An e-commerce management method and system based on big data analysis

Abstract

The present invention discloses an e-commerce management method and system based on big data analysis, which relates to the field of data management technology. The present invention realizes the stable operation and efficient data transmission of the e-commerce platform in a complex network environment by introducing big data analysis and dynamic slicing scheduling mechanism. With the help of real-time evaluation and classification processing of the network status of user terminals, the present invention can intelligently adapt to different types of e-commerce business needs, thereby realizing personalized transmission strategies, and effectively alleviates the network congestion problem during the promotion period through time-series differential encoding and interpolation compensation of high-priority data such as flash sales, thereby improving the response speed and user experience of key businesses. Combined with dynamic delay floating analysis and packet loss rate evolution judgment, the present invention can adaptively adjust the data sending rhythm, ensure the smooth scheduling of slice data when bandwidth fluctuates, and significantly improve business continuity and system fault tolerance.

Description

Technical Field

[0001] The present invention relates to the field of data management technology, and in particular to an e-commerce management method and system based on big data analysis. Background Art

[0002] As e-commerce platforms become increasingly complex, the types of data involved in e-commerce systems are becoming increasingly multi-source, time-series, and sensitive. In various promotional events, real-time inventory synchronization, and payment interactions, the varying network states of end users pose significant challenges to the stability of real-time transmission of critical business data. Especially in high-concurrency scenarios like flash sales, data transmission faces significant burst and real-time requirements. Traditional e-commerce systems generally employ unified transmission strategies and static caching mechanisms, which struggle to adapt to the data scheduling needs of dynamically changing user terminal network conditions. Furthermore, with the development of edge intelligence and big data analytics technologies, data processing mechanisms that dynamically configure terminal network states and adapt to business types are becoming a hot topic in research and engineering practice. There is an urgent need for intelligent slicing scheduling and adaptive reconstruction methods for business-sensitive data in complex scenarios to enhance critical business assurance capabilities under bandwidth constraints or network fluctuations.

[0003] CN117853146B provides a big data-based e-commerce operational data management system and method. Its technical focus is on optimizing e-commerce operational management efficiency by analyzing the impact of campaigns on operational service objectives and constructing an activity ranking model. This method primarily evaluates the contribution of e-commerce campaigns in both horizontal (between different campaigns) and vertical (the same campaign under different states) dimensions, improving the targeted nature of operational plan planning. However, this technical solution focuses on a posteriori analysis and ranking optimization of campaign operational data, without addressing differentiated processing of different business data in complex network environments or considering data transmission scheduling strategies for high-real-time data such as flash sale countdowns and payment requests. Therefore, while this method has some effectiveness in improving management efficiency, it lacks the ability to ensure stable real-time data transmission and prioritize services under varying terminal network conditions. This is particularly true in high-concurrency promotion scenarios, where it fails to effectively guarantee a user experience.

[0004] CN118608229A proposes an e-commerce data monitoring and order management method and system. By acquiring information such as user purchasing trends, sales forecasts, and order evaluation levels, this method optimizes product replenishment and order processing strategies. This method focuses on predicting user behavior and adjusting sales responses at the product level, significantly improving platform replenishment efficiency and order management quality. However, this solution relies on platform-wide data analysis and rule matching, and does not consider encoding, slicing, and scheduling strategies for different service types at the data transmission link level. For example, in scenarios where user network status fluctuates or latency spikes occur, the system lacks a differentiated priority transmission mechanism for time-sensitive data (such as payment information and inventory synchronization data), nor does it address adaptive decoding and timing compensation methods on the terminal side. Therefore, in terms of data management capabilities for real-time interactive services, this solution suffers from technical shortcomings such as low accuracy and untimely response, failing to meet the low latency and high reliability requirements of flash sales services. Summary of the Invention

[0005] In view of the problems existing in existing e-commerce management technologies when processing highly concurrent business data, the present invention is proposed.

[0006] Therefore, the problem to be solved by the present invention is how to generate a transmission configuration table through network status evaluation and service type mapping, perform differentiated slicing processing, and dynamically adjust the transmission rhythm based on delay and packet loss feedback, and finally realize data reorganization and interpolation compensation on the client through a sliding time window, thereby ensuring the transmission stability of key data and the continuity of the interactive experience.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides an e-commerce management method based on big data analysis, which includes forming a network status level according to the real-time network status parameters of the user terminal; forming a status mapping matrix in combination with the e-commerce business type; generating a slice configuration table based on the network status level and the status mapping matrix; performing differentiated slicing processing on different business data in the slice configuration table; during the promotion period, converting the second sale countdown data into a time difference format; adjusting the slice packet sending interval according to the feedback delay fluctuation and packet loss rate to ensure stable scheduling of slice data under bandwidth-constrained conditions; the client updates the local clock through a sliding time window, reorganizes the slices according to the business timing, prioritizes restoring payment and inventory data, and interpolates and compensates for the out-of-order second sale countdown data.

