Live Streaming Recommendation Methods, Devices, Equipment and Computer-Readable Storage Media

Abstract

This application provides a live streaming recommendation method, apparatus, device, and storage medium, relating to the fields of cloud technology and artificial intelligence technology. The method includes: responding to a trigger operation requesting the start of live streaming to initiate a live stream, and acquiring live stream conversion data in real time during the current live stream, as well as acquiring historical live stream conversion data within a preset historical time period; during the current live stream, determining at least one recommended product category based on the live stream conversion data at the current moment and the historical live stream conversion data at a preset predetermined frequency; generating live stream recommendation information based on the at least one recommended product category; and displaying the live stream recommendation information on the live stream interface. This application enables broadcasters to conveniently view optimization suggestions and product recommendations during the live stream, making live stream software more intelligent and convenient, and improving user experience.

Description

Technical Field

[0001] This application relates to the field of Internet technology, and includes, but is not limited to, a live streaming recommendation method, apparatus, device, and computer-readable storage medium. Background Technology

[0002] With the development of the internet and e-commerce, more and more merchants are starting to conduct live-streaming e-commerce sales. Merchants can use live-streaming platforms to increase their sales channels and exposure. However, during each live stream, including before, during, and after the stream, the terminal interface only displays the current number of viewers and the current live stream duration to the host. Or, after the live stream ends, it may display the total transaction amount. However, there is no summary or live stream suggestions regarding the live stream itself or even the live stream performance over a period of time. Therefore, current live-streaming software is not intelligent and convenient enough, and cannot intelligently provide the host with product selection and optimization suggestions during and after the live stream. Summary of the Invention

[0003] This application provides a live streaming recommendation method, apparatus, device, and computer-readable storage medium, relating to the fields of cloud technology and artificial intelligence technology. Based on live streaming conversion data during the current live stream and historical live streaming conversion data, at least one recommended product category is determined. Live streaming recommendation information is generated and displayed in real time during the live stream. Thus, by providing live stream suggestions and reminders to the streamer in real time through live streaming recommendation information, the streamer can conveniently view optimization suggestions and product recommendations during the live stream, making the live streaming software more intelligent and convenient, and improving the user experience.

[0004] The technical solution of this application embodiment is implemented as follows: This application provides a live streaming recommendation method, the method comprising: Responding to the trigger operation of the request to start live streaming to start live streaming, and obtaining live streaming conversion data in real time during the current live streaming process, as well as obtaining historical live streaming conversion data within a preset historical time period; During the current live stream, at least one recommended category is determined according to a preset frequency, based on the live stream conversion data at the current moment and the historical live stream conversion data. Generate live streaming recommendation information based on at least one recommended product category; The live stream recommendation information is displayed on the live stream interface.

[0005] This application embodiment provides a live streaming recommendation device, the device comprising: The response module is used to respond to the trigger operation of the request to start live streaming to start live streaming, and to obtain the live streaming conversion data in real time during the current live streaming process, as well as the historical live streaming conversion data within a preset historical time period. The determination module is used to determine at least one recommended category during the current live broadcast, based on the live broadcast conversion data at the current moment and the historical live broadcast conversion data, according to a preset determined frequency. The generation module is used to generate live streaming recommendation information based on the at least one recommended category; The display module is used to display the live stream recommendation information on the live stream interface.

[0006] This application provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions to implement the aforementioned live streaming recommendation method.

[0007] This application provides a live streaming recommendation device, including: a memory for storing executable instructions; and a processor for executing the executable instructions stored in the memory to implement the above-described live streaming recommendation method.

[0008] This application provides a computer-readable storage medium storing executable instructions, which are used to cause a processor to execute the executable instructions to implement the above-described live streaming recommendation method.

