A multi-source e-commerce order adaptive processing method and system based on big data

Abstract

The application discloses a kind of multi-source e-commerce order adaptive processing method and system based on big data, it is related to order adaptive processing technical field, the present application method includes order multi-source data acquisition, order multi-source data analysis, order adaptive processing and early warning prompt, based on obtaining multi-source heterogeneous order data, and then index is constructed, and e-commerce multi-source order adaptive processing model is established, whether the mechanism switching of adaptive processing model is correctly triggered is analyzed, while processing standard multi-source order data, whether there is abnormal risk in multi-source order is analyzed, and abnormal order is real-time eliminated, again the processing collection corresponding to effective order is analyzed, the execution of order corresponding processing collection is monitored, the processing result of e-commerce platform corresponding order adaptation is analyzed, and the adjustment of order corresponding adaptive processing problem is executed, to realize the efficient adaptive processing of multi-source e-commerce order, to guarantee that order is handled in time and efficiently.

Description

Technical Field

[0001] This invention relates to the field of order adaptive processing technology, specifically to a method and system for adaptive processing of multi-source e-commerce orders based on big data. Background Technology

[0002] With the diversified development of the e-commerce industry, multi-platform deployment has become the mainstream operating model for merchants. Multi-source e-commerce orders exhibit core characteristics such as heterogeneity, high concurrency, dynamic fluctuations, and high demand for end-to-end collaboration. Therefore, a big data-based adaptive processing method and system for multi-source e-commerce orders is proposed to achieve flexible processing of orders from multiple platforms, improve the efficiency and accuracy of multi-source e-commerce order processing, and realize efficient operation and maintenance of e-commerce.

[0003] Existing technologies, such as the invention patent application with publication number CN113538081A, disclose an e-commerce order system and its resource adaptive scheduling processing method. This method includes: receiving an order creation request sent by a user; obtaining user information based on the order creation request, the user information including the user's service level information; obtaining the concurrent quantity of orders corresponding to the order type requested by the user; determining whether the corresponding order module should directly create an order for the user based on the user's service level information and the concurrent quantity; if not, adding the user to a user queue for queuing; obtaining the concurrent processing pressure coefficient of the current user queue, and obtaining the bandwidth load and physical resource usage of each order module; adjusting the allocation ratio of bandwidth and physical resources among the order modules based on the concurrent processing pressure coefficient and the bandwidth load and physical resource usage of each order module. This method enables adaptive dynamic resource allocation.

[0004] Existing technology, such as the invention patent application with publication number CN117709900A, discloses an order shipment processing method and system. This method includes: when the expiration time in the expiration time marker is less than or equal to a second preset time value, determining whether the order information contains a logistics tracking number; if the order information contains a logistics tracking number, calling the order shipment processing interface and moving the order information into a document printing queue; if the order information does not contain a logistics tracking number, determining whether the order information contains an adaptive marker; if the order information contains an adaptive marker, obtaining a preset target logistics tracking number based on the logistics marker, assigning the target logistics tracking number to the order information to form adaptive self-order information, calling the order shipment processing interface, and moving the adaptive order information into the document printing queue. This technical solution enables automated order processing, improves order processing efficiency, and avoids or reduces order cancellation rates caused by untimely processing.

[0005] Current technologies only address the order creation process of a single online store, focusing on matching and scheduling user service levels with the number of concurrent requests. They do not cover the access and adaptive processing of multi-source heterogeneous orders, and cannot cope with order processing scenarios involving multiple platforms and high concurrency. They are limited to single-point, localized resource scheduling optimization and do not solve the problem of adaptive processing of multi-source orders. Existing technologies only focus on basic data processing or single-stage optimization of multi-source orders, resulting in insufficient adaptation to multi-source heterogeneous data and a lack of simplification. At the same time, traditional methods for order processing and analysis generally suffer from poor data fusion adaptability, static scheduling rules, and insufficient real-time performance. Therefore, it is easy for errors, omissions, or inventory synchronization delays to occur during order processing, making it difficult to meet the needs of merchants for efficient operation and reducing processing quality. Summary of the Invention

[0006] To address the aforementioned technical shortcomings, the present invention aims to provide a method and system for adaptive processing of multi-source e-commerce orders based on big data.

[0007] To solve the above technical problems, the present invention adopts the following technical solution: The present invention provides a multi-source e-commerce order adaptive processing method based on big data, including: S1, order multi-source data acquisition: acquiring multi-source heterogeneous order data from e-commerce platforms, thereby constructing an index system for the corresponding multi-source heterogeneous order data of e-commerce platforms, and establishing an e-commerce multi-source order adaptive processing model.

[0008] S2. Order Multi-Source Data Analysis: Collect order traffic data corresponding to e-commerce platforms, analyze whether the mechanism switching of the e-commerce multi-source order adaptive processing model is correctly triggered, and when the mechanism is successfully switched, input the standard multi-source order data corresponding to the e-commerce platform into the adaptive processing model to analyze whether there are any abnormal risks in the multi-source orders corresponding to the e-commerce platform.

