, the combined theoretical reference shipment volume will be very high. However, for merchants, inventory levels can affect their ability to handle excessively high sales volumes. Combining customer inventory data effectively avoids overly high predictions that do not reflect the merchant's actual capacity. For example, the model predicts an average daily shipment volume of 500 units, but based on the merchant's inventory and production capacity, the maximum possible shipment volume is only 200 units. Adjusting the weekly coefficient based on inventory data lowers the model's predicted average daily shipment volume, making it more consistent with the merchant's actual capabilities.

[0049] Please see Figure 7The seventh embodiment of the logistics lightweight customer identification and application review sorting method in this invention includes: S701. Based on the preset grading standards, the reference quantity, customer level, product attributes and outlet priority are quantified to obtain multiple scores for different indicators; S702. Based on the preset indicator weights and corresponding scores, calculate the priority score of the submitted application. S703. Sort all submitted review tasks according to priority scores from high to low. Review tasks with the same score are sorted according to the order of submission time to form the final review task sequence.

[0050] In this embodiment, the review tasks are dynamically sorted based on the number of reference documents submitted, customer level, product attributes, and branch priority. Before sorting, the above indicators need to be quantified.

[0051] First, it is necessary to determine the quantitative standards for the ranking indicators. For example, for reference volume, it can be divided into 5 levels based on the historical average volume; for customer level, it can be divided into 3 levels based on historical cooperation amount, fulfillment rate, etc.; for product attributes, it can be divided into 3 levels based on the seasonal urgency of demand; and for branch priority, it can be divided into 4 levels based on branch compliance score and service quality level. Each level of the indicator corresponds to a different score.

[0052] Then, a fixed weight is assigned to each quantified indicator. Preferably, the reference quantity has the highest weight, followed by customer level, then product attribute urgency, and finally branch priority. Each indicator level of each review task is converted into a corresponding score, and a comprehensive score is calculated based on the weights. The formula is: Comprehensive Score = Reference Quantity Score × Weight + Customer Level Score × Weight + Product Attribute Score × Weight + Branch Priority Score × Weight. All review tasks are sorted from highest to lowest comprehensive score, and tasks with the same score are sorted by submission time to form the final review task sequence.

[0053] The above describes the method for identifying and sorting customers involved in lightweight and easily damaged goods in logistics according to embodiments of the present invention. The following describes the device for identifying and sorting customers involved in lightweight and easily damaged goods in logistics according to embodiments of the present invention. Please refer to [link / reference]. Figure 8 One embodiment of the logistics lightweight customer identification and application review sorting device of the present invention includes: The first judgment module 10 is used to obtain the submission application, and use the customer code and branch code in the submission application as a combination to match and verify with all historical submission records to determine whether it is a duplicate submission, and output the first judgment result. The second judgment module 20 is used to obtain customer information if the first judgment result is non-duplicate declaration, determine whether the customer is a customer who throws things lightly based on the customer information, and output the second judgment result. The prediction module 30 is used to extract the customer's historical shipment data, industry seasonal characteristics, and logistics peak and off-peak season coefficients if the second judgment result is a customer who throws lightly, in order to predict the customer's shipment volume in the future period as a reference quantity. The sorting module 40 is used to score the priority of the submitted applications based on the reference quantity and the customer level, product attributes and branch priority in the submitted applications, and sort the review tasks of all bill of lading applications according to the scoring results.