[0009] As an optimal solution of the e-commerce management method based on big data analysis described in the present invention, the generation of a slice configuration table based on the network status level and the status mapping matrix includes: weighted aggregation of real-time network status parameters to obtain a comprehensive network scoring coefficient, and combined with preset status judgment rules, marking it as a network status level; for the network status level, combined with the current e-commerce business type identifier, extracting the corresponding transmission demand mapping vector from the preset status mapping matrix, including the recommended slice granularity, redundant structure and transmission priority level; generating a corresponding slice configuration table based on the transmission demand mapping vector.

[0010] As a preferred solution of the e-commerce management method based on big data analysis described in the present invention, the differentiated slicing processing includes: classifying the current data set to be sliced ​​into subset groups according to business type, and pre-slicing each subset group with boundary constraints according to the slicing granularity to generate an intermediate data block sequence; applying a verification algorithm to the intermediate data block sequence based on redundant structure parameters to generate a verification slice set; marking the intermediate data block sequence and the verification slice set with unique numbers respectively, and combining them to form a main slice set and a verification slice set.

[0011] As a preferred solution of the e-commerce management method based on big data analysis described in the present invention, in which: in the main film set, for high-priority business data with a transmission priority level greater than a preset threshold, a timestamp coding layer is attached and a business timing identifier is embedded.

[0012] As a preferred solution of the e-commerce management method based on big data analysis described in the present invention, the operation identification bit marked as promotion mode in the slice configuration table is detected. If it is in an activated state, a time series differential conversion is performed on the flash sale countdown data to obtain a differential sequence, and the differential sequence is embedded in the specified window structure of the current main slice set.

[0013] As an optimal solution of the e-commerce management method based on big data analysis described in the present invention, the adjustment of the slice packet sending interval includes: obtaining the average round-trip delay fluctuation and packet loss rate in the past cycle; constructing a multi-state migration diagram based on the historical transmission status and the current delay fluctuation trend to realize the logical judgment and strategy switching of the rhythm state; integrating the packet loss rate evolution trajectory and the business sensitivity weight to construct a strategy decision tree and dynamically determine the redundancy strategy and the sending rhythm strategy.

[0014] As an optimal solution of the e-commerce management method based on big data analysis described in the present invention, the construction of a multi-state transition diagram includes: establishing a transmission delay event sequence library to record the continuous delay fluctuation within a time window; extracting a trigger factor set based on the delay change trend and its previous and subsequent fluctuation factors at each time point in the transmission delay event sequence library; constructing a delay-state transition diagram, in which each node represents a transmission rhythm state, and each directed edge represents a state transition rule, whose trigger condition is driven by an element in the trigger factor set; performing a rhythm state evaluation every n seconds during the current system operation, matching the transfer path in the delay-state transition diagram according to the delay event sequence, and realizing the adjustment of the slice sending interval and the sending queue priority; wherein n is a constant.

[0015] In a second aspect, the present invention provides an e-commerce management system based on big data analysis, which includes: a configuration generation module, which forms a network status level according to the real-time network status parameters of the user terminal; forms a status mapping matrix based on the e-commerce business type; and generates a slice configuration table based on the network status level and the status mapping matrix;

[0016] The data slicing module performs differentiated slicing processing on different business data in the slicing configuration table; during the promotion period, the flash sale countdown data is converted into a time series difference format;

[0017] The scheduling and control module adjusts the interval between slice packets according to the feedback of delay fluctuation and packet loss rate to ensure stable scheduling of slice data under bandwidth-constrained conditions;

[0018] In the timing reorganization module, the client updates the local clock through a sliding time window, reorganizes slices according to business timing, prioritizes restoring payment and inventory data, and interpolates and compensates for out-of-order flash sale countdown data.

[0019] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the e-commerce management method based on big data analysis as described in the first aspect of the present invention are implemented.

[0020] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the e-commerce management method based on big data analysis as described in the first aspect of the present invention are implemented.

[0021] The present invention has the following beneficial effects: By introducing big data analysis and a dynamic slicing scheduling mechanism, it achieves stable operation and efficient data transmission for e-commerce platforms in complex network environments. By leveraging real-time assessment and classification of user terminal network status, the present invention intelligently adapts to different types of e-commerce business needs, thereby implementing personalized transmission strategies. By implementing time-series differential encoding and interpolation compensation for high-priority data such as flash sales, it effectively alleviates network congestion during promotional periods, improving the response speed and user experience of key services.

