Method and system for optimizing e-commerce transaction platform

Abstract

The invention relates to the field of e-commerce, and discloses an optimization method and system for an e-commerce transaction platform, and the method comprises the following steps: carrying out the dynamic service management; s1, real-time data processing; s3, user intention modeling is carried out; s4, dynamically arranging a strategy; and S5, full-link monitoring is carried out, in the aspect of system architecture, an innovative dynamic management framework reconstructs a traditional service management mode through a service grid technology, and automatic sensing of micro-service topology and intelligent distribution of flow are achieved. A self-adaptive fusing and service degradation mechanism guarantees the service continuity in a high-concurrency scene, a multi-dimensional gray release function remarkably reduces the online risk of a new function, the platform has excellent elasticity and stability, a data processing level and a streaming batch integrated architecture break through the limitation of traditional offline analysis, second-level response is achieved through a unified data pipeline, and the service performance of the platform is improved. Data consistency is ensured through real-time ETL and dimension association, real-time data support is provided for precision marketing through online feature engineering, and the operation decision-making efficiency is greatly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of electronic commerce, in particular to an optimization method and system of an electronic commerce transaction platform. BACKGROUND

[0002] The current electronic commerce transaction platform has entered the development stage of intelligence, personalization and full-channel integration, and its technical architecture is facing a transformation from function implementation to experience optimization. Modern e-commerce systems generally use cloud-native architecture, which realizes flexible expansion through micro-service and container deployment, and integrates advanced technologies such as big data analysis, artificial intelligence and Internet of Things. From the perspective of technical implementation, these platforms have formed a complete technology stack including interaction layer, business layer, data layer and infrastructure layer.

[0003] The core technical bottlenecks faced by the current electronic commerce platform mainly manifest in three aspects: the traditional service governance mode is difficult to meet the needs of rapid business iteration, static resource configuration leads to insufficient system stability during large promotion, and new functions have long online cycle and high risk; in terms of data processing, the mainstream platform still adopts an offline batch processing-based analysis mode, which leads to the dependence of operational decision-making on historical data, the inability to capture user behavior changes in real time, and the serious impact of feature engineering lag on personalized recommendation effect; the user experience is poor, the static recommendation strategy based on rule engine cannot understand the complex intention of the user, especially when the user hesitates to browse and deeply compares prices, there is a lack of effective response plan, and the contradiction between privacy protection requirements and personalized service needs is increasingly prominent. These technical shortcomings directly affect the operational efficiency, market competitiveness and user experience of the platform. Therefore, an optimization method and system of an electronic commerce transaction platform are proposed. SUMMARY

[0004] The purpose of the present application is to provide an optimization method and system of an electronic commerce transaction platform to solve the problems raised in the background art.

[0005] To achieve the above purpose, the present application provides the following technical solution: an optimization method of an electronic commerce transaction platform, comprising the following steps: S1, dynamic service governance; S1, real-time data processing; S3, user intention modeling; S4, dynamic strategy arrangement; S5, full-link monitoring.

[0006] As a preferred, in S1, the following sub-steps are included: S101, service topology awareness: automatically discovering and maintaining the call dependency graph between microservices; S102, elastic traffic distribution: dynamically adjusting the traffic weight according to the real-time load of the service instance; S103, adaptive fusing: dynamically adjusting the fusing strategy based on service response time, error rate and other indicators; S104, gray release: supporting service version gray release in multiple dimensions such as user characteristics and traffic proportion; S105, service degradation: automatically enabling alternative service path when the system is overloaded.

[0007] As preferred, in S2, the following sub-steps are included: S201, data source access: unified access to multi-source data such as database change log, message queue and application log; S202, streaming ETL: real-time cleaning, conversion and enrichment of raw data; S203, state management: maintaining the intermediate state of stream processing job, supporting exactly-once processing semantics; S204, dimension association: real-time association of user behavior data with business entities such as goods and orders; S205, feature engineering: online calculation of user behavior features and product features.