[0009] S3. Adaptive Order Processing: Adaptive processing is performed on the valid multi-source order data corresponding to the e-commerce platform through an adaptive model to obtain the processing set of valid orders corresponding to the e-commerce platform. The execution process of the processing set of valid orders corresponding to the e-commerce platform is monitored, monitoring data is obtained, and the adaptive processing results of orders corresponding to the e-commerce platform are analyzed. When there are processing omissions in the adaptive processing of orders corresponding to the e-commerce platform, the adjustment of the adaptive processing of orders corresponding to the e-commerce platform is analyzed.

[0010] S4. Warning prompt: A warning prompt will be issued when the mechanism corresponding to the e-commerce multi-source order adaptive processing model is not successfully switched or when there is a processing oversight in the order adaptive processing of the e-commerce platform.

[0011] In a second aspect, the present invention provides a multi-source e-commerce order adaptive processing system based on big data, comprising: an order multi-source data acquisition module, used to acquire multi-source heterogeneous order data from an e-commerce platform, thereby constructing an index system for the corresponding multi-source heterogeneous order data of the e-commerce platform, and establishing an e-commerce multi-source order adaptive processing model; The multi-source order data analysis module is used to collect order traffic data corresponding to e-commerce platforms, analyze whether the mechanism switching of the multi-source order adaptive processing model of e-commerce is correctly triggered, and analyze whether there are any abnormal risks in the multi-source orders corresponding to e-commerce platforms. The order adaptive processing module is used to adaptively process valid multi-source order data based on an adaptive model to obtain the processing set of valid orders corresponding to the e-commerce platform, monitor the execution process of the processing set of valid orders corresponding to the e-commerce platform, analyze the adaptive processing results of orders corresponding to the e-commerce platform, and analyze the adjustment of the adaptive processing of orders corresponding to the e-commerce platform when there are processing omissions in the adaptive processing of orders corresponding to the e-commerce platform. The early warning terminal is used to issue early warnings when the mechanism corresponding to the e-commerce multi-source order adaptive processing model fails to switch successfully or when there are processing omissions in the order adaptive processing of the e-commerce platform.

[0012] The beneficial effects of this invention are as follows: 1. This invention provides a method and system for adaptive processing of multi-source e-commerce orders based on big data. It acquires multi-source heterogeneous order data, constructs an index, establishes an adaptive processing model for multi-source e-commerce orders, analyzes whether the mechanism switching of the adaptive processing model is correctly triggered, processes standard multi-source order data, analyzes whether there are any abnormal risks in multi-source orders, removes abnormal orders in real time, analyzes the processing set corresponding to valid orders, monitors the execution process of the processing set corresponding to valid orders, analyzes the adaptive processing results of orders on the e-commerce platform, and adjusts the adaptive processing of orders accordingly. This achieves efficient adaptive processing of multi-source e-commerce orders, ensuring timely and efficient order processing, further improving order processing efficiency, and solving problems such as incorrect orders, missing orders, or delayed inventory synchronization in e-commerce orders. It also automates order processing, reducing manual workload and costs, and efficiently ensures traceability and control throughout the entire order lifecycle, improving order processing quality and efficiency.

[0013] 2. Obtain multi-source heterogeneous order data from e-commerce platforms, then construct an index system for the corresponding multi-source heterogeneous order data of e-commerce platforms, and establish an adaptive processing model for multi-source e-commerce orders, thereby realizing unified management of order data and providing comprehensive support for subsequent adaptive processing of multi-source orders.

[0014] 3. Collect order traffic data from e-commerce platforms, analyze whether the mechanism switching of the multi-source order adaptive processing model for e-commerce is correctly triggered, and when the mechanism switches successfully, analyze whether there are any abnormal risks in the multi-source orders corresponding to the e-commerce platform, further solve the problems of incorrect orders, missing orders, or delayed inventory synchronization in e-commerce orders, and intelligently improve the quality and efficiency of order processing.

[0015] 4. Based on the adaptive model, abnormal orders are removed in real time. The adaptive model is implemented to adaptively process valid multi-source orders, resulting in a processing set. The processing set is monitored and analyzed to determine the adaptive processing results of the corresponding orders on the e-commerce platform. When there are processing omissions in the adaptive processing of orders, the adjustment of the adaptive processing of the corresponding orders on the e-commerce platform is analyzed to achieve efficient adaptive processing of multi-source e-commerce orders. Timely adjustments are made to the processing issues to further ensure the accuracy of order processing. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the implementation steps of the method of the present invention.

[0018] Figure 2 This is a schematic diagram of the system structure connection of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1 As shown, a multi-source e-commerce order adaptive processing method based on big data includes: S1, order multi-source data acquisition: acquiring multi-source heterogeneous order data from e-commerce platforms, then constructing an index system for the corresponding multi-source heterogeneous order data of e-commerce platforms, and establishing an e-commerce multi-source order adaptive processing model.

