A method, device and medium for analyzing information from e-commerce live streaming sales.

By constructing a multidimensional association model and a dynamic weight allocation strategy, the problem of lacking multidimensional data integration in e-commerce live streaming sales was solved, enabling timely identification of abnormal interaction patterns and optimization of the live streaming process, thereby improving the live streaming effect.

CN120634600BActive Publication Date: 2026-11-17INSPUR ZHUOSHU BIG DATA IND DEV CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202510687598.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-05-27
Publication Date
2026-11-17
Estimated Expiration
2045-05-27

AI Technical Summary

Technical Problem

The lack of multi-dimensional data integration in e-commerce live streaming makes it impossible to identify abnormal interaction patterns in a timely manner, resulting in the inability to improve the live streaming effect.

Method used

By acquiring and combining information sets of livestreamers and products, a multi-dimensional correlation model is constructed to identify abnormal interaction patterns and optimize the livestreaming process through a dynamic weight allocation strategy.

Benefits of technology

It enables accurate analysis of live-streaming e-commerce information, timely detection of anomalies, optimization of the live-streaming process, and improvement of efficiency and effectiveness.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120634600B_ABST
    Figure CN120634600B_ABST
Patent Text Reader

Abstract

This invention provides a method, device, and medium for parsing e-commerce live-streaming sales information, belonging to the field of e-commerce live-streaming technology. It involves acquiring a set of anchor information and a set of product information to be compared; obtaining a pre-configured target anchor ID and a mapping information between the anchor ID and the corresponding preset time period anchor ID and product code; performing concatenation processing to obtain the corresponding anchor sales information; inserting the anchor sales information into a pre-constructed sales information verification table; and using a dynamic weight allocation strategy to prioritize abnormal interaction modes in real time, automatically triggering high-risk modes to specific partitions in the sales information verification table, while simultaneously generating multi-dimensional verification rules to drive adaptive optimization of the subsequent live-streaming process. This method can comprehensively and accurately analyze live-streaming sales information, promptly detect and handle abnormal situations, optimize the live-streaming process, and improve the efficiency and effectiveness of live-streaming sales.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of e-commerce live streaming technology, specifically relating to a method, device, and medium for parsing e-commerce live streaming sales information. Background Technology

[0002] Against the backdrop of the booming development of the e-commerce industry, live-streaming e-commerce, as an emerging marketing model, has rapidly become an important force in driving product sales due to its real-time interaction and intuitive display features.

[0003] In related technologies, e-commerce live-streaming sales data only collects basic product information or a brief introduction of the host, lacking integrated collection of multi-dimensional data such as the host's historical sales performance, fan demographics, product sales dynamics, and user feedback. This results in a lack of comprehensive data analysis, making it difficult to accurately assess the host's sales ability and the product's sales potential.

[0004] Current technologies lack effective anomaly detection methods for verifying information in live-streaming e-commerce. They often only allow for simple data comparisons and cannot identify complex abnormal interaction patterns. Existing e-commerce live-streaming processes are relatively fixed and lack the ability to dynamically adjust based on real-time data and anomalies. When problems such as poor product sales or negative user feedback occur during a live stream, the live-streaming strategy cannot be optimized in a timely manner, resulting in a failure to effectively improve the live-streaming effect. Summary of the Invention

[0005] This invention provides a method for analyzing e-commerce live-streaming sales information, which can comprehensively and accurately analyze live-streaming sales information, promptly detect and handle abnormal situations, optimize the live-streaming process, and improve the efficiency and effectiveness of live-streaming sales.

[0006] The methods include: S101: Obtain the set of information on the livestreamers and the set of information on the products to be compared; S102: Obtain the pre-configured target anchor ID from the preset anchor sales verification set, as well as the anchor ID and sales code mapping information corresponding to the target anchor ID for a preset time period; S103: Extract the anchor information and product information corresponding to the target anchor ID from the anchor information set and the product information set, respectively; S104: Based on the preset time period anchor ID and product code mapping information, the anchor information and product information are spliced ​​together to obtain the corresponding anchor product information; S105: Insert the livestreamer's sales information into a pre-built sales information verification form; S106: Execute a preset cross-validation algorithm to identify abnormal interaction patterns that differ from the livestreaming sales information in the livestreaming sales information verification form; S107: Based on the abnormal interaction mode, construct a multi-dimensional correlation model that includes the characteristics of the anchor's behavior, the dynamics of product sales and user feedback. Through a dynamic weight allocation strategy, prioritize the abnormal interaction mode in real time and automatically trigger the high-risk mode to a specific partition of the product information verification table. At the same time, generate multi-dimensional verification rules to drive the adaptive optimization of the subsequent live broadcast process.