[0054] Please see Figure 9 One embodiment of the logistics lightweight customer identification and application review sorting device of the present invention includes: The first judgment module 10 is used to obtain the submission application, and use the customer code and branch code in the submission application as a combination to match and verify with all historical submission records to determine whether it is a duplicate submission, and output the first judgment result. The second judgment module 20 is used to obtain customer information if the first judgment result is non-duplicate declaration, determine whether the customer is a customer who throws things lightly based on the customer information, and output the second judgment result. The prediction module 30 is used to extract the customer's historical shipment data, industry seasonal characteristics, and logistics peak and off-peak season coefficients if the second judgment result is a customer who throws lightly, in order to predict the customer's shipment volume in the future period as a reference quantity. The sorting module 40 is used to score the priority of the submitted applications based on the number of reference documents and the customer level, product attributes and branch priority in the submitted applications, and sort the review tasks of all bill of lading applications according to the scoring results. In this embodiment, the first determination module 10 includes: The first matching unit 11 is used to obtain the submission application, and to match the customer code and branch code combination in the submission application with the customer code and branch code combination in the historical submission record to obtain the first matching result. The second matching unit 12 is used to extract all associated store numbers based on the customer code if the first matching result is non-repeating, combine each store number with the customer code and the outlet code to form a combination as the matching target, and match it again with the historical reporting record to obtain the second matching result. The third matching unit 13 is used to, if the second matching result is non-repetition, use a semantic matching algorithm to match the customer name and store name in the application with historical application records to obtain the first judgment result. In this embodiment, the second determination module 20 includes: Extraction submodule 21 is used to obtain customer information if the first determination result is non-duplicate declaration, and extract customer features from the customer information. The customer features include basic information, product category tags, historical transaction records and store attributes. The processing submodule 22 is used to classify and process each customer feature to obtain multiple conversion features, and dynamically set weights for each conversion feature; The comparison submodule 23 is used to perform weighted summation calculation on each transformation feature to obtain a comprehensive judgment value, compare the comprehensive judgment value with a preset threshold, and output a second judgment result. In this embodiment, the processing submodule 22 includes: The first calculation unit 221 is used to match the product category tags with the preset core category list and calculate the tag matching degree; The second calculation unit 222 is used to calculate the correlation degree of the light-weight attribute based on the transaction proportion of light-weight goods in historical transaction records; Statistical unit 223 is used to count the historical frequency of cooperation with corresponding product category outlets in the basic information, as a cooperation association feature; Weight unit 224 is used to dynamically set weights for tag matching degree, light throw attribute correlation degree and historical cooperation frequency respectively. For stores whose attributes are specific category specialty stores, additional basic weights of features are added. In this embodiment, the prediction module 30 includes: Preprocessing submodule 31 is used to extract customers' historical shipment data, preprocess the historical shipment data, and form a continuous historical shipment sequence. The conversion submodule 32 is used to extract industry seasonal characteristics and logistics peak and off-peak season coefficients. Based on the seasonal sales patterns of customers' lightweight and portable goods, the industry seasonal characteristics are converted into monthly coefficients. Based on the fluctuations in the overall regional logistics business volume, the weekly coefficients of the logistics peak and off-peak season coefficients are set. The prediction submodule 33 is used to input the historical shipment sequence as the basic trend data, the monthly coefficient of the industry seasonal characteristics and the logistics peak and off-peak season coefficient as adjustment factors into the prediction model to obtain the daily average shipment volume prediction result for a future period of time. The daily average shipment volume prediction result is the reference shipment volume. In this embodiment, the conversion submodule 32 includes: Seasonal feature unit 321 is used to extract industry seasonal features, calculate monthly coefficients according to the subcategories of lightweight and portable goods, obtain climate data, and adjust the monthly coefficients using the climate data; The first adjustment unit 322 is used to extract the peak and off-peak season coefficient of logistics, and to set the weekly coefficient of the peak and off-peak season coefficient of logistics in combination with regional consumption habits, natural weekly fluctuations and event-driven factors. The second adjustment unit 323 is used to obtain customer inventory data and adjust the weekly coefficient of the logistics peak and off-peak season coefficient based on the customer inventory data. In this embodiment, the sorting module 40 includes: The quantification unit 41 is used to quantify the reference quantity, customer level, product attributes and outlet priority according to the preset grading standards, and obtain multiple scores for different indicators. The scoring calculation unit 42 is used to calculate the priority score of the submitted application based on the preset indicator weights and corresponding scores. The sorting unit 43 is used to sort all submitted review tasks according to priority scores from high to low. Review tasks with the same score are sorted according to the order of submission time to form the final review task sequence.