[0009] The embodiments of this application have the following beneficial effects: During the current live broadcast, live broadcast conversion data is acquired in real time, along with historical live broadcast conversion data within a preset historical time period. During the current live broadcast, at least one recommended product category is determined based on the current live broadcast conversion data and historical live broadcast conversion data at a preset frequency. Live broadcast recommendation information is then generated and displayed based on the recommended product category. Thus, because live broadcast recommendation information is generated and displayed in real time based on the live broadcast conversion data and historical live broadcast conversion data, timely live broadcast suggestions and reminders are provided to the broadcaster. This allows the broadcaster to conveniently view optimization suggestions and product recommendations during the live broadcast, making the live broadcast software more intelligent and convenient, and improving the user experience. Attached Figure Description

[0010] Figure 1 This is a schematic diagram illustrating an application scenario of the live streaming recommendation method provided in this application embodiment; Figure 2 This is a schematic diagram of the structure of the live streaming recommendation device provided in the embodiments of this application; Figure 3A This is an optional flowchart illustrating the live streaming recommendation method provided in this application embodiment; Figure 3B This is an optional interface diagram for displaying live stream recommendation information provided in an embodiment of this application; Figure 3C This is an interface diagram of the live streaming interface provided in the embodiments of this application; Figure 4 This is a schematic diagram of the live streaming interface provided in an embodiment of this application; Figure 5 This is an optional flowchart illustrating the live streaming recommendation method provided in this application embodiment; Figure 6 This is an optional flowchart illustrating the live streaming recommendation method provided in this application embodiment; Figure 7 This is an optional flowchart illustrating the live streaming recommendation method provided in this application embodiment; Figure 8 This is a screenshot of the live-streaming e-commerce interface provided in an embodiment of this application; Figure 9 This is a screenshot of the live stream ending as provided in an embodiment of this application; Figure 10 This is an interface diagram illustrating the recommended product categories in the product showcase provided in this embodiment of the application; Figure 11 This is an interface diagram of the product detail page for recommended product categories provided in the embodiments of this application; Figure 12 This is a schematic diagram of the control process of the live streaming recommendation method provided in the embodiments of this application. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0012] In the following description, references to "some embodiments" refer to a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same or different subsets of all possible embodiments and may be combined with each other without conflict. Unless otherwise defined, all technical and scientific terms used in the embodiments of this application have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiments of this application pertain. The terminology used in the embodiments of this application is for the purpose of describing the embodiments of this application only and is not intended to limit the application.

[0013] In the implementation of this application, the collection and processing of relevant data should strictly comply with the requirements of relevant laws and regulations, obtain the informed consent or separate consent of the personal information subject, and carry out subsequent data use and processing within the scope of laws and regulations and the authorization of the personal information subject.

[0014] Before explaining the embodiments of this application, the nouns and key terms involved in this application will be explained first: 1) Product SKU: Stock Keeping Unit (SKU) is a unit for measuring inventory inflows and outflows, such as a piece or box. SKU is now often used as a shorthand for a unique product identification number; each product has a unique SKU number. For a given product, if its brand, model, configuration, grade, unit, production date, shelf life, use, price, place of origin, etc., differ from other products, it can be considered a single item, and this single item has a unique SKU number.

[0015] 2) Traffic-driving products: These are product categories with high exposure and conversion rates. Traffic-driving products represent products with large purchase volumes and high market demand. Regardless of their profit margin, these products can bring cash flow and traffic to the company.

[0016] 3) High-profit products: These are product categories that have very low exposure but extremely high conversion rates. These products are worth trying to increase exposure for and are potential winners among products.

[0017] 4) Future Star Products: These are product categories with low exposure and low conversion rates. If these products have high profit margins, you can try to increase their exposure. If the profit margins are not high, you can consider removing them from the shelves.

[0018] 5) High exposure, low conversion rate products: These are product categories with high exposure but low conversion rate. Regardless of profit margin, these products should be abandoned and the space should be reserved for products with better conversion rates.

[0019] Currently, during each live stream, the host can only see standard live stream data on the live stream interface, such as the current number of viewers and the current live stream duration. They cannot see which products are selling best. Furthermore, to know the sales performance, they need to manually check the sales data of each product in the backend and then manually compare the results. After the live stream ends, the host needs to actively check and review the product data in the backend to select and optimize products for the next live stream. This is time-consuming and labor-intensive for both hosts and merchants, and is neither intuitive nor convenient.