[0021] In a specific embodiment, the acquisition process of multi-source heterogeneous order data from an e-commerce platform is as follows: the multi-source heterogeneous order data includes basic order information, payment information, logistics information, and user behavior information. Preprocessing is performed on the multi-source heterogeneous order data, including data format standardization conversion, redundant data deduplication, missing field filling, noise data filtering, and platform data time sequence alignment, to obtain standard multi-source order data.

[0022] It should be noted that basic order information includes order number, store code, order time, order status, product details, shipping address, order amount, order source, and order type; payment information includes payment transaction number, payment method, payment status, payment time, actual payment amount, refund information, and discount information; logistics information includes delivery node status, timestamps for each stage, delivery timeliness, and logistics anomaly information; user behavior data packets include IP address, historical transaction characteristics, abnormal behavior characteristics, and user risk indicators; data preprocessing is performed to eliminate processing biases caused by heterogeneous data, ensuring the correctness of adaptive order processing.

[0023] In a specific embodiment, the construction process of the index system for multi-source heterogeneous order data corresponding to the e-commerce platform is as follows: Based on the preprocessed standard multi-source order data, a multi-dimensional joint index is constructed according to the order identifier, e-commerce platform source, time series, product SKU, delivery address, payment status and fulfillment status. At the same time, the data is stored and indexed in layers according to hot data, real-time data and historical data to form a multi-dimensional layered index structure for association positioning and batch aggregation, thus obtaining the index system for multi-source heterogeneous order data corresponding to the e-commerce platform.

[0024] It should be noted that the preprocessed standard multi-source order data is divided into hot data, real-time data, and historical data according to access frequency, real-time nature, and business lifecycle. Among them, hot data is online order data that is currently pending processing, pending payment, pending shipment, pending fulfillment, and frequently accessed; real-time data is incremental order data that is pushed in real time by various e-commerce platforms, generated in the recent period, and requires immediate calculation and feature extraction; historical data is existing order data that has been completed, cancelled, archived, has low access frequency, and is used for model training, statistical analysis, and reconciliation backtracking. Based on the above-mentioned hierarchical results, partitioned indexes and hierarchical storage structures are constructed to achieve low-latency access to high-frequency data and efficient management of full data, providing efficient data access and query support for subsequent feature extraction, model scheduling, and adaptive order processing; SKU is the inventory holding unit.

[0025] In a specific embodiment, the establishment process of the e-commerce multi-source order adaptive processing model is as follows: The main features of order processing are extracted from the index system of multi-source heterogeneous order data corresponding to the e-commerce platform. These main features are then normalized, discretized, and subjected to high-dimensional feature reduction to form the feature vector corresponding to the e-commerce multi-source order adaptive processing model. Simultaneously, the standard multi-source order data is divided into training, validation, and test sets. A model network structure is constructed based on the integration of reinforcement learning algorithms and e-commerce order processing business rules. Order efficiency, abnormal order identification accuracy, and resource utilization are used as the core dimensions of the model's reward function. The training set is input to perform iterative training of the model. During training, the model's fitting effect is monitored in real time based on the validation set. After training is complete, the test set is input into the model for multi-dimensional validation. If the model's output indicators are consistent with the preset standard indicators, the construction of the e-commerce multi-source order adaptive processing model is complete, and the parameters corresponding to successful model validation are used as the initial running parameters. Otherwise, iterative training of the model continues.

[0026] It should be noted that the main features include concurrency features, average order value features, and fulfillment timeliness features at the order level; interface response features and traffic fluctuation features at the platform level; and behavioral profile features and transaction credit features at the user level. The dataset will be partitioned by professional staff, and the partitioning ratio will be preset based on requirements; the specific partitioning ratio is not limited here. The training set is used for model parameter learning, the validation set is used for iterative parameter tuning, and the test set is used for final effect verification. Normalization, discretization, high-dimensional feature dimensionality reduction, reinforcement learning, and early stopping are all existing technologies and will not be elaborated upon further here. During training, based on the validation... The system integrates real-time monitoring of model fitting performance and employs early stopping to avoid overfitting. Pre-defined standard indicators are set by professional staff, such as a standard abnormal order identification accuracy of 98%, a standard order processing delay of 500ms, and a standard resource utilization rate of 85%. If the model output indicators are greater than or equal to 98% and less than or equal to 500ms, or greater than or equal to 85%, the model construction indicators are deemed to meet the standards, and the corresponding model construction is completed. Conversely, if any indicator is inconsistent with the preset results, the model hyperparameters are iteratively tuned using methods such as grid search or Bayesian optimization until they meet the standards.

[0027] S2. Order Multi-Source Data Analysis: Collect order traffic data corresponding to e-commerce platforms, analyze whether the mechanism switching of the e-commerce multi-source order adaptive processing model is correctly triggered, and when the mechanism is successfully switched, input the standard multi-source order data corresponding to the e-commerce platform into the adaptive processing model to analyze whether there are any abnormal risks in the multi-source orders corresponding to the e-commerce platform.