[0007] Preferably, step S101 specifically includes: obtaining a set of information to be compared, which includes a set of spatiotemporal behavioral features of the anchor, a set of semantic associations of products, a set of user interaction heatmaps, and a set of multimodal features of the scene; The anchor's spatiotemporal behavior feature set includes historical live broadcast time distribution vectors and high-frequency word matrix; the product semantic association set includes cross-platform product description similarity vector space and competitor feature comparison matrix; the user interaction heat set consists of bullet screen emotional polarity time sequence stream and like behavior heat matrix; and the scene multimodal feature set integrates live broadcast audio and video semantic tag sequence and scene change detection vector.

[0008] Preferably, step S102 further includes: by integrating the anchor's historical sales records, real-time product inventory data, and user profile tags, based on the traffic characteristics of the live broadcast period and the product sales trend, constructing a dynamically adjusted mapping relationship between the anchor ID and the sales code, and verifying the real-time nature and accuracy of the mapping information through a cross-platform data interface, so as to support the accurate analysis and verification of subsequent live broadcast sales information.

[0009] Preferably, step S103 further integrates the anchor's real-time behavior data, product dynamic attributes, and user interaction information, and dynamically verifies and extracts structured anchor information and product information based on the anchor ID and product code mapping relationship for a preset time period.

[0010] Preferably, step S104 integrates the anchor's real-time behavior data, product dynamic attributes, and user interaction information, and dynamically verifies and splices the structured anchor's sales information based on the anchor ID and product code mapping relationship for a preset time period.

[0011] Preferably, before step S102, the method further includes: determining whether a configuration request triggered by a user corresponding to the set of goods to be verified by the livestreamer has been received; If so, display the preset configuration interface; Receive configuration information input by the user in the configuration interface; Based on the preset construction strategy, the corresponding configuration verification table is constructed using the configuration information; The set of products to be verified by the livestreamer is stored and processed.

[0012] Preferably, step S106 further verifies and generates a live-streaming sales analysis report by integrating the anchor's real-time behavior data, product dynamic attributes, and user interaction information, based on the anchor ID and product code mapping relationship for a preset time period.

[0013] Preferably, in step S107, the three types of data—anchor behavior, product sales data, and user feedback—are used as a multidimensional correlation model. The dynamic weight allocation strategy defines a risk coefficient, with the formula: Risk coefficient = (Frequency of anchor recommendations × Product return rate) ÷ Average user dwell time, and dynamically adjusts the weight parameters based on real-time data; For abnormal patterns with a risk coefficient ≥ 0.8, prioritize inserting them into the high-risk partition of the verification table and trigger the early warning mechanism; For patterns with a risk coefficient of 0.5 ≤ risk coefficient < 0.8, delay insertion into low-risk partitions; The adaptive optimization-driven approach links abnormal interaction modes with the live streaming process control module.

[0014] According to another embodiment of this application, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the e-commerce live-streaming sales information parsing method.

[0015] According to another embodiment of this application, a storage medium is also provided, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the steps of the e-commerce live-streaming sales information parsing method.

[0016] As can be seen from the above technical solutions, the present invention has the following advantages: The e-commerce live-streaming sales information parsing method provided in this application acquires a set of anchor and product information to be compared through multi-source data collection, ensuring the comprehensiveness and richness of the data. Obtaining the target anchor ID and mapping information from a pre-set set of anchor sales to be verified allows for focusing on key anchor and product combinations, making subsequent data processing more targeted. Based on the mapping information, corresponding anchor and product information is extracted, further narrowing the data scope and accurately acquiring data directly related to the target anchor's sales. This reduces data processing complexity and quickly locates key information. The anchor and product information are concatenated to form a complete anchor sales information unit, improving the data's logic and readability. The concatenated anchor sales information is inserted into a pre-built verification table, providing a unified storage platform for data verification and analysis. This facilitates rapid data querying, retrieval, and management. Executing a cross-validation algorithm to identify abnormal interaction patterns enables timely detection of data anomalies during the live-streaming sales process. A multi-dimensional association model is constructed based on these abnormal interaction patterns, prioritizing data through a dynamic weight allocation strategy and generating verification rules to drive live-streaming process optimization. It enables tiered management of anomalies, prioritizing high-risk anomalies to minimize business losses; dynamic weighting makes risk assessments more aligned with real-world scenarios; and generated verification rules drive continuous improvement of the livestreaming process, enhancing livestreaming sales performance and user experience. It comprehensively and accurately analyzes livestreaming sales information, promptly identifies and handles anomalies, optimizes the livestreaming process, and improves the efficiency and effectiveness of livestreaming sales. Attached Figure Description

[0017] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying 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.