[0020] To address the aforementioned issues with current live streaming software, this application provides a live streaming recommendation method. The system intelligently determines the current recommended product categories during and after the live stream, and generates a live streaming recommendation interface (which can be displayed as a live stream showcase) after the stream ends. This not only allows streamers and merchants to easily monitor the live stream's progress but also facilitates review of the sales data, providing real-time and convenient access to optimization suggestions for the next broadcast and product recommendations for the showcase. The system's intelligent recommendations and data analysis can help streamers and merchants accurately improve sales conversion rates.

[0021] The live streaming recommendation method provided in this application embodiment first starts the live stream in response to a trigger operation requesting the start of the live stream, and acquires the live stream conversion data during the current live stream process in real time, as well as the historical live stream conversion data within a preset historical time period; then, during the current live stream process, at least one recommended product category is determined according to the current live stream conversion data and the historical live stream conversion data at a preset predetermined frequency; live stream recommendation information is generated based on the at least one recommended product category; finally, the live stream recommendation information is displayed on the live stream interface. Thus, because live stream recommendation information is generated and displayed in real time based on the live stream conversion data and historical live stream conversion data during the live stream, timely live stream suggestions and reminders are provided to the streamer through the live stream recommendation information, making it convenient for the streamer to view optimization suggestions and product selection recommendations during the live stream, making the live stream software more intelligent and convenient, and improving the user experience.

[0022] The following describes exemplary applications of the live streaming recommendation device according to embodiments of this application. In one implementation, the live streaming recommendation device provided in this application embodiment can be implemented as any terminal with image display and live streaming functions, such as a laptop, tablet, mobile device (e.g., mobile phone, portable music player, personal digital assistant, dedicated messaging device, portable gaming device), or intelligent robot. In another implementation, the live streaming recommendation device provided in this application embodiment can also be implemented as a server. The following will describe exemplary applications when the live streaming recommendation device is implemented as a terminal.

[0023] See Figure 1 , Figure 1This is a schematic diagram illustrating an application scenario of the live streaming recommendation method provided in this application embodiment. To realize a live streaming sales event, the live streaming recommendation system 10 in this application embodiment includes at least a terminal 100, a network 200, and a system server 300. The terminal 100 receives a user request to start a live stream and responds to the request to start a live stream to begin the live stream. During the current live stream, the terminal 100 collects live stream conversion data in real time and obtains historical live stream conversion data within a preset historical time period, and sends the collected live stream conversion data and historical live stream conversion data to the system server 300. The live stream conversion data can be data obtained by the terminal through a private channel established with the platform server 400 of the platform where the goods sold in the current live stream are located. In some embodiments, the terminal 100 may also only collect live stream conversion data and send it to the system server 300, and the system server 300 obtains historical live stream conversion data from a preset storage unit. After obtaining the live conversion data and historical live conversion data, the system server 300 determines at least one recommended product category based on the live conversion data and historical live conversion data at a preset frequency; generates live recommendation information based on the at least one recommended product category; and sends the generated live recommendation information to the terminal 100 through the network 200. When the terminal 100 receives the live recommendation information, it displays the live recommendation information on the live interface.

[0024] In some embodiments, the live streaming recommendation information can also be generated by the terminal 100 itself, that is, the client on the terminal 100 determines at least one recommended category based on the live streaming conversion data and historical live streaming conversion data; and generates live streaming recommendation information based on at least one recommended category.

[0025] The live streaming recommendation method provided in this application embodiment can also be implemented based on a cloud platform and through cloud technology. For example, the system server 300 mentioned above can be a cloud server, or the platform server 400 can be a cloud server. Alternatively, it can also have a cloud storage to store historical live streaming conversion data and live streaming conversion data collected in real time during the current live streaming process. In this way, during subsequent live streaming processes, data can be directly obtained from the cloud server in real time to generate live streaming recommendation information.

[0026] It's important to note that cloud technology refers to a hosting technology that unifies hardware, software, and network resources within a wide area network (WAN) or local area network (LAN) to achieve data computation, storage, processing, and sharing. Cloud technology is a collective term for network technologies, information technologies, integration technologies, management platform technologies, and application technologies applied in the cloud computing business model. It can form resource pools, providing flexible and convenient on-demand access. Cloud computing technology will become a crucial support. Backend services of technical network systems require substantial computing and storage resources, such as video websites, image websites, and many portal websites. With the rapid development and application of the internet industry, every item may have its own identification mark in the future, requiring transmission to backend systems for logical processing. Data at different levels will be processed separately, and various industry data will require robust system support, which can only be achieved through cloud computing.