[0028] In a specific embodiment, the analysis process for determining whether the switching mechanism of the e-commerce multi-source order adaptive processing model is correctly triggered is as follows: Order traffic data includes daily stable traffic, promotional peak traffic, and off-peak traffic. The order traffic data packets corresponding to the multi-source orders on the e-commerce platform are compared with the reference order traffic sets corresponding to each switching mechanism stored in the database. The comparison yields the matching switching mechanism corresponding to the order traffic data. When the multi-source orders on the e-commerce platform are daily stable traffic, the basic steady-state switching mechanism is triggered. When the multi-source orders on the e-commerce platform are promotional peak traffic, the high-concurrency expansion switching mechanism is triggered. When the multi-source orders on the e-commerce platform are off-peak traffic, the resource-saving switching mechanism is triggered.

[0029] When the adaptive processing model triggers a mechanism switch, the corresponding running data of the model after the trigger is obtained, including the switch response time and the order processing delay time after the switch. If either the switch response time or the order processing delay time after the switch is greater than the preset reference time value, it is determined that the mechanism corresponding to the e-commerce multi-source order adaptive processing model has not been successfully switched; otherwise, the switch is successful.

[0030] It should be noted that the system collects the order volume, payment request volume, and concurrent user number of the e-commerce platform in real time within a preset time period. It also samples and uses a sliding window to count the current traffic value and traffic fluctuation rate, and compares them with the historical reference traffic range stored in the database. If the traffic is within the historical reference traffic range, it is considered normal daily traffic. If it exceeds the historical reference traffic range, it is considered peak traffic during promotions. If it does not reach the historical reference traffic range, it is considered low-peak and low-peak traffic.

[0031] It should be noted that, based on the historical order traffic data and order traffic switching mechanisms of the e-commerce platform, a reference order traffic set corresponding to each switching mechanism is constructed. This set serves as a reference for the adaptive processing model when matching switching mechanisms to multiple-source orders on the e-commerce platform, ensuring the correct mechanism operation. Daily stable traffic refers to orders in a stable trend, hence a basic steady-state switching mechanism is adopted to ensure timely and effective processing of orders in a stable state. Promotional peak traffic refers to orders experiencing a cumulative increase or a phased surge, hence a high-concurrency scaling switching mechanism is used to ensure timely and orderly processing of orders during periods of high volume, avoiding order failures or delayed processing. Low-peak and low-valley traffic refers to orders that are scattered or few in number, hence a resource-saving switching mechanism is used. To ensure timely order processing while minimizing unnecessary resource consumption, the basic steady-state switching mechanism maintains the core adaptive processing main model without loading additional sub-models. It dynamically fine-tunes model feature weights, such as order priority and resource allocation coefficients, to maintain low resource consumption and stability, ensuring efficient processing of regular orders. The high-concurrency scaling switching mechanism automatically unloads non-core computing tasks, optimizes order batch processing logic, and improves rules for quickly identifying abnormal orders. The resource-saving switching mechanism reduces the number of model computing nodes, merges batch processing tasks for low-frequency orders, and lowers the resource occupancy rate of the model. Preset reference duration values ​​are set by professional staff based on the anticipated needs of the model during switching; specific limits are not specified here to allow for a more intuitive assessment of whether the model is switching using the correct mechanism.

[0032] In a specific embodiment, the analysis process for determining whether there are abnormal risks in multi-source orders corresponding to an e-commerce platform is as follows: Standard multi-source order data corresponding to the e-commerce platform is input into an adaptive processing model. Based on the presentation of the standard multi-source order data in the adaptive processing model after input, order performance features are extracted and compared with the fraudulent order behavior feature library corresponding to the adaptive model. If the order performance features of a certain order in the standard multi-source order data of the e-commerce platform are included in the fraudulent order behavior feature library corresponding to the adaptive model, then it is determined that there are abnormal risks in the multi-source orders corresponding to the e-commerce platform. This order in the standard multi-source order data is then recorded as a malicious order and removed. Conversely, if the abnormal risks are not present in the standard multi-source order data, then it is determined that there are no abnormal risks in the multi-source orders corresponding to the e-commerce platform, and adaptive processing will continue to be performed on the standard multi-source order data.

[0033] It should be noted that the feature library of order-brushing behavior corresponding to the adaptive model is statistically obtained based on the data training results of the adaptive model during continuous execution. It serves as a reference for judging whether there are abnormal risks in multi-source orders on e-commerce platforms, facilitating timely removal and identification of malicious orders. The specific performance characteristics of the orders include high-frequency ordering and refund rate, etc.

[0034] S3. Adaptive Order Processing: Adaptive processing is performed on the valid multi-source order data corresponding to the e-commerce platform through an adaptive model to obtain the processing set of valid orders corresponding to the e-commerce platform. The execution process of the processing set of valid orders corresponding to the e-commerce platform is monitored, monitoring data is obtained, and the adaptive processing results of orders corresponding to the e-commerce platform are analyzed. When there are processing omissions in the adaptive processing of orders corresponding to the e-commerce platform, the adjustment of the adaptive processing of orders corresponding to the e-commerce platform is analyzed.