[0018] Figure 1 Flowchart for information analysis methods in e-commerce live streaming sales; Figure 2 This is a schematic diagram of an electronic device. Detailed Implementation

[0019] The following describes in detail the e-commerce live-streaming sales information parsing method involved in this application. Specific details, such as particular system structures and technologies, are presented for illustrative purposes rather than limiting, to ensure a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application can also be implemented in other embodiments without these specific details.

[0020] It should be understood that, when used in this specification, the term "comprising" indicates the presence of the described feature, integral, step, operation, element, and / or component, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0021] It should be understood that "one or more" as mentioned in this application refers to one, two, or more, and "multiple" as mentioned in this application refers to two or more. In the description of this application, unless otherwise stated, " / " means "or," for example, A / B can mean A or B. The "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone.

[0022] The terms "one embodiment" or "some embodiments" used in this application mean that one or more embodiments of this application include the specific features, structures, or characteristics described in that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this application do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized.

[0023] In embodiments of the present invention, computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof. These programming languages ​​include, but are not limited to, object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or power server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (exemplarily using an Internet service provider for Internet connection).

[0024] 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.

[0025] Please see Figure 1 The diagram shows a flowchart of an e-commerce live-streaming sales information parsing method in a specific embodiment. The method includes: S101: Obtain the set of information about the anchors and products to be compared.

[0026] In some embodiments, data is collected by establishing data interfaces with multiple data sources, such as e-commerce platform databases, live streaming platform backends, and social media platforms. The streamer information set includes not only basic information such as streamer name, age, and number of followers, but also historical performance data such as the number of live streams, average number of viewers, interaction rate, sales conversion rate, and preferred sales techniques, as well as fan profile data such as geographic distribution, age group, and consumption habits. The product information set, in addition to basic attributes such as product name, price, and specifications, also includes inventory change records, sales trend data, user reviews, and comparisons of competitor prices and features. Furthermore, data related to the live streaming scenario is collected, such as the live streaming room setup style and lighting and sound effects parameters. This allows for a multi-dimensional reflection of the actual situation of the streamer's sales performance. A comprehensive data foundation helps in more accurately analyzing the streamer's sales performance and the product's sales potential.

[0027] S102: Obtain the pre-configured target anchor ID from the preset anchor sales verification set, as well as the anchor ID and sales code mapping information corresponding to the target anchor ID for a preset time period.

[0028] In some embodiments, the preset set of anchors to be verified for product promotion is a set of anchors to be prioritized and monitored, pre-defined by operations personnel based on business needs, and stored in a specific data table or file. By querying this set, the target anchor ID is obtained. Simultaneously, the mapping relationship between anchor IDs and product codes within a preset time period corresponding to the target anchor ID is found in the associated data table, clarifying the specific product code that each target anchor is responsible for promoting during that time period. This mapping relationship may also include information such as the product promotion time period and the main product identifier.

[0029] It should be noted that, based on the streamer's participation in the activity and key streamer tags, target streamer IDs are filtered from the pool of streamers awaiting verification for product sales. Then, based on the target streamer ID, a related query is performed in the associated mapping table to obtain the corresponding streamer ID and product code mapping information, achieving precise data location and extraction. This avoids indiscriminate processing of all streamer and product data, improves data processing efficiency, and allows for more targeted analysis of the sales performance of key streamers.

[0030] S103: Extract the anchor information and product information corresponding to the target anchor ID from the anchor information set and the product information set, respectively.

[0031] In some embodiments, based on the mapping information obtained in step S102, a filtering operation is performed on the large set of anchor information and product information. In the anchor information set, the live streaming performance data of the target anchor within a preset time period is filtered out, such as the change curve of the number of viewers, interactive message content, sales order data, etc.; simultaneously, the anchor's script text is extracted, and their speaking style, high-frequency vocabulary, etc., are analyzed. In the product information set, detailed attributes, price adjustment records, real-time inventory changes, user review details, and other information of the corresponding product are extracted based on the product code.