[0055] The logistics lightweight customer identification and application review and sorting device of the present invention first obtains the customer's application, uses the customer code and branch code in the application as a combination key, and performs a matching and verification with all historical application records in the system to determine whether there is a duplicate application and outputs a first judgment result, thereby eliminating duplicate applications in the system and saving unnecessary duplicate processing steps; if the first judgment result is a non-duplicate application, the device further retrieves the customer information, determines whether the customer is a lightweight customer based on the information, and outputs a second judgment result, thereby accurately screening out non-lightweight customers, avoiding invalid resource occupation, and improving overall processing efficiency; if the second judgment result is... If a customer is identified as a low-volume customer, the system integrates the customer's historical shipping data, seasonal characteristics of the footwear and apparel industry, and logistics peak and off-peak season coefficients to predict the customer's shipping volume for a future period and use this as a reference volume. This reference volume will become the core basis for determining whether an application should be prioritized for review. Finally, the system combines the reference volume with key factors such as customer level, product attributes, and branch priority stated in the application to conduct a comprehensive priority score. Based on the score results, the review tasks of all applications are ranked to accurately identify core low-volume customers, and their applications are then prioritized for processing, providing targeted and refined services.

[0056] The above describes the logistics lightweight customer identification and application review and sorting device in the embodiments of the present invention from the perspective of modular functional entities. The following describes the logistics lightweight customer identification and application review and sorting device in the embodiments of the present invention from the perspective of hardware processing.

[0057] Figure 10This is a schematic diagram of a logistics lightweight customer identification and application review and sorting device 900 provided in an embodiment of the present invention. The device can vary significantly due to different configurations or performance characteristics. It may include one or more central processing units (CPUs) 910 (e.g., one or more processors) and a memory 920, and one or more storage media 930 (e.g., one or more mass storage devices) storing application programs 933 or data 932. The memory 920 and storage media 930 can be temporary or persistent storage. The program stored in the storage media 930 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the logistics lightweight customer identification and application review and sorting device 900. Furthermore, the processor 910 may be configured to communicate with the storage media 930 and execute the series of instruction operations in the storage media 930 on the logistics lightweight customer identification and application review and sorting device 900 to implement the steps of the logistics lightweight customer identification and application review and sorting methods provided in the above-described method embodiments.

[0058] The logistics lightweight customer identification and application review sorting device 900 may also include one or more power supplies 940, one or more wired or wireless network interfaces 950, one or more input / output interfaces 960, and / or one or more operating systems 931, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 10 The illustrated structure of the customer identification and application review sorting device for light-duty logistics is not intended to limit the scope of the device. It may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0059] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the logistics lightweight customer identification and application review sorting method.

[0060] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the device or apparatus described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0061] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0062] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0063] It is understood that those skilled in the art can make equivalent substitutions or modifications to the technical solution and inventive concept of the present invention, and all such substitutions or modifications should fall within the protection scope of the appended claims.

Claims

1. A method for identifying and sorting applications from customers handling lightweight and easily disposable goods in logistics, characterized in that, The method comprises the following steps: Obtaining a report application, matching and checking all historical report records with the combination of the customer code and the network point code in the report application, judging whether it is a repeated report, and outputting a first judgment result; If the first judgment result is a non-repeated report, obtaining customer information, judging whether the customer is a light-throwing customer according to the customer information, and outputting a second judgment result; If the second judgment result is a light-throwing customer, extracting the historical delivery data of the customer, the industry seasonal characteristics and the logistics slack season coefficient to predict the delivery quantity of the customer in a future period of time as a reference quantity; According to the reference quantity and the customer level, product attributes and network priority in the report application, the priority of the report application is scored, and the audit tasks of all bill applications are sorted according to the scoring result.