[0035] In a specific embodiment, the specific analysis process for obtaining the processing set of valid orders corresponding to the e-commerce platform is as follows: input the valid multi-source order data corresponding to the e-commerce platform into the adaptive model, extract the main features corresponding to the valid multi-source order data, including order features, real-time resource features and fulfillment features. After extraction, the processing set corresponding to the valid orders is generated according to the processing criteria corresponding to the adaptive mode, namely priority, fulfillment resource matching and fulfillment path.

[0036] Based on the order data covered by the order features, a weighted scoring mechanism is used to assign weighted scores to the order data corresponding to valid multi-source orders, thereby obtaining a comprehensive priority score for each valid multi-source order. The comprehensive priority scores are then sorted in descending order to determine the priority for processing each order.

[0037] Based on the real-time inventory quantity and picking efficiency included in the real-time resource characteristics corresponding to valid multi-source orders, the content of each retail warehouse's corresponding product SKU and picking efficiency are obtained. The content and picking efficiency are then sorted in descending order. If a retail warehouse has the highest content and picking efficiency for both its corresponding orders, it is designated as the fulfillment resource matching warehouse. If the ranking results of the content and picking efficiency for each retail warehouse's corresponding orders are inconsistent, the content is used as the evaluation criterion to determine the fulfillment resource matching warehouse.

[0038] Finally, based on the congestion rate covered by the fulfillment characteristics corresponding to valid multi-source orders, the congestion rates of each path are sorted and compared, and the path with the lowest congestion rate is selected as the fulfillment execution path. Combining the above, a set of processing for valid orders corresponding to the e-commerce platform is obtained.

[0039] It should be noted that order data includes order amount, order timeliness, and user level; fulfillment characteristics include congestion rate and distance; the weighted scoring mechanism: multiple features affecting priority are normalized individually (0-1 points), then multiplied by preset weights, with the weights summed to obtain a comprehensive priority score. Higher scores result in higher priority. For example, order amount (weight 0.25), user level (weight 0.20), timeliness (weight 0.25), promotional tag (weight 0.15), and fulfillment urgency (weight 0.15): Order amount: High → Normalized score 1.0; User level: VIP → Score 1.0; Timeliness: Hourly delivery → Score 1.0; Promotional tag: During a major promotion → Score 1.0; Fulfillment urgency: High → Score 1.0; Comprehensive score = 1.0 × 0.25 +1.0×0.20+1.0×0.25+1.0×0.15+1.0×0.15=1.0 point, with professional staff setting the weights based on demand; the actual number of items picked per unit time and the standard number of items picked per unit time are obtained from the database, and the picking efficiency is obtained by dividing the actual number of items picked per unit time by the standard number of items picked per unit time and then multiplying by the percentage; the current amount to be delivered, the maximum capacity of the area, the number of overdue orders, the total amount to be delivered, and the historical congestion coefficient are obtained from the database, and the congestion rate is obtained by using (current amount to be delivered divided by the maximum capacity of the area * 0.5) + (number of overdue orders divided by the total amount to be delivered * 0.3) + historical congestion coefficient * 0.2, with a value range of 0 to 1, the closer to 1 the more congested, and the closer to 0 the smoother.

[0040] In a specific embodiment, the analysis of the adaptive processing results of orders corresponding to the e-commerce platform is carried out as follows: The monitoring data includes the order processing node completion rate, the fulfillment execution progress, and the deviation between the actual processing time and the expected processing time. If the order processing node completion rate is less than the preset expected order processing node completion rate, or the fulfillment execution progress is lower than the preset expected fulfillment execution progress, or the deviation between the actual processing time and the expected processing time exceeds the preset reference deviation range, it is determined that the adaptive processing of orders corresponding to the e-commerce platform has a processing omission and does not meet the normal processing standard. If the order processing node completion rate is greater than or equal to the preset expected order processing node completion rate, the fulfillment execution progress is higher than the preset expected fulfillment execution progress, and the deviation between the actual processing time and the expected processing time does not exceed the preset reference deviation range, it is determined that the adaptive processing of orders corresponding to the e-commerce platform does not have a processing omission and meets the normal processing standard.

[0041] It should be noted that professional staff set the expected order processing node completion rate, expected fulfillment progress, and reference deviation range based on the multi-source order requirements of the e-commerce platform. These settings serve as references for analyzing whether the e-commerce platform's adaptive order processing results are correctly handled, allowing for timely adjustments to ensure correct order processing. The number of completed valid nodes and the total number of standard nodes for the order are obtained from the database. The order processing node completion rate is calculated by multiplying the number of completed valid nodes by the total number of standard nodes for the order. The fulfillment progress is the weighted cumulative value of the completed nodes. For example, if order completion is 10%, payment completion is 15%, packaging completion is 10%, delivery is 5%, and receipt completion is 10%, then the total is 10% + 15% + 10% + 5% + 10% = 50%. The order completion time and actual start time are obtained from the database. The actual start time and actual time consumed are subtracted from the order completion time, and then the expected processing time is subtracted from the actual time consumed to obtain the deviation value.