[0027] The live streaming recommendation method provided in this application also relates to the field of artificial intelligence (AI) technology. It generates live streaming recommendation information using AI technology, specifically by analyzing live streaming conversion data and historical live streaming conversion data to determine at least one recommended product category. Alternatively, it generates live streaming recommendation information based on at least one recommended product category or generates a live streaming recommendation interface at the end of the live stream, and then optimizes and updates the interface. In some embodiments, an interface generation model can be trained using AI technology to generate the live streaming recommendation interface at the end of the live stream.

[0028] In this embodiment, implementation can be achieved at least through machine learning and computer vision technologies within artificial intelligence. Machine learning (ML) is a multidisciplinary field involving probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory, specifically studying how computers can simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to give computers intelligence; its applications span all areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and instructional learning. Computer vision (CV) is a science that studies how to enable machines to "see." More specifically, it refers to machine vision, which uses cameras and computers to replace human eyes for target recognition, tracking, and measurement, and further performs image processing to create images more suitable for human observation or transmission to instruments for detection. As a scientific discipline, computer vision researches related theories and technologies, attempting to establish artificial intelligence systems capable of extracting information from images or multidimensional data. Computer vision technology typically includes image processing, image recognition, image semantic understanding, image retrieval, OCR, video processing, video semantic understanding, video content / behavior recognition, 3D object reconstruction, 3D technology, virtual reality, augmented reality, simultaneous localization and mapping, and other technologies, as well as common biometric recognition technologies such as face recognition and fingerprint recognition.

[0029] Figure 2 This is a schematic diagram of the structure of the live streaming recommendation device provided in the embodiments of this application. Figure 2 The live streaming recommendation device shown includes at least one processor 310, a memory 350, at least one network interface 320, and a user interface 330. The various components in the live streaming recommendation device are coupled together via a bus system 340. It is understood that the bus system 340 is used to implement communication between these components. In addition to a data bus, the bus system 340 also includes a power bus, a control bus, and a status signal bus. However, for clarity, ... Figure 2 The general labeled all buses as Bus System 340.

[0030] The processor 310 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0031] User interface 330 includes one or more output devices 331 that enable the presentation of media content, including one or more speakers and / or one or more visual displays. User interface 330 also includes one or more input devices 332, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

[0032] Memory 350 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state memory, hard disk drives, optical disk drives, etc. Memory 350 may optionally include one or more storage devices physically located remote from processor 310. Memory 350 may include volatile memory or non-volatile memory, or both. Non-volatile memory may be read-only memory (ROM), and volatile memory may be random access memory (RAM). The memory 350 described in this application embodiment is intended to include any suitable type of memory. In some embodiments, memory 350 is capable of storing data to support various operations, examples of which include programs, modules, and data structures, or subsets or supersets thereof, as illustrated below.

[0033] Operating system 351 includes system programs for handling various basic system services and performing hardware-related tasks, such as the framework layer, core library layer, driver layer, etc., for implementing various basic business functions and handling hardware-based tasks; The network communication module 352 is used to reach other computing devices via one or more (wired or wireless) network interfaces 320, exemplary network interfaces 320 including: Bluetooth, WiFi, and Universal Serial Bus (USB), etc. The input processing module 353 is used to detect and translate one or more user inputs or interactions from one or more input devices 332.

[0034] In some embodiments, the apparatus provided in this application may be implemented in software. Figure 2 A live streaming recommendation device 354 stored in memory 350 is shown. This device 354 can be a live streaming recommendation component within a live streaming recommendation device, and can be software in the form of programs and plugins. It includes the following software modules: a response module 3541, a determination module 3542, a generation module 3543, and a display module 3544. These modules are logically connected and can therefore be arbitrarily combined or further separated according to their implemented functions. The functions of each module will be described below.