[0022] Combining dynamic delay fluctuation analysis with packet loss rate evolution judgment, the present invention can adaptively adjust the data sending rhythm, ensuring smooth scheduling of sliced ​​data when bandwidth fluctuates, and significantly improving business continuity and system fault tolerance.

[0023] In summary, the present invention effectively solves the problems of insufficient data processing capabilities and unstable transmission in traditional e-commerce systems under high concurrency and weak network environments, and greatly improves the data scheduling efficiency, user interaction experience and overall business stability of the e-commerce platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0025] Figure 1 This is a flow chart of the e-commerce management method based on big data analysis.

[0026] Figure 2 A flowchart for adjusting the slice package sending interval for an e-commerce management method based on big data analysis.

[0027] Figure 3 This is a structural diagram of the e-commerce management system based on big data analysis. DETAILED DESCRIPTION

[0028] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0029] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0030] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0031] As mentioned in the above background technology, in various promotional activities, real-time inventory synchronization and payment interaction scenarios, the differences in the network status of terminal users pose a great challenge to the stability of real-time transmission of key business data. Especially in high-concurrency scenarios such as flash sales, there are obvious bursts and real-time requirements for data transmission. Traditional e-commerce systems generally adopt unified transmission strategies and static caching mechanisms, which are difficult to adapt to the data scheduling requirements under the dynamic changes in user terminal network conditions. In addition, with the development of edge intelligence and big data analysis technologies, data processing mechanisms based on dynamic configuration of terminal network status and business type adaptation have gradually become hot topics in research and engineering practice. There is an urgent need for an intelligent slicing scheduling and adaptive reconstruction method for business-sensitive data in complex scenarios to improve the key business guarantee capabilities under conditions of bandwidth limitations or network fluctuations.

[0032] Figure 1 FIG is a flow chart of an e-commerce management method based on big data analysis according to an embodiment of the present invention. Figure 1 As shown, the e-commerce management method based on big data analysis includes:

[0033] S1: Form a network status level based on the real-time network status parameters of the user terminal; form a status mapping matrix based on the e-commerce business type; generate a slice configuration table based on the network status level and the status mapping matrix.

[0034] In an embodiment of the present invention, the process of generating a slice configuration table includes:

[0035] (1) Perform weighted aggregation on the real-time network status parameters to obtain a comprehensive network scoring coefficient, which is then combined with the preset status judgment rules and marked as the network status level.

[0036] Specifically, real-time network status parameters include the average round-trip delay, bandwidth fluctuation rate, and instantaneous packet loss rate of the user terminal's current network status parameters; among them, the average round-trip delay is calculated based on the sliding window mechanism, the bandwidth fluctuation rate is calculated by comparing the change trends of the three most recent bandwidth peaks, and the instantaneous packet loss rate is the ratio of the number of packet losses per unit time to the total number of packets sent.

[0037] Using weighted aggregated network scoring coefficients instead of conventional threshold judgments can improve the ability to finely classify network status and avoid misjudgments caused by isolated parameters.

[0038] (2) Based on the network status level and the current e-commerce business type identifier, the corresponding transmission demand mapping vector is extracted from the preset status mapping matrix, including the recommended slice granularity, basic redundancy structure and initial priority level.

[0039] Among them, the status judgment rule sets a gradient interval for the network scoring coefficient, corresponding to five levels of "excellent, good, medium, poor, and extremely poor" to quantify the current network transmission capacity as the basis for subsequent data configuration.

[0040] It's worth noting that the network status rating doesn't use the traditional three-stage grading (e.g., excellent, fair, and poor). Instead, it introduces a nonlinear five-level distribution. Each level is set at unequal intervals based on the distribution density of the comprehensive network score coefficient, providing stronger boundary discrimination capabilities. For example, the score intervals could be set to: [0.0-0.2], [0.2-0.4], [0.4-0.6], [0.6-0.8], and [0.8-1.0], corresponding to excellent, good, fair, poor, and extremely poor, respectively.

[0041] By introducing this five-level gradient range, we can more effectively respond to intermediate network conditions with differentiated responses, avoiding policy duplication and response delays in edge states. This approach significantly improves adaptability in complex e-commerce scenarios, ensuring that critical services receive priority even in challenging network conditions.

[0042] (3) Generate the corresponding slice configuration table based on the transmission demand mapping vector.

[0043] Specifically, based on the network status level and the current e-commerce business type identifier (including orders, payments, logs, and inventory), the corresponding transmission demand mapping vector is extracted from the preset status mapping matrix. The transmission demand mapping vector includes the recommended slice granularity, redundancy structure, and transmission priority level.