[0008] As preferred, in S3, the following sub-steps are included: S301, behavior sequence modeling: using Transformer architecture to model the long-term dependence of user behavior sequence; S302, context fusion: integrating time, place, device and other context information; S303, sentiment analysis: analyzing the sentiment tendency in user comments and customer service conversations; S304, intent prediction: predicting the current most likely shopping intent of the user; S305, portrait update: real-time updating of user portrait to reflect the latest behavior.

[0009] As preferred, in S4, the following sub-steps are included: S401, state awareness: monitoring system resources, service health and business indicators; S402, strategy candidate generation: generating a set of candidate strategies based on the current state; S403, revenue estimation: estimating the expected revenue of each candidate strategy; S404, strategy selection: selecting the optimal strategy using the multi-armed bandit algorithm; S405, strategy execution: issuing the selected strategy to each service node.

[0010] As preferred, in S5, the following sub-steps are included: S501, tracking and burying points: deploying tracking points in key business paths; S502, Metric Calculation: Real-time calculation of service quality metrics such as success rate and latency; S503, Root Cause Analysis: Automatically analyzes the root causes of service anomalies; S504, Alarm Generation: Generate hierarchical alarms based on anomaly detection algorithms; S505, Visualization: Provides a multi-dimensional service quality visualization dashboard.

[0011] As a preferred option, the above also includes: S6, edge intelligent collaboration: deploying lightweight AI models on user terminals; S6 specifically includes the following sub-steps: S601, Local Intent Recognition: Initial user intent recognition is completed on the terminal device; S602. Privacy Protection: Federated learning technology is used to protect user privacy; S603, Hot Model Update: Dynamically update the terminal model without interrupting service; S604, Offline Service: Provides basic service capabilities when the network is unavailable; S605, Edge Cache: Preload resources that users may access.

[0012] As a preferred option, the above also includes: S7, blockchain evidence storage: implemented using consortium blockchain technology; S7 specifically includes the following sub-steps: S701, Operation Audit: Record key business operations to the blockchain; S702, Smart Contracts: Automatically execute business rules through smart contracts; S703, Data Ownership Confirmation: Realizing the confirmation of ownership of user data; S704, Anti-tampering certificate storage: Provides business certificates that cannot be tampered with; S705, Cross-platform mutual recognition: Supports mutual recognition and trust with other platforms.

[0013] This application also provides an e-commerce transaction platform optimization system, including: Service governance layer: Enables dynamic governance and traffic management of microservices; Real-time computing layer: Provides integrated stream and batch data processing capabilities; User understanding layer: Building dynamic user profiles and intent models; Strategy engine layer: Generates and executes optimal service strategies; Monitoring and operations layer; enabling end-to-end monitoring and alerting; Edge nodes are deployed on user terminals to enable edge computing; Blockchain networks provide notarization and smart contract functionality; The service governance layer is connected to the real-time computing layer, the real-time computing layer is connected to the user understanding layer, the user understanding layer is connected to the strategy engine layer, the strategy engine layer is connected to the monitoring and maintenance layer, the monitoring and maintenance layer is connected to the edge nodes, and the edge nodes are connected to the blockchain network.

[0014] Preferably, the aforementioned edge nodes support dynamic loading and updating of models, and the service governance layer, real-time computing layer, user understanding layer, policy engine layer, monitoring and maintenance layer, edge nodes, and blockchain network are all interconnected through a unified service bus.