[0035] In other embodiments, the apparatus provided in this application can be implemented in hardware. As an example, the apparatus provided in this application can be a processor in the form of a hardware decoding processor, which is programmed to execute the live recommendation method provided in this application. For example, the processor in the form of a hardware decoding processor can be one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.

[0036] The following will describe the live streaming recommendation method provided in this application embodiment, using exemplary applications and implementations of the live streaming recommendation device. The live streaming recommendation device can be any terminal with video input, voice input, and interface display functions, or it can be a server. That is, the live streaming recommendation method in this application embodiment can be executed by a terminal, by a server, or by interaction between a terminal and a server. See also... Figure 3A , Figure 3A This is an optional flowchart illustrating the live streaming recommendation method provided in this application embodiment. The following will combine... Figure 3A The steps shown will be explained. It should be noted that... Figure 3A The live streaming recommendation method in this article is implemented by using the terminal as the execution entity.

[0037] Step S301: In response to the trigger operation of requesting to start live streaming, start live streaming, and obtain live streaming conversion data in real time during the current live streaming process, as well as obtain historical live streaming conversion data within a preset historical time period.

[0038] Here, the user (i.e., the streamer) has a live streaming application installed on their terminal. The streamer can trigger an operation on the live streaming application client to request to start the live stream. When the terminal receives the streamer's trigger operation, it responds to the trigger operation and starts the live stream. During the live stream, the streamer can introduce any product and interact in real time with the viewers (who can be fans who have followed the streamer or non-fans who have not followed the streamer), answering their questions. At the same time, viewers can purchase corresponding products through product links on the live stream interface.

[0039] When users watching a live stream purchase products introduced and recommended by the streamer, corresponding live stream conversion data is generated. This data includes, but is not limited to, at least one of the following: current sales volume, current page views, current conversion rate, current profit margin, current selling price, current sales revenue, and other sales-related data during the current live stream. In this embodiment, live stream conversion data can be acquired in real time, meaning it is acquired during the current live stream. The acquired live stream conversion data is data from the current live stream up to the moment before the current time.

[0040] In this embodiment of the application, historical live streaming conversion data within a preset historical time period can also be obtained. The historical live streaming conversion data includes, but is not limited to, at least one of the following: historical sales volume, historical page views, historical conversion rate, historical profit margin, historical sales price, historical sales amount, and other data related to historical product sales during the historical live streaming process.

[0041] In some embodiments, when this is the first live stream, historical live stream conversion data may not be required; or, when this is the first live stream recommendation of a certain currently recommended object, historical live stream conversion data corresponding to that currently recommended object may not be required. In this embodiment, not only can live stream recommendation information be generated based solely on the live stream conversion data during the current live stream, but it can also be generated based on both the live stream conversion data during the current live stream and historical live stream conversion data within a preset historical time period. When generating live stream recommendation information based solely on the live stream conversion data during the current live stream, the obtained historical live stream conversion data is 0 or empty.

[0042] The preset historical time period can be any length of historical time period set by the user through the live streaming application, or it can be any length of historical time period preset by the system. For example, the preset historical time period can be a historical time period within a week before the current time, or within a month before the current time, etc.

[0043] Step S302: During the current live stream, at least one recommended product category is determined based on the current live stream conversion data and historical live stream conversion data at a preset frequency.

[0044] Here, when both live stream conversion data and historical live stream conversion data are obtained simultaneously, at least one recommended product category is determined based on the current live stream conversion data and historical live stream conversion data. When only live stream conversion data is obtained, at least one recommended product category can be determined based on the current live stream conversion data. Recommended product categories include, but are not limited to, traffic-driving products, high-profit products, future star products, and high-exposure, low-conversion products.

[0045] The preset frequency can be determined based on a certain duration, or it can be reached when the current live broadcast situation meets preset conditions. For example, in this embodiment, the recommended product category is determined according to the preset frequency in the following two ways.

[0046] Method one involves periodically determining at least one recommended product category based on current and historical live stream conversion data. This means filtering live stream conversion data within a defined time interval and then using this filtered data along with historical data to identify at least one recommended product category. For example, a recommended product category can be determined every 10 minutes, based on the live stream conversion data from those 10 minutes and historical data.