[0044] It should be noted that the state mapping matrix is ​​constructed with the network state level and the e-commerce business type identifier as dual input dimensions, and each matrix unit maps an adaptation parameter set.

[0045] Furthermore, based on the transmission demand mapping vector, a corresponding slice configuration table is generated, which includes the setting parameters of slice granularity, redundancy structure and transmission priority under each type of e-commerce business.

[0046] It should be noted that this invention specifically proposes a dynamic caching mechanism for configuration tables. After a configuration table is generated, it is temporarily stored in a local cache area and tagged with the corresponding network status level. When the network status repeats or is similar in the future, the existing configuration table can be directly called without repeated calculations. This mechanism significantly reduces the system's computational burden under high-concurrency conditions, improves response speed, and enhances configuration reuse, supporting configuration migration and sharing across different terminals.

[0047] S2: Perform differentiated slicing on different business data in the slicing configuration table; during the promotion period, convert the flash sale countdown data into a time series difference format.

[0048] S2.1: Perform differentiated slicing processing on different business data in the slicing configuration table.

[0049] In an embodiment of the present invention, performing differential slicing processing includes the following steps:

[0050] (a) The data set to be sliced ​​is classified into subset groups according to business type, and each subset group is pre-sliced ​​with boundary constraints according to the slicing granularity to generate an intermediate data block sequence.

[0051] In the slice preprocessing stage, the original business data stream is first divided into multiple data subset groups with the same logical function according to the business type identifier contained in the generated slice configuration table.

[0052] During the segmentation process, the granularity alignment filling strategy is adopted to perform structural zero filling on the last block that is less than the slice granularity length.

[0053] After completing the data classification, the data block pre-segmentation operation is performed according to the slice granularity parameters matched by each subset group:

[0054] This pre-segmentation uses boundary constraint logic to ensure that the length of each data block is neatly segmented at a set granularity, avoiding data loss during cross-segment transmission.

[0055] In conventional practice, data often ends up with redundant patches or incomplete information due to incomplete granularity. To address this issue, the present invention employs a granularity-aligned padding strategy. This involves precisely padding the last block of data that falls short of the standard slice granularity through structural zero-filling. This strategy automatically determines the optimal padding length by detecting the mapping between the offset position of the last block of data in the original service data and the logical fields of the original data. This balances padding integrity with space efficiency, avoiding the waste of data space and extended parsing time caused by blind zero-filling.

[0056] This padding operation not only adheres to the same field specifications as the previously mentioned data structure, but also introduces a preset marker bit in the padding value, allowing the receiving end to identify and effectively remove the padding value during parsing, avoiding misunderstandings about data validity. This structured padding offers greater parsing stability and protocol compatibility than traditional static zero-padding, making it particularly suitable for business data with variable-length fields.

[0057] (b) Based on the redundant structure parameters, a checksum algorithm is applied to the intermediate data block sequence to generate a set of checksum pieces.

[0058] It should be noted that, unlike traditional static redundancy strategies, the dynamic selection mechanism of the coding algorithm of the present invention allows for flexible adaptation of the optimal verification strategy according to the current network level and service type, thereby improving the trade-off control capability between redundancy and computational complexity.

[0059] Preferably, the verification algorithm is dynamically loaded according to the coding identifier of the redundant structure parameter, such as using algorithm templates such as GF field linear transformation, local parity check, finite field coding, etc.

[0060] Specifically, if the coding identifier in the redundant structure parameter indicates the use of GF domain linear transformation, the original data block is mapped to a polynomial matrix by setting the linear combination coefficients on the finite field to generate a redundant check block with linear recovery characteristics; this algorithm is suitable for poor network status and has strong error correction performance and fast parallel computing capabilities. If the coding identifier is a local parity check, some main blocks are selected to construct a local check unit, and low-overhead local check fragments are quickly generated through XOR to repair lightweight network jitter. If finite field coding is used, a highly fault-tolerant data redundancy structure is further implemented through a class template in the GF domain, which is widely applicable to the transmission of large-scale order-related businesses. When the network status is poor, it automatically switches to a low-complexity redundancy algorithm.

[0061] The implementation of this dynamic loading mechanism is based on the preset parameter index in the slice configuration table. It searches for the function template corresponding to the coding identifier and automatically adjusts the algorithm parameters with the network level parameters of the transmission requirement dimension as input variables. It is highly adaptable and intelligent.

[0062] (c) The intermediate data block sequence and the check piece set are marked with unique numbers respectively, and combined to form the main piece set and the check piece set.

[0063] After encoding, to ensure accurate identification and recovery of data during complex network transmission, this invention identifies and numbers all generated data slices by type (primary slice, check slice). This numbering system combines a globally unique primary key with a business logic sequence number. This not only uniquely identifies each data slice, but also enables sequence recovery and error localization during the reassembly phase.