[0042] In a specific embodiment, the analysis of the adjustment of the adaptive processing of orders corresponding to the e-commerce platform is carried out as follows: When there is a processing oversight in the adaptive processing of orders corresponding to the e-commerce platform, the performance features and problem-oriented dimensions corresponding to the order processing oversight are extracted. The adaptive model is corrected based on the performance features and problem-oriented dimensions, including the running parameters, feature weights and processing rules of the adaptive processing model, so as to obtain the processing adjustment direction of the adaptive model for multi-source orders, including order priority, SKU resource allocation and fulfillment path, and comprehensively complete the adjustment of the adaptive processing of orders corresponding to the e-commerce platform.

[0043] S4. Warning Prompt: A warning prompt will be issued when the mechanism corresponding to the e-commerce multi-source order adaptive processing model fails to switch successfully or when there is a processing oversight in the e-commerce platform's order adaptive processing. Please see Figure 2 As shown, a multi-source e-commerce order adaptive processing system based on big data includes a multi-source order data acquisition module, a multi-source order data analysis module, an order adaptive processing module, an early warning terminal, and a database.

[0044] The multi-source order data acquisition module is connected to the multi-source order data analysis module and the database, the multi-source order data analysis module is connected to the adaptive order processing module and the early warning terminal, the adaptive order processing module is connected to the early warning terminal and the database, and the multi-source order data analysis module is connected to the database.

[0045] The multi-source order data acquisition module is used to acquire multi-source heterogeneous order data from e-commerce platforms, thereby constructing an index system for the corresponding multi-source heterogeneous order data of e-commerce platforms, and establishing an adaptive processing model for multi-source e-commerce orders. The multi-source order data analysis module is used to collect order traffic data corresponding to e-commerce platforms, analyze whether the mechanism switching of the multi-source order adaptive processing model of e-commerce is correctly triggered, and analyze whether there are any abnormal risks in the multi-source orders corresponding to e-commerce platforms. The order adaptive processing module is used to adaptively process valid multi-source order data based on an adaptive model to obtain the processing set of valid orders corresponding to the e-commerce platform, monitor the execution process of the processing set of valid orders corresponding to the e-commerce platform, analyze the adaptive processing results of orders corresponding to the e-commerce platform, and analyze the adjustment of the adaptive processing of orders corresponding to the e-commerce platform when there are processing omissions in the adaptive processing of orders corresponding to the e-commerce platform. The early warning terminal is used to issue early warnings when the mechanism corresponding to the e-commerce multi-source order adaptive processing model fails to switch successfully or when there are processing omissions in the order adaptive processing of the e-commerce platform.

[0046] The database is used to store multi-source heterogeneous order data, key features, order flow data, standard multi-source order data, valid multi-source order data, and monitoring data.

[0047] This invention, based on acquiring multi-source heterogeneous order data, constructs an index and establishes an adaptive processing model for e-commerce multi-source orders. It analyzes whether the mechanism switching of the adaptive processing model is correctly triggered, processes standard multi-source order data, analyzes whether there are any abnormal risks in multi-source orders, and removes abnormal orders in real time. It then analyzes the processing set corresponding to valid orders, monitors the execution process of the processing set corresponding to valid orders, analyzes the adaptive processing results of orders on the e-commerce platform, and adjusts the adaptive processing of orders accordingly. This achieves efficient adaptive processing of multi-source e-commerce orders, ensuring timely and efficient order processing, further improving order processing efficiency, and solving problems such as incorrect orders, missing orders, or delayed inventory synchronization in e-commerce. It also automates order processing, reducing manual workload and costs, and efficiently ensures traceability and control throughout the entire order lifecycle, improving order processing quality and efficiency.

[0048] The examples described in this invention are not limited to the specific embodiments listed above. The examples are merely illustrative to facilitate understanding of the invention and do not constitute a limitation on the scope of protection of this invention. Any modifications, equivalent substitutions, etc., made within the spirit and principles of this invention should be included within the scope of protection.

[0049] The above description is merely an example and illustration of the concept of the present invention. 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 concept of the invention or exceed the scope defined in this specification, they should all fall within the protection scope of the present invention.