[0032] This embodiment narrows the data scope, extracting data directly related to the target livestreamer's sales efforts while removing a large amount of irrelevant data, thus reducing the complexity and workload of subsequent data processing. It enables a rapid focus on key information, improving the accuracy and efficiency of data analysis.

[0033] As an example of this application, step S103 further integrates the anchor's real-time behavior data, product dynamic attributes, and user interaction information, and dynamically verifies and extracts structured anchor information and product information based on the anchor ID and product code mapping relationship for a preset time period.

[0034] S104: Based on the preset time period anchor ID and product code mapping information, the anchor information and product information are spliced ​​together to obtain the corresponding anchor product information.

[0035] In some embodiments, the anchor information and product information extracted in step S103 are integrated according to the correspondence determined by the anchor ID and product code mapping information for a preset time period. For example, the anchor's live broadcast script and interaction during a certain time period are combined with the product introduction and promotional information corresponding to the product code during that time period to form a complete record. The record contains anchor-related information, product-related information, and the relationship between the two, such as the product sales time and sales method.

[0036] This embodiment uses mapping information as the link and basis for data splicing. By merging data records with the same associated target anchor ID and product code in the anchor information and product information, and combining relevant information fields according to a pre-designed data structure and format, structured anchor sales information is constructed.

[0037] In some specific embodiments, step S104 integrates the anchor's real-time behavior data, product dynamic attributes, and user interaction information, and dynamically verifies and splices the structured anchor's sales information based on the anchor ID and product code mapping relationship for a preset time period.

[0038] This embodiment uses a preset time-period mapping information between the streamer's ID and the product code to concatenate streamer information and product information, incorporating real-time streamer behavior, product inventory changes, price fluctuations, and user feedback. By comparing the data with the preset mapping information in real time and adjusting the concatenation strategy based on historical sales conversion rates and user dwell time, for example, automatically prioritizing the concatenation of alternative product information during periods of high inventory, or optimizing the product display order during periods of low user activity, the timeliness and completeness of the streamer's sales information are ensured.

[0039] S105: Insert the livestreamer's sales information into a pre-built sales information verification form.

[0040] In some embodiments, the pre-built product verification table is a database table structure designed based on the characteristics of the livestreamer's product information and the needs of subsequent verification and analysis. It includes fields for storing various aspects of the livestreamer's product information, such as livestreamer ID, product ID, product broadcast time, sales data, and user reviews. The livestreamer's product information assembled in step S104 is then inserted into the corresponding positions in the table according to the field requirements of the verification table, completing the data storage operation.

[0041] This embodiment utilizes the database's insert operation function, employing the insert data statements provided by the database management system to write structured livestreamer sales information records into a pre-created sales information verification table. During the insertion process, the database checks and verifies the data according to the table structure constraints, ensuring data integrity and accuracy. This facilitates quick data retrieval and access.

[0042] S106: Execute a preset cross-validation algorithm to identify abnormal interaction patterns that differ from the livestreaming sales information in the livestreaming sales information verification form.

[0043] In some embodiments, the preset cross-validation algorithm is a set of validation rules trained by business experts and data analysts based on the business rules, industry standards, and historical data of e-commerce live streaming. These rules include product description consistency rules (e.g., the product features introduced by the streamer must be consistent with the product details page), rules requiring verification of abnormal sales growth within a short period, and user interaction logic rules. By comparing and analyzing each streamer's sales information in the verification table against these rules, the logical relationships between data, numerical reasonableness, and semantic consistency are checked to identify abnormal interaction patterns that do not conform to the rules, such as streamers exaggerating product efficacy or sales data suspected of being manipulated.

[0044] This embodiment sequentially matches and judges the data records in the product information verification table against preset cross-validation rules. For each data record, the data fields are checked and analyzed according to the conditions and logical relationships set by the rules. Through conditional judgment, logical operations, and other methods, data records that do not conform to the rules are identified, and abnormal interaction patterns are determined. This allows for the timely detection of errors, false information, or unreasonable situations in product information, preventing erroneous information from misleading e-commerce live streaming business decisions and ensuring the accuracy and reliability of the data.