[0015] Compared with the prior art, the present invention, employing the above technical solution, has the following technical effects: In terms of system architecture, the innovative dynamic governance framework reconstructs the traditional service management model through service mesh technology, enabling automatic awareness of microservice topology and intelligent traffic allocation. Adaptive circuit breaking and service degradation mechanisms ensure business continuity in high-concurrency scenarios, while multi-dimensional canary release functionality significantly reduces the risk of launching new features, giving the platform excellent elasticity and stability. Second, in terms of data processing, the integrated stream and batch architecture breaks through the limitations of traditional offline analysis. It achieves second-level response through a unified data pipeline, ensures data consistency through real-time ETL and dimension correlation, and provides real-time data support for precision marketing through online feature engineering, which greatly improves the efficiency of operational decision-making. Third, in terms of user experience, multimodal intent recognition combined with a reinforcement learning strategy engine enables deeply personalized services, edge intelligence technology balances response speed and privacy protection, and end-to-end monitoring ensures continuous optimization of service quality.

[0016] These innovations collectively build a user-centric intelligent service system, which not only improves platform operational efficiency but also enhances user stickiness and market competitiveness, providing solid technical support for the sustainable development of e-commerce platforms. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the process 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] Example 1 Please see Figure 1 This invention provides a technical solution: an optimization method for an e-commerce transaction platform, comprising the following steps: S1. Dynamic Service Governance Steps: Construct a microservice governance framework based on a service mesh to achieve dynamic configuration of service discovery, traffic management, and circuit breaker mechanisms, including the following sub-steps: S101, Service Topology Awareness: Automatically discover and maintain the call dependency graph between microservices; S102, Elastic Traffic Allocation: Dynamically adjusts traffic weights based on the real-time load of service instances; S103, Adaptive Circuit Breaker: Dynamically adjusts the circuit breaker strategy based on metrics such as service response time and error rate; S104, Gray Release: Supports gray release of service versions based on multiple dimensions such as user characteristics and traffic ratio; S105, Service Degradation: Automatically enable alternative service paths when the system is overloaded.

[0021] S2. Real-time data processing steps: Establish a unified stream and batch data processing pipeline, integrating the Flink real-time computing engine with the offline data warehouse, including the following sub-steps: S201, Data Source Access: Unified access to multiple data sources such as database change logs, message queues, and application logs; S202, Streaming ETL: Real-time cleaning, transformation, and enrichment of raw data; S203, State Management: Maintain the intermediate state of stream processing jobs and support exact one-time processing semantics; S204, Dimensional Association: Real-time association of user behavior data with business entities such as products and orders; S205, Feature Engineering: Online calculation of user behavior features and product features.

[0022] S3. User Intent Modeling Steps: Employing multimodal deep learning technology, this step integrates user behavior sequences, contextual environment, and sentiment tendencies to construct a dynamic user profile, including the following sub-steps: S301. Behavioral Sequence Modeling: Modeling long-term dependencies of user behavior sequences using the Transformer architecture; S302, Context Fusion: Integrating contextual information such as time, location, and device; S303. Sentiment Analysis: Analyze the sentiment trends in user reviews and customer service conversations; S304. Intent Prediction: Predict the user's most likely current shopping intent; S305. User Profile Update: Update user profiles in real time to reflect the latest behavior.

[0023] S4. Dynamic Strategy Orchestration Steps: Based on a reinforcement learning-based strategy engine, the optimal service strategy is generated according to the real-time system state and user profiles, including the following sub-steps: S401, Status Awareness: Monitor system resources, service health, and business metrics; S402, Policy Candidate Generation: Generate a set of candidate policies based on the current state; S403, Profit Forecast: Forecast the expected profit of each candidate strategy; S404, Strategy Selection: Use the multi-armed slot machine algorithm to select the optimal strategy; S405, Policy Execution: Distribute the selected policy to each service node.

[0024] S5. End-to-End Monitoring Steps: Implement end-to-end tracing from the user end to the server end, and establish a multi-dimensional service quality evaluation system, including the following sub-steps: S501, Tracking Points: Deploy tracking points in critical business paths; S502, Metric Calculation: Real-time calculation of service quality metrics such as success rate and latency; S503, Root Cause Analysis: Automatically analyzes the root causes of service anomalies; S504, Alarm Generation: Generate hierarchical alarms based on anomaly detection algorithms; S505, Visualization: Provides a multi-dimensional service quality visualization dashboard.