[0047] Method two involves real-time collection of the broadcaster's voice information during the live stream, followed by voice analysis to obtain the analysis results. When the voice analysis indicates that live stream recommendations are needed, at least one recommended product category is determined based on the current and historical live stream conversion data. For example, during a live stream, if the broadcaster says, "Which product should I broadcast next?", the voice analysis results can determine that the broadcaster currently wants the live stream recommendation system to provide recommendations. Therefore, based on the current and historical live stream conversion data, at least one recommended product category is determined for live stream recommendations.

[0048] Step S303: Generate live streaming recommendation information based on at least one recommended product category.

[0049] Here, the live stream recommendation information includes, but is not limited to, the name of the product corresponding to the recommended category and the product's link, that is, obtaining the products under the determined recommended category.

[0050] Step S304: Display live stream recommendation information on the live stream interface.

[0051] In this embodiment of the application, when live streaming recommendation information is generated, the live streaming recommendation information is displayed on the current live streaming interface of the streamer in a timely manner to remind the streamer of the live streaming process at the current moment and provide live streaming recommendation suggestions. If the current moment corresponds to the end of the live streaming, product selection recommendations can be made for the streamer in the next live streaming.

[0052] In some embodiments, if the current time corresponds to the end time of the live stream, the system can receive the live stream end operation input by the broadcaster in the live streaming application client. In response to the live stream end operation, it generates annotation information and corresponding identifiers for each recommended category, and generates and displays a live stream recommendation interface based on at least one recommended category. That is, when the broadcaster finishes the live stream, they can request to end the current live stream by performing a live stream end operation on the live streaming application client. When the terminal receives the live stream end operation, it ends the current live stream in response to the operation, and simultaneously displays the generated live stream recommendation interface on the live stream end screen, so that the broadcaster can view the recommended categories and suggestions determined by the system when ending the current live stream.

[0053] In this embodiment, the live stream recommendation interface is used to display the determined recommended categories, so that the streamer can easily view the data during the live stream and the suggested recommended objects selected by the live stream recommendation system based on the data of the live stream. After determining the recommended categories, these recommended categories, annotation information for each recommended category, the current recommended object corresponding to each recommended category in the current live stream, the associated objects of each recommended category (i.e., suggested recommended objects belonging to this recommended category), and the description information of each suggested recommended object can be added to the live stream recommendation interface. The description information of each suggested recommended object includes, but is not limited to, the name, link, ID, unit price, historical sales volume, user reviews, number of favorites, and inventory of the suggested recommended object.

[0054] In some embodiments, the live stream recommendation interface may also display live stream data for the current live stream, including but not limited to: number of viewers, live stream duration, transaction amount, etc. When the current live stream ends, the live stream data for the current live stream is obtained and added to the live stream recommendation interface so that the streamer can view it when the current live stream ends.

[0055] In some embodiments, when a view operation corresponding to the annotation information is received on the live stream recommendation interface, the annotation information of each recommended product category can be displayed on the live stream recommendation interface in response to the view operation.

[0056] In some embodiments, to avoid the impact on the live streamer's live stream by displaying live stream recommendation information for an extended period of time, the live stream recommendation information may be displayed for a preset duration, for example, 10 seconds or 15 seconds; or, the displayed live stream recommendation information may correspond to a close button, and the live stream recommendation information disappears when the streamer clicks the close button.

[0057] In some embodiments, live streaming recommendation information may be displayed in the form of pop-ups or bubbles. If the host clicks on a pop-up or bubble for a product, the live streaming recommendation system will automatically recommend the product's link to the live streaming interface during the current live stream. Alternatively, in some embodiments, if the host clicks on a pop-up or bubble for a product, a pre-recorded explanation video for that product will be played, or an explanation video of the product that the host explained before the current moment and during the current live stream will be played.

[0058] Figure 3B This is an optional interface diagram for displaying live stream recommendation information provided in an embodiment of this application, such as... Figure 3B As shown, once the live stream recommendation information is determined, it can be displayed as a product guide during the current live stream, on the streamer's interface 301. This recommendation information 302 helps the streamer quickly determine which product to discuss next. It should be noted that the recommendation information 302 can be displayed only on the streamer's live stream application client interface, or simultaneously on both the streamer's and viewers' live stream application client interfaces. A close button 303 corresponds to the recommendation information 302, which the streamer can click to stop displaying.