[0064] The combination process packages the numbered main fragment set and the corresponding check fragment set according to a predetermined structure to form the final transmission data packet cluster.

[0065] Furthermore, in the main slice set, for high-priority service data with a transmission priority level greater than a preset threshold, a timestamp coding layer is added and a service timing identifier is embedded to facilitate subsequent reorganization and delay compensation.

[0066] S2.2: Convert the flash sale countdown data into time series difference format.

[0067] In an embodiment of the present invention, the operation flag marked as "promotion mode" in the slice configuration table is detected. If it is in an activated state, a time series differential conversion is performed on the flash sale countdown data to obtain a differential sequence, and the differential sequence is embedded in the specified window structure of the current main slice set.

[0068] Specifically, all transmitted data is checked for business information containing a countdown field and its location within the slice structure is automatically identified. Upon detection, the original countdown sequence is encoded using temporal differencing (TD) logic. This differencing process significantly reduces data redundancy by recording the time intervals between consecutive time nodes (rather than absolute times). For example, if the original countdown is 60, 59, 58, 57..., after differentiation, it is simply represented as -1, -1, -1..., compressing the transmitted data dimension to a fraction of the original structure.

[0069] The structural layer transformation logic is triggered by the promotion logo, and the redundancy of the flash sale data is reduced through time difference, so that it can be efficiently transmitted in a limited-bandwidth environment, while providing a standard structural interface for subsequent interpolation and recovery.

[0070] S3: Adjust the slice packet sending interval according to the feedback delay fluctuation and packet loss rate to ensure stable scheduling of slice data under bandwidth-constrained conditions, such as Figure 3 As shown, the details are as follows.

[0071] S3.1: Synchronously obtain the average round-trip delay fluctuation ΔRTT and packet loss rate PLR ​​in the past cycle from the transmission feedback channel, and record this set of network feedback parameters in the dynamic parameter window.

[0072] Since conventional slice scheduling methods are mostly based on fixed intervals or static priority rules and lack the ability to respond to network status in real time, problems such as increased service response delay and frequent packet loss of core tasks occur in complex business scenarios. To this end, the present invention uses real-time feedback of round-trip delay fluctuation ΔRTT and packet loss rate PLR ​​and other parameters to dynamically adjust the sending interval of data slices in a communication environment where network bandwidth resources are limited and dynamically fluctuating, thereby achieving a more stable scheduling mechanism with business differentiation awareness.

[0073] To prevent isolated parameter mutations (such as a one-time packet loss peak or instantaneous congestion) from affecting the stability of the overall scheduling strategy, this step introduces a sliding window mechanism to form a dynamic time window with the feedback data of the last N frames.

[0074] This window has two functions: first, it aggregates the sliding mean and deviation of ΔRTT and PLR, and second, it provides short-term statistical analysis of their trend changes. The value of N can be set empirically based on the frequency of network fluctuations, typically between 5 and 10. This mechanism provides the system with "short-term memory," buffering sudden anomalies at the data flow level and improving the consistency and stability of scheduling policy responses.

[0075] S3.2: Based on the historical transmission status and the current delay fluctuation trend, a multi-state migration diagram is constructed to implement logical judgment and strategy switching of the rhythm state.

[0076] In the embodiment of the present invention, the specific operations include the following steps:

[0077] S3.2.1: Establish a transmission delay event sequence library to record the continuous delay fluctuations within a time window, including the value of the delay ΔRTT at each time point, the direction of state change (increasing, decreasing, stable), and the corresponding business type (such as payment, inventory, flash sales, etc.).

[0078] S3.2.2: Extract a set of trigger factors based on the delay change trend and fluctuation factors at each point in the transmission delay event sequence library. Taking payment tasks as an example, if the ΔRTT increases for two consecutive cycles and the fluctuation range exceeds 30ms, it can be identified as a high-latency intervention point. For inventory tasks, if the request frequency per second exceeds a threshold (such as 10 times) and the delay continues to increase, it is determined to be a bandwidth contention-type fluctuation.

[0079] S3.2.3: Construct a delay-state transition graph. Each node in the graph represents a transmission rhythm state, such as rhythm drift, congestion accumulation, bottleneck outbreak, or recovery state. Each directed edge represents a state transition rule, and its triggering condition is driven by an element in the trigger factor set.

[0080] The above uses business activity as one of the driving parameters for state transition and introduces an important business driving mechanism. For example, when the increase in latency is small, but the activity of key businesses (such as payment) suddenly increases, it is still possible to enter the warning state in advance, thereby realizing forward-looking strategic defense.