[0045] S107: Based on the abnormal interaction mode, construct a multi-dimensional correlation model that includes the characteristics of the anchor's behavior, the dynamics of product sales and user feedback. Through a dynamic weight allocation strategy, prioritize the abnormal interaction mode in real time and automatically trigger the high-risk mode to a specific partition of the product information verification table. At the same time, generate multi-dimensional verification rules to drive the adaptive optimization of the subsequent live broadcast process.

[0046] In some embodiments, for the abnormal interaction patterns identified in step S106, multi-dimensional data related to them is collected. Regarding the anchor's behavioral characteristics, data such as recent changes in livestream duration, repetition rate of statements, and frequency of interaction with viewers are collected; regarding product sales dynamics, data such as product sales fluctuation curves, refund rate changes, and inventory turnover rate are obtained; regarding user feedback, information such as user comments, evaluation texts, and complaint records are compiled. A multi-dimensional correlation model is constructed based on this data, with each dimension in the model being interconnected. Simultaneously, a dynamic weight allocation strategy is used to assign weights to each dimension based on the degree of impact of different dimension data on the livestream business. By calculating the comprehensive score of each abnormal interaction pattern under the multi-dimensional model, the abnormal interaction patterns are prioritized in real time. Abnormal interaction patterns with high scores and high risk levels are automatically transferred to specific partitions of the product information verification table for focused monitoring. Furthermore, based on the characteristics and causes of the abnormal interaction patterns, new multi-dimensional verification rules are generated, such as real-time keyword monitoring rules for anchors' exaggerated claims, to optimize subsequent livestream processes.

[0047] In some specific embodiments, when inserting difference data, analysis of abnormal interaction patterns is added, for example, inferring potential problems with the live streamer's product promotion strategy through user stay duration and click-through rate, rather than simply recording differences. By combining product sales trends with the live streamer's historical performance, the abnormal patterns of high-value products to be processed preferentially are dynamically adjusted.

[0048] In S103, in addition to extracting live streamer information and product information, bullet screen keywords, likes, share rates, as well as product inventory changes and price fluctuations are further integrated. Through multi-modal data fusion, a relationship graph of live streamers, products and users is constructed to identify the status in the live streamer's product promotion information. For example, when a live streamer recommends Product A, the proportion of users who actually purchase Product B is significantly abnormal.

[0049] In S107, the insertion of abnormal interaction patterns needs to be combined with dynamic weight allocation. Higher priority is assigned to patterns that occur frequently. For low-frequency but high-risk patterns, an alert is triggered and insertion is performed preferentially.

[0050] The high-frequency pattern is that the live streamer's sales pitch leads to an increase in the return rate of a certain type of product. The low-frequency but high-risk pattern is the contradiction between the sudden out-of-stock of product inventory and the live streamer's continuous promotion.

[0051] The dynamic weight allocation strategy of this embodiment is to define a risk coefficient, and the specific formula is: risk coefficient = (live streamer recommendation frequency × product return rate) ÷ average user stay duration, and the weight parameters are dynamically adjusted according to real-time data. For example, the weight of user stay duration is increased during peak hours.

[0052] For abnormal patterns with a risk coefficient ≥ 0.8, they are preferentially inserted into the high-risk partition of the verification form, and the alert mechanism is triggered; for patterns with 0.5 ≤ risk coefficient < 0.8, insertion into the low-risk partition is delayed to avoid resource waste.

[0053] In some specific embodiments, the live streamer recommendation frequency refers to the number of verbal recommendations for a specific product made by the live streamer per unit live broadcast duration. This data is generated by parsing the live broadcast audio sentence by sentence through speech recognition technology, identifying the pitch segments containing product names and promotion information, and performing statistics in combination with timestamps. For example, if a live streamer mentions a certain product 15 times in a 2-hour live broadcast, the recommendation frequency is 7.5 times per hour.

[0054] The product return rate is the ratio of the number of returned orders of a specific product to the total number of sales orders in a certain period, calculated based on the e-commerce platform order data. The system will automatically exclude return records caused by non-quality issues to ensure that the data reflects the real product acceptance.

[0055] The average user stay duration is obtained through the user behavior logs of the live streaming platform, which records the time interval from when a user enters the live room to when they leave. After excluding invalid data with excessively short stay duration (e.g., less than 10 seconds), the average stay duration of all viewing users is calculated. For example, if a live stream has a total of 1000 viewing users and the total stay duration is 5000 minutes, the average stay duration is 5 minutes.