[0025] This application also provides an e-commerce transaction platform optimization system, including: Service governance layer: Enables dynamic governance and traffic management of microservices; Real-time computing layer: Provides integrated stream and batch data processing capabilities; User understanding layer: Building dynamic user profiles and intent models; Strategy engine layer: Generates and executes optimal service strategies; Monitoring and Operations Layer: Enables end-to-end monitoring and alerting; Edge nodes, deployed on user terminals, enable edge computing. Edge nodes support dynamic loading and updating of models. The blockchain network provides notarization and smart contract functions. The service governance layer, real-time computing layer, user understanding layer, strategy engine layer, monitoring and maintenance layer, edge nodes and blockchain network are all interconnected through a unified service bus. The service governance layer is connected to the real-time computing layer, the real-time computing layer is connected to the user understanding layer, the user understanding layer is connected to the policy engine layer, the policy engine layer is connected to the monitoring and operation layer, the monitoring and operation layer is connected to the edge nodes, and the edge nodes are connected to the blockchain network.

[0026] In summary, Embodiment 1, as the basic architecture, fully encompasses all the technical features of claims 1-6, constructing a dynamically responsive e-commerce platform optimization system. It achieves intelligent management of the microservice architecture through service mesh technology, reconstructing the traditional service governance model: when the platform conducts major promotional activities, the order service faces sudden traffic surges. At this time, the system can automatically identify overloaded nodes based on real-time monitoring data, dynamically adjust traffic allocation strategies, divert some requests to backup service instances, and trigger a circuit breaker mechanism to protect the core transaction chain from being dragged down. At the data processing level, this embodiment breaks through the limitations of traditional offline batch processing, designing a real-time computing architecture integrating stream and batch processing. Taking user payment behavior as an example, successful payment events immediately enter the stream processing pipeline through change data capture technology, completing data cleaning, dimension association, and feature calculation within millisecond latency, and updating key indicators such as user profiles and product sales in real time. For user intent understanding, a multimodal deep learning model is adopted, not only analyzing conventional clickstream data but also innovatively introducing fine-grained interaction features such as page heatmaps and mouse trajectories. When the system detects that a user is repeatedly comparing parameters on the product details page but has not placed an order, it will automatically trigger intelligent customer service intervention or push customized coupons. This deep intent understanding capability significantly improves the conversion rate.

[0027] Example 2 An optimization method for an e-commerce transaction platform, based on Embodiment 1, further includes: S6, Edge intelligent collaboration: Deploying a lightweight AI model on the user terminal; S6 specifically includes the following sub-steps: S601, Local Intent Recognition: Initial user intent recognition is completed on the terminal device; S602. Privacy Protection: Federated learning technology is used to protect user privacy; S603, Hot Model Update: Dynamically update the terminal model without interrupting service; S604, Offline Service: Provides basic service capabilities when the network is unavailable; S605, Edge Cache: Preload resources that users may access.

[0028] In summary, Embodiment 2 adds the edge intelligent collaboration technology of claim 7 to Embodiment 1, with in-depth optimization specifically for mobile user experience. This embodiment deploys a specially optimized lightweight AI model on the user terminal, enabling preliminary user behavior analysis and intent prediction locally on the device. For example, when a user browses products using an app, the terminal model can analyze subtle interaction features in real time, such as the user's scrolling speed, image zoom area, and dwell time, accurately determining the strength of their purchase intention. This edge computing mode reduces intent recognition latency to milliseconds. Regarding privacy protection, this embodiment employs an innovative federated learning framework. Sensitive user data is always stored on the terminal device; only anonymized model parameters are uploaded to the cloud for global model training. The system also features a differential privacy mechanism, adding carefully calculated noise to the data to ensure that individual information cannot be inferred from the aggregated results. For model updates, an innovative incremental hot update technology is used. When a new version of the recommendation model is trained in the cloud, only the differences are compressed and transmitted, silently updating when the user's device is idle. The entire process requires no user intervention and does not affect normal use. When network conditions are poor, edge nodes can continue to provide basic services based on local caching and lightweight models, ensuring that users can still have a smooth shopping experience in weak network environments.