[0081] S3.2.4: During the current system operation, a rhythm state evaluation is performed every n seconds. The transition path in the delay-state transition diagram is matched according to the delay event sequence to adjust the slice sending interval and the sending queue priority; where n is a constant.

[0082] Exemplarily, the rhythm status of the received network events is periodically evaluated with a set evaluation period of n seconds. Its main goal is to compare and match the delay event sequence with the preset delay-state transition diagram to determine whether the current network has entered one of the four states: rhythm drift, congestion accumulation, bottleneck outbreak or recovery state.

[0083] To implement the above evaluation logic, we first need to standardize the description of the delay events that occur within the cycle. Expressed as ,in: The time when the event occurred; is the delay difference between this moment and the previous event; The business priority weight corresponding to the event.

[0084] Form a sequence of events in each n-second period , generate feature maps according to the following rules:

[0085] When m delay increase events occur consecutively in the LES and the average service priority weights are all greater than the threshold, the cycle is determined to be a rhythm drift-accumulation precursor state and matched to the D0→D1 path in the delay-state transition diagram;

[0086] If at least two Event pairs with opposite transition directions that occur within a time period with an average interval less than τ are identified as the rhythm adjustment critical interval and matched to the D1→D2 path to monitor whether the rhythm adjustment mechanism has entered the adaptive phase; τ is the interval threshold.

[0087] If all events in LES There is no negative jump (no downward trend) in a continuous number of seconds, and the total If the growth exceeds the convergence threshold, the rhythm will be out of control and the D2→D3 path will be matched;

[0088] If in two consecutive cycles, the LES If the delay shows a monotonically decreasing trend and the delay variance is less than the set stability threshold, the state is determined to have returned to S0 (normal), and the D3→D0 or D2→D0 path is matched.

[0089] Among them, the above D0~D3 correspond to the four states of recovery, rhythm drift, congestion accumulation and bottleneck outbreak respectively.

[0090] It should be clarified that terms such as rhythm drift-accumulation precursor, rhythm adjustment critical interval, and rhythm out of control are essentially nouns used to describe the judgment stage or trigger conditions in the state evolution process. Semantically, they can be regarded as supplements to the detailed trigger logic of D1~D3 states, rather than independent state nodes.

[0091] The specific mapping is as follows: the rhythm drift-accumulation precursor corresponds to the transition trigger condition from D0 to D1, describing the precursory pattern of the initial appearance of rhythm drift and the obstruction of high-priority services; the rhythm adjustment critical interval corresponds to the transition trigger condition from D1 to D2, indicating that the system is in a stage of adaptive rhythm adjustment but the effect is uncertain; the rhythm out of control corresponds to the state from D2 to D3, that is, the adjustment has completely failed and the rhythm is out of control.

[0092] Through the above-mentioned method based on matching event sequences with state graph paths, periodic self-assessment of the network rhythm state is achieved.

[0093] S3.3: Integrate the packet loss rate evolution trajectory and service sensitivity weight, build a policy decision tree, and dynamically determine the redundancy strategy and sending rhythm strategy.

[0094] S3.3.1: Establish a historical packet loss rate trajectory sequence. This sequence is based on business type and records the evolution history of packet loss rates for various types of business data packets and the associated events with key performance anomalies (such as inventory synchronization failures, payment response interruptions, etc.).

[0095] S3.3.2: Determine transmission sensitivity weight labels for different types of services, such as 1.0 for payment data, 0.8 for inventory data, and 0.4 for advertising recommendations (this is only an example setting and needs to be set according to actual operations), to form a weight table.

[0096] S3.3.3: Construct a policy decision tree based on PLR, service weight, and packet loss evolution pattern.

[0097] The decision-making path is divided into three layers. The first layer uses the PLR ​​threshold change trend as the initial screening branch (for example, PLR continues to rise / fall / fluctuate stably). The second layer uses business sensitivity weights for weighted ranking (for example, payment types prioritize redundancy, while recommendation types allow for loose policies). The third layer uses historical evolution patterns to identify whether it is a critical fluctuation cycle (for example, there have been three PLR ​​surges in the past five minutes).

[0098] S3.3.4: After the decision is executed, the results are recorded and an impact-feedback comparison table is formed to be used for subsequent optimization of similar situations (for example, whether a certain strategy reduces the PLR ​​or improves the response rate of the main business).

[0099] Through the above operations, it is no longer a static threshold judgment, but a dynamic decision path is constructed by integrating packet loss trajectory, business sensitivity and evolution direction, which is conducive to adaptive scheduling of slicing strategies in complex scenarios.

[0100] S4: The client updates the local clock through a sliding time window, reorganizes the slices according to the business time sequence, prioritizes restoring payment and inventory data, and interpolates and compensates for out-of-order flash sale countdown data.