[0056] The dynamic adjustment of weight parameters means that during peak hours, the weight of the average user stay duration is increased to 1.5 times the original coefficient, because the user traffic is large but the attention is scattered during this period, and the stay duration can better reflect the attractiveness of the product; while during the new product promotion period, the weight of the anchor recommendation frequency is increased to highlight the impact of promotion intensity on sales conversion. The adjustment process is automatically executed by the rule engine, which monitors signals such as live streaming period tags and promotion activity types in real time to trigger weight changes.

[0057] The adaptive optimization-driven method links the abnormal interaction mode with the live streaming process control module, for example: when it is identified that "the anchor's wording causes a sharp surge in the return rate of a certain product", the live streaming script is automatically adjusted to block relevant wording keywords; when it is detected that "users' interest in product A has decreased", the recommended product B is dynamically switched, and the mapping information of the product promotion code is updated.

[0058] It can be seen that this embodiment comprehensively applies data modeling, weight distribution and rule generation technologies. In terms of data modeling, multi-dimensional data is used to build a correlation model, and potential relationships and patterns between data are discovered through data mining and machine learning algorithms. The dynamic weight distribution strategy adjusts the weight of each dimension of data in real time according to business requirements and data changes, ensuring that the model can accurately reflect the severity of abnormal situations. The priority sorting algorithm sorts the abnormal interaction modes according to the comprehensive score calculated by the model. For high-risk modes, they are transferred to a specific partition through database operations. The rule generation module extracts key features and rules based on the analysis of abnormal modes to generate new verification rules. High-risk modes are centrally monitored to facilitate operation personnel to quickly locate and handle problems. The generated multi-dimensional verification rules can continuously improve the review and management mechanism of the live streaming process, realize continuous optimization and adaptive adjustment of the live streaming process, and improve the effect of live streaming product sales and user experience.

[0059] In one embodiment of the present invention, based on step S102, a possible embodiment will be given below to provide a non-limiting illustration of its specific implementation solution.

[0060] Before step S102, the method further includes: determining whether a configuration request corresponding to the livestreaming sales verification set triggered by a user has been received; if so, displaying a preset configuration interface; receiving the configuration information input by the user in the configuration interface; constructing a corresponding configuration verification table using the configuration information based on a preset construction strategy; and storing the livestreaming sales verification set.

[0061] Step S102 also includes: by integrating the anchor's historical sales records, real-time product inventory data, and user profile tags, based on the traffic characteristics of the live broadcast period and the product sales trend, constructing a dynamically adjusted mapping relationship between the anchor ID and the sales code, and verifying the real-time nature and accuracy of the mapping information through a cross-platform data interface to support the accurate analysis and verification of subsequent live broadcast sales information.

[0062] Specifically, the system continuously monitors user actions to determine if it has received a configuration request from a user regarding the verification set for livestream sales. This request can be triggered by the user clicking a specific button or entering a command in the system management backend. If a request is detected, the system immediately displays a preset configuration interface containing various interactive elements, such as text boxes for inputting target livestreamer filtering criteria, time selectors for choosing verification time periods, and drop-down menus for setting verification priorities. Users input detailed configuration information in the configuration interface, such as specifying key livestreamer IDs, setting specific verification time periods, and marking certain product categories as requiring priority verification. After obtaining the configuration information, the system categorizes and organizes the information according to a preset construction strategy, filling it into corresponding fields to build a configuration verification table. This table records the various parameters and relationships configured by the user. Finally, the system stores the constructed livestream sales verification set in a designated database table or storage area for easy retrieval later.

[0063] In this embodiment, when acquiring the target streamer ID and mapping information, the system first integrates multi-dimensional data. It retrieves the streamer's historical sales records from the historical data storage area, including the product categories promoted in past live streams, sales performance, and user feedback; it obtains real-time inventory data through a real-time data interface to understand the current available quantity and inventory warning status of products; and it extracts user profile tags from the user profile database to understand user consumption preferences, purchasing power, and other characteristics. Next, the system analyzes the traffic characteristics of the live stream, such as the number of users online at different times and the traffic source channels; it studies product sales trends, including recent sales changes of similar products and the distribution of peak sales periods. Based on these data analysis results, the system dynamically constructs a mapping relationship between streamer IDs and product codes. For example, during peak traffic periods, popular products are assigned to streamers with strong sales capabilities. After construction, the system interacts with e-commerce platforms and live streaming platforms through cross-platform data interfaces to verify the real-time performance and accuracy of the streamer and product data in the mapping information, ensuring that the information is consistent with the actual situation.