Claims

1. A multi-source e-commerce order adaptive processing method based on big data, characterized in that, Includes the following steps: S1. Acquisition of multi-source order data: Acquire multi-source heterogeneous order data from e-commerce platforms, then construct an index system for the corresponding multi-source heterogeneous order data of e-commerce platforms, and establish an adaptive processing model for multi-source e-commerce orders. S2. Multi-source order data analysis: Collect order traffic data corresponding to e-commerce platforms, analyze whether the mechanism switching of the multi-source order adaptive processing model of e-commerce is correctly triggered. When the mechanism is successfully switched, input the standard multi-source order data corresponding to the e-commerce platform into the adaptive processing model to analyze whether there are any abnormal risks in the multi-source orders corresponding to the e-commerce platform. S3. Adaptive Order Processing: Adaptive processing is performed on the valid multi-source order data corresponding to the e-commerce platform through an adaptive model to obtain the processing set of valid orders corresponding to the e-commerce platform. The execution process of the processing set of valid orders corresponding to the e-commerce platform is monitored, monitoring data is obtained, and the adaptive processing results of orders corresponding to the e-commerce platform are analyzed. When there are processing omissions in the adaptive processing of orders corresponding to the e-commerce platform, the adjustment of the adaptive processing of orders corresponding to the e-commerce platform is analyzed. S4. Warning prompt: A warning prompt will be issued when the mechanism corresponding to the e-commerce multi-source order adaptive processing model is not successfully switched or when there is a processing oversight in the order adaptive processing of the e-commerce platform.

2. The adaptive processing method for multi-source e-commerce orders based on big data according to claim 1, characterized in that, The specific process for obtaining multi-source heterogeneous order data from e-commerce platforms is as follows: Multi-source heterogeneous order data includes basic order information, payment information, logistics information, and user behavior information. Preprocessing is performed on the multi-source heterogeneous order data, including data format standardization conversion, redundant data deduplication, missing field imputation, noise data filtering, and platform data time sequence alignment, to obtain standard multi-source order data.

3. The adaptive processing method for multi-source e-commerce orders based on big data according to claim 2, characterized in that, The specific construction process for building an index system for multi-source heterogeneous order data corresponding to the e-commerce platform is as follows: Based on the preprocessed standard multi-source order data, a multi-dimensional joint index will be constructed according to order identifier, e-commerce platform source, time series, product SKU, delivery address, payment status and fulfillment status. At the same time, the data will be stored and indexed in layers according to hot data, real-time data and historical data to form a multi-dimensional layered index structure for association positioning and batch aggregation, thus obtaining the index system of multi-source heterogeneous order data corresponding to the e-commerce platform.

4. The adaptive processing method for multi-source e-commerce orders based on big data according to claim 3, characterized in that, The specific process for establishing the adaptive processing model for multi-source e-commerce orders is as follows: The main features of order processing are extracted from the index system of multi-source heterogeneous order data corresponding to e-commerce platforms. The extracted main features are normalized, discretized, and subjected to high-dimensional feature reduction to form the feature vector corresponding to the adaptive processing model of e-commerce multi-source orders. At the same time, the standard multi-source order data is divided into training set, validation set, and test set. Based on the integration of reinforcement learning algorithm and e-commerce order processing business rules, the model network structure is constructed. Order efficiency, abnormal order identification accuracy, and resource utilization are used as the core dimensions of the model reward function. The training set is used to perform iterative training of the model. During the training process, the model fitting effect is monitored in real time based on the validation set. After the training is completed, the test set is input into the model for multi-dimensional validation. If the model output index is consistent with the preset standard index result, the construction of the adaptive processing model of e-commerce multi-source orders is completed, and the parameters corresponding to the successful validation of the model are used as the initial running parameters of the model. Otherwise, the iterative training of the model continues.

5. The adaptive processing method for multi-source e-commerce orders based on big data according to claim 4, characterized in that, The analysis process regarding whether the mechanism switching of the e-commerce multi-source order adaptive processing model is correctly triggered is as follows: Order traffic data includes daily stable traffic, promotional peak traffic, and off-peak traffic. The order traffic data package corresponding to multi-source orders on the e-commerce platform is compared with the reference order traffic set corresponding to each switching mechanism stored in the database. The comparison yields the matching switching mechanism corresponding to the order traffic data. When the multi-source orders on the e-commerce platform are daily stable traffic, the basic steady-state switching mechanism is triggered. When the multi-source orders on the e-commerce platform are promotional peak traffic, the high-concurrency expansion switching mechanism is triggered. When the multi-source orders on the e-commerce platform are off-peak traffic, the resource-saving switching mechanism is triggered. When the adaptive processing model triggers a mechanism switch, the corresponding running data of the model after the trigger is obtained, including the switch response time and the order processing delay time after the switch. If either the switch response time or the order processing delay time after the switch is greater than the preset reference time value, it is determined that the mechanism corresponding to the e-commerce multi-source order adaptive processing model has not been successfully switched; otherwise, the switch is successful.

6. The adaptive processing method for multi-source e-commerce orders based on big data according to claim 5, characterized in that, The analysis process for determining whether there are any abnormal risks in multi-source orders corresponding to the e-commerce platform is as follows: The standard multi-source order data corresponding to the e-commerce platform is input into the adaptive processing model. Based on the presentation of the standard multi-source order data corresponding to the adaptive processing model after input, the order performance features are extracted and compared with the order fraud behavior feature library corresponding to the adaptive model. If the order performance features of a certain order data corresponding to the standard multi-source order data of the e-commerce platform are included in the order fraud behavior feature library corresponding to the adaptive model, it is determined that the multi-source order of the e-commerce platform has an abnormal risk. In this case, the order data in the standard multi-source order data is recorded as a malicious order and removed. Otherwise, it is determined that the multi-source order of the e-commerce platform does not have an abnormal risk, and adaptive processing will continue to be performed on the standard multi-source order data.