[0059] The live streaming recommendation method provided in this application acquires live streaming conversion data in real time during the current live stream, and also acquires historical live streaming conversion data within a preset historical time period. During the current live stream, at least one recommended product category is determined based on the current live streaming conversion data and historical live streaming conversion data at a preset frequency. Live streaming recommendation information is then generated and displayed based on the recommended product category. Thus, by generating and displaying live streaming recommendation information in real time based on live streaming conversion data and historical live streaming conversion data during the live stream, timely live streaming suggestions and reminders are provided to the streamer. This allows the streamer to easily view optimization suggestions and product recommendations during the live stream, making the live streaming software more intelligent and convenient, and improving the user experience.

[0060] In some embodiments, users watching a live stream can input comments on the live stream interface through a live streaming application client. These comments can be of any type, such as interactions with the streamer, inquiries about products, or requests for the streamer to explain a particular product. In this embodiment, user-inputted comments can be collected in real-time, periodically, or periodically over a given period to obtain each comment within a specific time period. After obtaining the comment information for that specific time period, the comments are first filtered according to preset filtering rules. For example, purely emoji comments or meaningless comments can be filtered out, resulting in filtered comments, which can be plain text. Then, each filtered comment is analyzed to obtain the analysis results. Based on these results, the recommended object type for each comment is determined. For example, if the user input is "down jacket," the recommended object type is determined to be "clothing." Next, the type of recommended object with the highest interaction frequency corresponding to the comment information within the specific time period is determined, and the link of the object to be recommended corresponding to the recommended object type is obtained from the preset recommended object library; finally, the link of the object to be recommended is displayed on the live broadcast interface for a first preset duration, where the first preset duration can be any preset duration, such as 10 seconds or 15 seconds.

[0061] In some embodiments, multiple counters can be preset, each corresponding to a recommended object type. After each link of the object to be recommended is displayed on the live broadcast interface, the counters are reset to zero. Then, in the next specific time period, each time a recommended object type is determined based on a comment, the counter for that recommended object type is incremented by one. After all the comment information in the specific time period has been counted, the recommended object type corresponding to the counter with the highest count result is determined based on the current count results of the multiple counters. This is the recommended object type with the highest interaction frequency.

[0062] In some embodiments, a product search box is also displayed on the live streaming interface. Figure 3C This is a screenshot of the live streaming interface provided in the embodiments of this application, such as... Figure 3C As shown, viewers of the live stream can enter search information in the product search box 311 to find products they want to buy. The search can be conducted from all currently recommended products during the live stream. Figure 3CThe search can be performed within the shopping bag 312 or within a preset product library. After finding the corresponding product, the link to the found product can be displayed on the live stream interface. In this embodiment, the method for searching products includes the following steps: Step S11: Receive search information entered through the product search box.

[0063] Step S12: When a corresponding recommended object is matched in the preset recommended object library based on the search information, or when a corresponding recommended object is matched in all the current recommended objects in the current live broadcast process based on the search information, the link of the matched recommended object is displayed on the live broadcast interface for a second preset duration.

[0064] Here, the preset recommendation object library includes at least one recommendation object. The preset recommendation object library can be a preset product library, which stores multiple products listed by the streamer, or it can store all the products in a store.

[0065] In some embodiments, a pre-broadcast reminder can be given to the broadcaster before responding to the trigger operation, i.e., before starting the live broadcast. The method for giving a pre-broadcast reminder includes the following steps: Step S13: Before responding to the trigger operation, obtain at least one currently recommended object in the current live broadcast process.

[0066] Here, you can access information about multiple products that the streamer will be showcasing in this live broadcast.

[0067] Step S14: Obtain historical live streaming conversion data for each currently recommended object.

[0068] Step S15: Generate current live stream optimization suggestions based on the historical live stream conversion data corresponding to each current recommended object.

[0069] Step S16: Display the suggestion information corresponding to the current live streaming optimization suggestion on the interface before starting the live stream.