[0064] In step S106 of this embodiment, the anchor's real-time behavior data, product dynamic attributes, and user interaction information are integrated to verify and generate a live-streaming sales analysis report based on the anchor ID and product code mapping relationship for a preset time period.

[0065] By collecting host behavior data through cameras and microphones deployed on the live streaming platform, computer vision algorithms are used to identify body movements and facial expressions, while speech recognition and sentiment analysis technologies analyze tone and emotion. Product dynamic attribute data is obtained through a real-time database synchronization interface. User interaction information is captured in real-time from the live streaming platform's message queue, and natural language processing technology is used for semantic analysis and sentiment classification. Based on the verification results, the system calls a preset report template to transform the data into visual charts and textual analysis conclusions. Integrating multi-dimensional data enables comprehensive monitoring of the entire live-streaming e-commerce process, not only identifying data logic errors but also uncovering potential problems from the perspectives of user emotions and host performance.

[0066] As an example of the above method: A beauty brand conducts a live-streaming sales event on the Douyin platform and needs to verify the sales information of the anchor "ID007" during the time period of 20:00-22:00.

[0067] Based on step S101: The system captures real-time bullet comments in the live stream room of the anchor, such as 1,000 comments per minute, and the SKU information of "foundation liquid - matte" in the product database is 600 units in stock, with an original price of 300 yuan and a live stream price of 200 yuan, and integrates logistics data.

[0068] Based on step S102: According to the preset time period configuration table, confirm that the live streamer's product code rule for this time period is "brand + category + color", and generate product code "MH20230915-YZ-M".

[0069] Based on step S103: extract "average conversion rate of beauty category in the past 30 days: 15%" from the anchor information and "current inventory: 400 pieces" from the product information, and combine them into structured data.

[0070] Based on step S104: merge the data into JSON format, attach the anchor's signature and timestamp, and write it into the verification table.

[0071] Based on step S105: the verification form records show that the product code is not duplicated in historical data, and there is no conflict between the inventory and price fields.

[0072] Based on step S106: cross-validation revealed that the inventory of "gift cotton pads" associated with "foundation" in this live broadcast was only 10 pieces, but the product information marked "buy one get one free", triggering an inventory discrepancy anomaly.

[0073] Based on step S107: Build an association model analysis and find that the anomaly may lead to an increase in user complaint rate. Automatically mark the problem as "high risk", write it into the hot zone and trigger the rule: suspend the deduction of gift inventory and simultaneously notify the operations staff to replenish the stock.

[0074] It can be seen that the system completes the abnormal interception within 20 seconds, avoiding overselling losses. At the same time, through adaptive rule optimization, the system automatically verifies the ratio of gift inventory to main product inventory in subsequent similar sessions, reducing manual intervention.

[0075] like Figure 2 As shown, this application also provides an electronic device, including a display module 103, a memory 102, a processor 101, and a computer program stored in the memory and executable on the processor 101. When the processor 101 executes the program, it implements the steps of the e-commerce live streaming sales information parsing method.

[0076] In embodiments of the present invention, electronic devices include, but are not limited to, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments described and / or claimed herein.

[0077] In this embodiment, processor 101 may be implemented using at least one of an application-specific integrated circuit, a programmable logic device, a field-programmable gate array, a processor, a controller, a microcontroller, a microprocessor, or an electronic unit designed to perform the functions described herein. In some cases, such an implementation may be implemented within a controller. For software implementation, implementations such as processes or functions may be implemented with separate software modules that allow the performance of at least one function or operation. Software code may be implemented by a software application (or program) written in any suitable programming language, and the software code may be stored in memory and executed by the controller.

[0078] The display module 103 is used to display information input by the user or information provided to the user. The display module 103 may include a display panel, which may be configured in the form of a liquid crystal display, an organic light-emitting diode, or the like.

[0079] The memory 102 can be used to store software programs and various data. The memory 102 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0080] This application also provides a storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the e-commerce live-streaming sales information parsing method.