2. The method of claim 1, wherein, The method comprises the following steps: Obtaining a report application, matching and checking all historical report records with the combination of the customer code and the network point code in the report application, judging whether it is a repeated report, and outputting a first judgment result; If the first judgment result is a non-repeated report, obtaining customer information, judging whether the customer is a light-throwing customer according to the customer information, and outputting a second judgment result; If the second judgment result is a light-throwing customer, extracting the historical delivery data of the customer, the industry seasonal characteristics and the logistics slack season coefficient to predict the delivery quantity of the customer in a future period of time as a reference quantity; 3. The method of claim 1, wherein, According to the reference quantity and the customer level, product attributes and network priority in the report application, the priority of the report application is scored, and the audit tasks of all bill applications are sorted according to the scoring result. The method comprises the following steps: Obtaining a report application, matching and checking all historical report records with the combination of the customer code and the network point code in the report application, judging whether it is a repeated report, and outputting a first judgment result; If the first judgment result is a non-repeated report, obtaining customer information, judging whether the customer is a light-throwing customer according to the customer information, and outputting a second judgment result; 4. The method of claim 3, wherein, If the second judgment result is a light-throwing customer, extracting the historical delivery data of the customer, the industry seasonal characteristics and the logistics slack season coefficient to predict the delivery quantity of the customer in a future period of time as a reference quantity; According to the reference quantity and the customer level, product attributes and network priority in the report application, the priority of the report application is scored, and the audit tasks of all bill applications are sorted according to the scoring result. The method comprises the following steps: Obtaining a report application, matching and checking all historical report records with the combination of the customer code and the network point code in the report application, judging whether it is a repeated report, and outputting a first judgment result; If the first judgment result is a non-repeated report, obtaining customer information, judging whether the customer is a light-throwing customer according to the customer information, and outputting a second judgment result; 5. The method of claim 1, wherein, If the second judgment result is a light-throwing customer, extracting the historical delivery data of the customer, the industry seasonal characteristics and the logistics slack season coefficient to predict the delivery quantity of the customer in a future period of time as a reference quantity; Extract the historical delivery data of the customer, preprocess the historical delivery data, and form a continuous historical delivery sequence; Extract the industry seasonal characteristics and the logistics off-season coefficient, convert the industry seasonal characteristics into monthly coefficients according to the seasonal sales law of the customer's light-throwing goods, and set the weekly coefficient of the logistics off-season coefficient according to the overall business volume fluctuation of the regional logistics; The historical delivery sequence is used as the basis trend data, the monthly coefficient of the industry seasonal characteristics and the logistics off-season coefficient are used as the adjustment factor, and the prediction model is input to obtain the daily average delivery quantity prediction result of a future period of time, which is the reference quantity.

6. The method of claim 5, wherein, The extraction of the industry seasonal characteristics and the logistics off-season coefficient, the conversion of the industry seasonal characteristics into monthly coefficients according to the seasonal sales law of the customer's light-throwing goods, and the setting of the weekly coefficient of the logistics off-season coefficient according to the overall business volume fluctuation of the regional logistics, specifically includes: Extract the industry seasonal characteristics, calculate the monthly coefficient according to the light-throwing goods subcategory, obtain the climate data, and adjust the monthly coefficient using the climate data; Extract the logistics off-season coefficient, set the weekly coefficient of the logistics off-season coefficient in combination with regional consumption habits, natural fluctuations and event-driven; Obtain the customer inventory data, and adjust the weekly coefficient of the logistics off-season coefficient according to the customer inventory data.

7. The method of claim 1, wherein, The priority of the submission application is scored according to the reference quantity and the customer level, product attribute and network priority in the submission application, and the audit tasks of all the submission applications are sorted according to the scoring results, specifically including: According to the preset grading standard, the reference quantity, customer level, product attribute and network priority are quantified respectively to obtain multiple scores for different indicators; According to the preset index weight and the corresponding score, the priority score of the submission application is calculated; The audit tasks of all the submission applications are sorted in descending order of priority score, and the audit tasks with the same score are sorted according to the submission time to form the final audit task sequence.

8. A logistics light-throw customer identification and application review sequencing device, characterized by, It includes: The first judgment module is used to obtain the submission application, and the customer code and network code in the submission application are combined to match and verify all historical submission records to determine whether it is a repeated report, and the first judgment result is output; The second judgment module is used to obtain the customer information if the first judgment result is not a repeated report, and to determine whether the customer is a light-throwing customer according to the customer information, and the second judgment result is output; The prediction module is used to extract the historical delivery data of the customer, the industry seasonal characteristics and the logistics off-season coefficient if the second judgment result is a light-throwing customer, to predict the delivery quantity of the customer in a future period of time as the reference quantity; The sorting module is used to score the priority of the submission application according to the reference quantity and the customer level, product attribute and network priority in the submission application, and to sort the audit tasks of all the submission applications according to the scoring results.

9. A logistics light-touch customer identification and application review sequencing apparatus, characterized by, It includes a memory and at least one processor, and the memory stores computer readable instructions; The at least one processor invokes the computer readable instructions in the memory to perform the steps of the method for sorting logistics light-throw customer identification and application review as claimed in any one of claims 1-7.

10. A computer-readable storage medium having stored thereon computer-readable instructions, wherein, The computer readable instructions, when executed by the processor, implement the steps of the method for sorting logistics light-throw customer identification and application review as claimed in any one of claims 1-7.

Images