[0101] The client maintains a sliding time window. When a new slice arrives and it is detected that it contains a timing difference or an absolute timestamp, the client will initiate a correction operation of the local synchronized clock.

[0102] The specific process is as follows: First, the drift rate is calculated based on the time deviation between the new slice and the historical slice, and the clock offset is dynamically smoothed using an exponentially weighted average algorithm. Second, the local synchronization clock update amplitude is determined based on the drift rate and the length of the data update time window to avoid drastic jumps that may cause subsequent reorganization logic disorder. This mechanism effectively avoids the sudden changes that traditional hard-overwrite updates may cause, ensuring that the client's logical clock gradually and consistently matches the server's rhythm.

[0103] After updating the synchronized clock, the client sorts all slices to be reassembled in the sliding window in ascending order according to the business sequence identifiers in the slices, and performs a group reconstruction operation to generate a set of chunks. High-priority businesses (such as payment confirmations and inventory changes) are reassembled by the priority scheduling queue and trigger early processing. The main goal of this reassembly operation is to merge multiple data slices that belong to the same business process but were split and sent due to network reasons into a complete business record.

[0104] If a time difference is missing or misordered in a service slice, the interpolation compensation mechanism is activated: the previous frame and the next frame of the slice are searched in the sliding window, their timing information is extracted respectively, and the missing timing of the slice is estimated using the linear interpolation algorithm. Estimated time The calculation method can be expressed as:

[0105]

[0106] in, and is the time between the previous frame and the next frame; and If the interpolated frame is not updated for more than 2 seconds, the server is forced to request the full data.

[0107] Furthermore, because the display rhythm of data fields containing countdown semantics, such as the "countdown seconds" in promotional services such as limited-time sales and flash sales, needs to be strictly synchronized with the service trigger clock, when the reorganized data contains fields with countdown semantics (such as promotional flash sales), the client constructs a decreasing timing buffer pool within the sliding time window. Within this buffer pool, each record will decrease according to the standard clock rhythm and is used to drive the countdown display on the front-end page. For those frames with countdown values ​​interrupted due to missing timing or lost slices, the countdown value is rewritten at equal intervals. That is, based on the most recently available countdown frame, the intermediate value is estimated over time and filled in, thereby achieving a visually continuous and smooth countdown effect and enhancing the user experience.

[0108] During actual operation, if a sudden change frame (i.e., a frame in which the countdown value suddenly regresses or significantly adjusts) is received, it is interpreted as a timing correction signal from the time master. At this point, the corresponding buffer pool field value is immediately refreshed, and the starting point parameter of the interpolation rewrite function is updated to ensure that the logical timing in the buffer pool always aligns with the actual server progress rhythm, thus preventing misjudgment or premature display of the countdown due to network instability.

[0109] Furthermore, this embodiment also provides an e-commerce management system based on big data analysis, such as Figure 3 Shown, including,

[0110] The configuration generation module 100 generates a network status level based on the real-time network status parameters of the user terminal; generates a status mapping matrix based on the e-commerce business type; and generates a slice configuration table based on the network status level and the status mapping matrix.

[0111] The data slicing module 200 performs differentiated slicing on different business data in the slicing configuration table; during the promotion period, the flash sale countdown data is converted into a time series difference format;

[0112] The scheduling and control module 300 adjusts the slice packet sending interval according to the feedback delay fluctuation and packet loss rate to ensure stable scheduling of slice data under bandwidth-constrained conditions;

[0113] In the timing reorganization module 400, the client updates the local clock through a sliding time window, reorganizes the slices according to the business timing, prioritizes restoring payment and inventory data, and interpolates and compensates for the out-of-order flash sale countdown data.

[0114] This embodiment also provides a computer device suitable for the e-commerce management method based on big data analysis, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the e-commerce management method based on big data analysis proposed in the above embodiment.

[0115] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0116] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the e-commerce management method based on big data analysis proposed in the above embodiment is implemented.

[0117] In summary, this invention, through the introduction of big data analysis and a dynamic slicing scheduling mechanism, achieves stable operation and efficient data transmission for e-commerce platforms in complex network environments. By leveraging real-time assessment and classification of user terminal network status, this invention intelligently adapts to different types of e-commerce business needs, thereby implementing personalized transmission strategies. By implementing time-series differential encoding and interpolation compensation for high-priority data such as flash sales, network congestion during promotional periods is effectively alleviated, improving the response speed and user experience of key services.

[0118] Combining dynamic delay fluctuation analysis and packet loss rate evolution judgment, the present invention can adaptively adjust the data sending rhythm, ensure the smooth scheduling of slice data when bandwidth fluctuates, and significantly improve business continuity and system fault tolerance.