[0029] Example 3 An optimization method for an e-commerce transaction platform, based on Embodiment 2, further includes: S7, blockchain evidence storage: implemented using consortium blockchain technology; S7 specifically includes the following sub-steps: S701, Operation Audit: Record key business operations to the blockchain; S702, Smart Contracts: Automatically execute business rules through smart contracts; S703, Data Ownership Confirmation: Realizing the confirmation of ownership of user data; S704, Anti-tampering certificate storage: Provides business certificates that cannot be tampered with; S705, Cross-platform mutual recognition: Supports mutual recognition and trust with other platforms.

[0030] In summary, Embodiment 3 combines the basic architecture of Embodiment 1 with the blockchain evidence storage technology of claim 8 to construct a trustworthy and transparent e-commerce transaction environment. This embodiment designs a distributed evidence storage system based on a consortium blockchain, recording key business operations such as order creation, payment confirmation, and logistics status updates on the blockchain to form tamper-proof electronic certificates. In the event of a transaction dispute, all parties can quickly clarify the attribution of responsibility based on the on-chain evidence storage. Regarding smart contract applications, this embodiment achieves automated execution of business rules. The data ownership confirmation mechanism is another major innovation of this embodiment. By storing user authorization information on the blockchain, it ensures that users have complete control over their personal data, and any third party using user data must obtain permission from the on-chain authorization record. The cross-platform mutual recognition mechanism, through the consensus characteristics of blockchain, allows different e-commerce platforms to securely share data such as credit ratings, breaking down data silos between platforms and avoiding the single point of failure risk of centralized data warehouses.

[0031] In summary, the three embodiments together constitute a complete technology evolution roadmap, addressing the core pain points of e-commerce platform development layer by layer, from infrastructure optimization to enhanced intelligent experience and the construction of a trusted environment. In terms of technical implementation, these three embodiments share fundamental capabilities such as dynamic service governance, real-time data processing, and deep user understanding, ensuring the consistency and scalability of the technical solution. Embodiments two and three, while maintaining the unchanged infrastructure, extend the system's capability boundaries through edge computing and blockchain technology, respectively. This modular design allows the platform to flexibly choose functional combinations according to business needs. From a business value perspective, this solution not only improves the operational efficiency and user experience of traditional e-commerce platforms but also effectively supports the development needs of emerging business models such as live-streaming e-commerce and social e-commerce, possessing broad application prospects and market potential.

[0032] Those skilled in the art will understand that the features described in the various embodiments and / or claims of the present invention can be combined or combined in various ways, even if such combinations or combinations are not explicitly described in the present invention. In particular, the features described in the various embodiments and / or claims of the present invention can be combined or combined in various ways without departing from the spirit and teachings of the present invention. All such combinations and / or combinations fall within the scope of the present invention.

Claims

1. An optimization method of an e-commerce transaction platform, characterized in that, Comprising the following steps: S1, dynamic service governance; S1, real-time data processing; S3, user intent modeling; S4, dynamic policy orchestration; S5, full-link monitoring.

2. The method of claim 1, wherein, In S1, the following sub-steps are included: S101, service topology awareness: automatically discover and maintain the call dependency graph between microservices; S102, elastic traffic distribution: dynamically adjust traffic weights according to the real-time load of service instances; S103, adaptive circuit breaking: dynamically adjust the circuit breaking strategy based on service response time, error rate, etc. indicators; S104, gray release: support multi-dimensional service version gray release according to user characteristics, traffic proportion, etc.; S105, service degradation: automatically enable alternative service paths when the system is overloaded.