[0081] The storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example,, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0082] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for parsing information from e-commerce live-streaming sales, characterized in that, The methods include: S101: Obtain the set of information on the livestreamers and the set of information on the products to be compared; S102: Obtain the pre-configured target anchor ID from the preset anchor sales verification set, as well as the anchor ID and sales code mapping information corresponding to the target anchor ID for a preset time period; S103: Extract the anchor information and product information corresponding to the target anchor ID from the anchor information set and the product information set, respectively; S104: Based on the preset time period anchor ID and product code mapping information, the anchor information and product information are spliced ​​together to obtain the corresponding anchor product information; S105: Insert the livestreamer's sales information into a pre-built sales information verification form; S106: Execute a preset cross-validation algorithm to identify abnormal interaction patterns that differ from the livestreaming sales information in the livestreaming sales information verification form; S107: Based on the abnormal interaction mode, construct a multi-dimensional correlation model that includes the characteristics of the anchor's behavior, the dynamics of product sales and user feedback. Through a dynamic weight allocation strategy, prioritize the abnormal interaction mode in real time and automatically trigger the high-risk mode to a specific partition of the product information verification table. At the same time, generate multi-dimensional verification rules to drive the adaptive optimization of the subsequent live broadcast process. In step S107, the three types of data—anchor behavior, product sales data, and user feedback—are used as a multidimensional correlation model. The dynamic weight allocation strategy defines a risk coefficient, with the formula: Risk coefficient = (Frequency of anchor recommendations × Product return rate) ÷ Average user dwell time, and dynamically adjusts the weight parameters based on real-time data; For abnormal patterns with a risk coefficient ≥ 0.8, prioritize inserting them into the high-risk partition of the verification table and trigger the early warning mechanism; For patterns with a risk coefficient of 0.5 ≤ risk coefficient < 0.8, delay insertion into low-risk partitions; The adaptive optimization-driven approach links abnormal interaction modes with the live streaming process control module.

2. The e-commerce live-streaming sales information parsing method according to claim 1, characterized in that, Step S101 specifically includes: obtaining a set of comparison information containing the anchor's spatiotemporal behavior feature set, the product semantic association set, the user interaction heat set, and the scene multimodal feature set; The anchor's spatiotemporal behavior feature set includes historical live broadcast time distribution vectors and high-frequency word matrix; the product semantic association set includes cross-platform product description similarity vector space and competitor feature comparison matrix; the user interaction heat set consists of bullet screen emotional polarity time sequence stream and like behavior heat matrix; and the scene multimodal feature set integrates live broadcast audio and video semantic tag sequence and scene change detection vector.

3. The e-commerce live-streaming sales information parsing method according to claim 1, characterized in that, Step S102 also includes: by integrating the anchor's historical sales records, real-time product inventory data, and user profile tags, based on the traffic characteristics of the live broadcast period and the product sales trend, constructing a dynamically adjusted mapping relationship between the anchor ID and the sales code, and verifying the real-time nature and accuracy of the mapping information through a cross-platform data interface to support the accurate analysis and verification of subsequent live broadcast sales information.

4. The e-commerce live-streaming sales information parsing method according to claim 1, characterized in that, Step S103 further integrates the anchor's real-time behavior data, product dynamic attributes, and user interaction information, and dynamically verifies and extracts structured anchor information and product information based on the anchor ID and product code mapping relationship for a preset time period.

5. The e-commerce live-streaming sales information parsing method according to claim 1, characterized in that, Step S104 integrates the anchor's real-time behavior data, product dynamic attributes, and user interaction information, and dynamically verifies and splices the structured anchor's sales information based on the anchor ID and product code mapping relationship for a preset time period.

6. The e-commerce live-streaming sales information parsing method according to claim 1, characterized in that, Before step S102, the method further includes: determining whether a configuration request triggered by a user corresponding to the set of goods to be verified by the livestreamer has been received; If so, display the preset configuration interface; Receive configuration information input by the user in the configuration interface; Based on the preset construction strategy, the corresponding configuration verification table is constructed using the configuration information; The set of products to be verified by the livestreamer is stored and processed.

7. The e-commerce live-streaming sales information parsing method according to claim 1, characterized in that, Step S106 further integrates the anchor's real-time behavior data, product dynamic attributes, and user interaction information, and verifies and generates a live-streaming sales analysis report based on the anchor ID and product code mapping relationship for a preset time period.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the e-commerce live-streaming sales information parsing method as described in any one of claims 1 to 7.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the e-commerce live-streaming sales information parsing method as described in any one of claims 1 to 7.

Citation Information

Patent Citations

  • Electronic resource issuing verification method and device, electronic equipment and storage medium

    CN115190328A

  • Real-time data-based live broadcast content risk control processing system and supervision method

    CN117763010A