[0119] In summary, the present invention effectively solves the problems of insufficient data processing capabilities and unstable transmission in traditional e-commerce systems under high concurrency and weak network environments, and greatly improves the data scheduling efficiency, user interaction experience and overall business stability of the e-commerce platform.

[0120] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An e-commerce management method based on big data analysis, characterized by: include: Forming a network status level based on real-time network status parameters of the user terminal; Form a state mapping matrix based on e-commerce business types; Generate a slice configuration table based on the network status level and the status mapping matrix; Performing differentiated slicing processing on different business data in the slicing configuration table; During the promotion period, convert the flash sale countdown data into time series difference format; Adjust the interval between slice packets based on the feedback of latency fluctuation and packet loss rate to ensure stable scheduling of slice data under bandwidth-constrained conditions; The client updates the local clock using a sliding time window, reorganizes the slices according to business chronology, prioritizes restoring payment and inventory data, and interpolates and compensates for out-of-order flash sale countdown data. Generating a slice configuration table based on the network status level and the status mapping matrix includes: Perform weighted aggregation on real-time network status parameters to obtain a comprehensive network scoring coefficient, which is then combined with preset status judgment rules to be marked as a network status level; Based on the network status level and the current e-commerce business type identifier, the corresponding transmission demand mapping vector is extracted from the preset status mapping matrix, including the recommended slice granularity, redundancy structure and transmission priority level; Generate a corresponding slice configuration table based on the transmission requirement mapping vector.

2. The e-commerce management method based on big data analysis according to claim 1, characterized in that: The differentiated slicing process includes: The current data set to be sliced ​​is classified into subset groups according to business type, and each subset group is pre-sliced ​​with boundary constraints according to the slicing granularity to generate an intermediate data block sequence; Applying a checksum algorithm to the intermediate data block sequence based on the redundant structure parameters to generate a checksum set; The intermediate data block sequence and the check piece set are respectively marked with unique numbers and combined to form a main piece set and a check piece set.

3. The e-commerce management method based on big data analysis according to claim 2, characterized in that: In the master slice set, for high-priority service data whose transmission priority level is greater than a preset threshold, a timestamp coding layer is added and a service timing identifier is embedded.

4. The e-commerce management method based on big data analysis according to claim 3, characterized in that: Detect the operation flag marked as promotion mode in the slice configuration table. If it is in an activated state, perform time series difference conversion on the flash sale countdown data to obtain a differential sequence, and embed the differential sequence into the specified window structure of the current main slice set.

5. The e-commerce management method based on big data analysis according to claim 1, characterized in that: The adjusting of the slice packet sending interval includes: Obtain the delay fluctuation and packet loss rate in the recent period; Based on historical transmission status and current latency fluctuation trends, a multi-state migration diagram is constructed to implement logical judgment and strategy switching of rhythm status. Integrate the packet loss rate evolution trajectory with the service sensitivity weight to build a policy decision tree and dynamically determine the redundancy strategy and sending rhythm strategy; The constructing of the multi-state transition diagram includes: Establish a transmission delay event sequence library to record the continuous delay fluctuations within a time window; Extract the trigger factor set based on the delay fluctuation trend and its preceding and following fluctuation factors at each time point in the transmission delay event sequence library; Construct a multi-state transition graph, where each node represents a transmission rhythm state, and each directed edge represents a state transition rule, whose triggering condition is driven by an element in the trigger factor set; During the current system operation, a rhythm state evaluation is performed every n seconds. The transfer path in the multi-state transition diagram is matched according to the delay event sequence to adjust the slice sending interval and the sending queue priority; where n is a constant.

6. An e-commerce management system based on big data analysis, based on the e-commerce management method based on big data analysis according to any one of claims 1 to 5, characterized in that: Also includes: A configuration generation module forms a network status level based on the real-time network status parameters of the user terminal; Form a state mapping matrix based on e-commerce business types; Generate a slice configuration table based on the network status level and the status mapping matrix; The data slicing module performs differentiated slicing processing on different business data in the slicing configuration table; During the promotion period, convert the flash sale countdown data into time series difference format; The scheduling and control module adjusts the interval between slice packets according to the feedback of delay fluctuation and packet loss rate to ensure stable scheduling of slice data under bandwidth-constrained conditions; In the timing reorganization module, the client updates the local clock through a sliding time window, reorganizes slices according to business timing, prioritizes restoring payment and inventory data, and interpolates and compensates for out-of-order flash sale countdown data.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the e-commerce management method based on big data analysis described in any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the e-commerce management method based on big data analysis described in any one of claims 1 to 5 are implemented.

Images