3. The method of claim 2, wherein, In S2, the following sub-steps are included: S201, data source access: unified access to database change logs, message queues, and application logs, etc. multi-source data; S202, stream ETL: real-time cleaning, transformation, and enrichment of raw data; S203, state management: maintain the intermediate state of stream processing jobs, support exactly-once processing semantics; S204, dimension association: real-time association of user behavior data with goods, orders, and other business entities; S205, feature engineering: Online calculation of user behavior features and product features.

4. The method of claim 3, wherein, In S3, the following sub-steps are included: S301, behavior sequence modeling: use the Transformer architecture to model the long-term dependencies of user behavior sequences; S302, context fusion: integrate time, location, device, and other context information; S303, sentiment analysis: analyze the sentiment tendency in user reviews and customer service conversations; S304, intent prediction: predict the user's current most likely shopping intent; S305, portrait update: real-time update of user portrait to reflect the latest behavior.

5. The method of claim 4, wherein, In S4, the following sub-steps are included: S401, state awareness: monitor system resources, service health, and business indicators; S402, candidate policy generation: generate a set of candidate policies based on the current state; S403, revenue estimation: estimate the expected revenue of each candidate policy; S404, policy selection: use the multi-armed bandit algorithm to select the optimal policy; S405, policy execution: issue the selected policy to each service node.

6. The method of Claim 5, wherein, In S5, the following sub-steps are included: S501, tracking and point setting: deploy tracking points on key business paths; S502, indicator calculation: real-time calculation of success rate, delay, and other service quality indicators; S503, root cause analysis: automatically analyze the root cause of service exceptions; S504, alarm generation: generate hierarchical alarms based on anomaly detection algorithms; S505, visualization: provide multi-dimensional service quality visualization dashboards.

7. The optimization method of the e-commerce transaction platform according to claim 6, further comprising: S6, edge intelligence collaboration: deploying lightweight AI models on user terminals; In S6, the following sub-steps are included: S601, local intent recognition: preliminary user intent recognition on terminal devices; S602, privacy protection: protect user privacy using federated learning technology; S603, model hot update: dynamically update terminal models without interrupting service; ​ S604, Offline service: Provide basic service capabilities when the network is unavailable; S605, Edge caching: Preload resources that users may access.

8. The method of claim 7, further comprising: S7, Blockchain storage: Implemented using consortium chain technology; In S7, the following sub-steps are included: S701, Operation audit: Record key business operations to the blockchain; S702, Smart contract: Automatically execute business rules through smart contracts; S703, Data rights: Realize the ownership confirmation of user data; S704, Anti-tamper storage: Provide tamper-proof business credentials; S705, Cross-platform mutual recognition: Support mutual recognition and trust with other platforms. Comprise:

9. An e-commerce transaction platform optimization system, characterized in that, Service governance layer: Realize the dynamic governance and traffic management of microservices; Real-time computing layer: Provide flow batch integrated data processing capability; User understanding layer: Build dynamic user portrait and intent model; Strategy engine layer: Generate and execute optimal service strategy; Monitoring and operation layer: Realize full-link monitoring and alarm; Edge node: Deployed in user terminals to realize edge computing; Blockchain network: Provide storage and smart contract functions; The service governance layer is connected with the real-time computing layer, the real-time computing layer is connected with the user understanding layer, the user understanding layer is connected with the strategy engine layer, the strategy engine layer is connected with the monitoring and operation layer, the monitoring and operation layer is connected with the edge node, and the edge node is connected with the blockchain network.

10. The system of claim 9, wherein the edge node supports dynamic loading and updating of models, and the service governance layer, real-time computing layer, user understanding layer, strategy engine layer, and monitoring and operation layer, edge node, and blockchain network are interconnected through a unified service bus. ​ ​

Images