7. The adaptive processing method for multi-source e-commerce orders based on big data according to claim 6, characterized in that, The specific analysis process for obtaining the processing set of valid orders from the e-commerce platform is as follows: The effective multi-source order data corresponding to the e-commerce platform is input into the adaptive model. Based on the extraction of the main features corresponding to the effective multi-source order data, including order features, real-time resource features and fulfillment features, after extraction, the processing set corresponding to the effective order will be generated according to the processing criteria corresponding to the adaptive mode, namely priority, fulfillment resource matching and fulfillment path. Based on the order data covered by the order features, a weighted scoring mechanism is used to assign weighted scores to the order data corresponding to valid multi-source orders, so as to obtain the comprehensive priority score assigned to the valid multi-source orders. The comprehensive priority scores are then sorted in descending order to obtain the priority of order processing. Based on the real-time inventory quantity and picking efficiency contained in the real-time resource characteristics corresponding to valid multi-source orders, the content of each retail warehouse corresponding to the product SKU and the picking efficiency are obtained. The content and picking efficiency are sorted in descending order. If the content and picking efficiency of an order corresponding to a certain retail warehouse are both the highest, then the retail warehouse is designated as the fulfillment resource matching warehouse. If the ranking results of the content and picking efficiency of the order corresponding to the retail warehouse are inconsistent, the content is used as the evaluation criterion to obtain the fulfillment resource matching warehouse. Finally, based on the congestion rate covered by the fulfillment characteristics corresponding to valid multi-source orders, the congestion rates of each path are sorted and compared, and the path with the lowest congestion rate is selected as the fulfillment execution path. Combining the above, a set of processing for valid orders corresponding to the e-commerce platform is obtained.

8. The adaptive processing method for multi-source e-commerce orders based on big data according to claim 7, characterized in that, The analysis of the adaptive processing results of corresponding orders on the e-commerce platform is as follows: The monitoring data includes order processing node completion rate, fulfillment progress, and deviation between actual and expected processing time. If the order processing node completion rate is lower than the preset expected order processing node completion rate, or the fulfillment progress is lower than the preset expected fulfillment progress, or the deviation between actual and expected processing time exceeds the preset reference deviation range, then the e-commerce platform's adaptive order processing is deemed to have a processing oversight and does not meet normal processing standards. If the order processing node completion rate is greater than or equal to the preset expected order processing node completion rate, the fulfillment progress is higher than the preset expected fulfillment progress, and the deviation between actual and expected processing time does not exceed the preset reference deviation range, then the e-commerce platform's adaptive order processing is deemed to have no processing oversight and meets normal processing standards.

9. The adaptive processing method for multi-source e-commerce orders based on big data according to claim 8, characterized in that, The analysis of the adaptive order processing adjustments on the e-commerce platform is as follows: When there are oversights in the adaptive order processing of an e-commerce platform, the performance characteristics and problem-indicating dimensions corresponding to these oversights are extracted. The adaptive model is then corrected based on these performance characteristics and problem-indicating dimensions, including the operating parameters, feature weights, and processing rules of the adaptive processing model. This results in the adjustment of the adaptive model for processing multi-source orders, including order priority, SKU resource allocation, and fulfillment paths. The overall adjustment of the adaptive order processing of the e-commerce platform is then completed.

10. A multi-source e-commerce order adaptive processing system utilizing the multi-source e-commerce order adaptive processing method based on big data as described in any one of claims 1-9, characterized in that, include: The multi-source order data acquisition module is used to acquire multi-source heterogeneous order data from e-commerce platforms, thereby constructing an index system for the corresponding multi-source heterogeneous order data of e-commerce platforms, and establishing an adaptive processing model for multi-source e-commerce orders. The multi-source order data analysis module is used to collect order traffic data corresponding to e-commerce platforms, analyze whether the mechanism switching of the multi-source order adaptive processing model of e-commerce is correctly triggered, and analyze whether there are any abnormal risks in the multi-source orders corresponding to e-commerce platforms. The order adaptive processing module is used to adaptively process valid multi-source order data based on an adaptive model to obtain the processing set of valid orders corresponding to the e-commerce platform, monitor the execution process of the processing set of valid orders corresponding to the e-commerce platform, analyze the adaptive processing results of orders corresponding to the e-commerce platform, and analyze the adjustment of the adaptive processing of orders corresponding to the e-commerce platform when there are processing omissions in the adaptive processing of orders corresponding to the e-commerce platform. The early warning terminal is used to issue early warnings when the mechanism corresponding to the e-commerce multi-source order adaptive processing model fails to switch successfully or when there are processing omissions in the order adaptive processing of the e-commerce platform.

Images