[0070] Here, current live streaming optimization suggestions can be made by identifying historical live streaming information such as lead-driving products, high-profit products, future star products, and high-exposure, low-conversion products based on historical live streaming conversion data. Then, based on the historical live streaming information, determine which products should be mainly broadcast in this live stream, how long each product should be allocated for playback, the order of products in the shopping bag, the label information added to each product in the shopping bag, and the order in which the products are explained in this live stream.

[0071] In some embodiments, a live streaming recommendation system for implementing the live streaming recommendation method of the present application includes at least: a terminal, a system server, and a platform server. The terminal has a live streaming application installed, allowing users to start and end live streaming via the application's client and sell goods during the live stream. The system server is either the server in the live streaming recommendation system or the server of the live streaming application. It generates live streaming recommendation information in real-time during the live stream or generates the final live streaming recommendation interface at the end of the live stream. It then sends the live streaming recommendation information to the client on the terminal during the live stream or sends the live streaming recommendation interface to the client at the end of the live stream, so that the live streaming recommendation information is displayed on the terminal in real-time or the live streaming recommendation interface is displayed at the end of the live stream. When the live streaming application does not have a product sales function, i.e., it does not have a shopping platform function, the live streaming recommendation system includes a platform server. This platform server can be the server of a third-party shopping platform, used to collect user purchase records on that shopping platform. When the live streaming application has a product sales function, i.e., it has a shopping platform function, the platform server is the aforementioned system server, meaning the live streaming recommendation system may not need to include a platform server.

[0072] For example, in one scenario, a streamer can broadcast live on a live streaming application A and sell products during the broadcast. However, live streaming application A is not a shopping platform and does not have shopping functionality. Therefore, server A1 of live streaming application A can interact with server B1 of shopping platform B to authorize the display of product links from shopping platform B within live streaming application A. This allows users of live streaming application A to be redirected to shopping platform B to purchase products while using live streaming application A. When displaying product link C from shopping platform B within live streaming application A, it could be that a streamer displays product link C on their live streaming interface during the broadcast, and the streamer's viewers can purchase products through product link C while watching the broadcast.

[0073] In another scenario, a streamer can broadcast live on a live streaming application A and sell products during the broadcast. This application not only has live streaming functionality but also shopping capabilities; that is, application A is also a shopping platform, providing entry points for purchasing products, i.e., product links. In this way, users of application A can simultaneously purchase products while using the live streaming function. Specifically, when displaying a product link C in application A, a streamer could directly display the product link C on their live stream interface, allowing their viewers to purchase the product through link C while watching the broadcast.

[0074] It's important to note that when a streamer recommends products during a live stream, they can either add the product link themselves to their own live stream interface, or other merchants can add the product link. In other words, a streamer can have their own shop within the live streaming application A, and can sell products from that shop during the live stream. Therefore, the streamer can push links through the client of live streaming application A to add product links from their shop to their own live stream interface. Alternatively, a streamer may not have their own shop within live streaming application A. During the live stream, the streamer can send a sales request to another user D (i.e., the shop owner) within live streaming application A through the client of live streaming application A. Upon receiving the sales request, user D responds through the client of live streaming application A, authorizing the streamer to add the product link to the product to be sold, thus adding the product link to the streamer's live stream interface and selling products from user D's shop during the live stream.

[0075] like Figure 4 The diagram shown is a schematic of a live streaming interface provided in an embodiment of this application. The current interface 401 of the live streaming application not only displays the live streaming content 402, but also displays a shopping bag 403. The shopping bag 403 contains at least one product link, and these product links can be product links from third-party shopping platforms or product links from the live streaming application. Furthermore, the product links can be product links recommended by the streamer through a push link operation on the client, or product links attached after the streamer sends a sales request to other users in the live streaming application on the client and the other users respond to the sales request on the client.

[0076] The following will use a live streaming recommendation system, including a terminal, a system server, and a platform server, as an example to illustrate the live streaming recommendation method of this application. Figure 5 This is an optional flowchart illustrating the live streaming recommendation method provided in this application embodiment, such as... Figure 5 As shown, the method includes the following steps: Step S501: The terminal receives a user request to start the live broadcast.

[0077] Here, a live streaming application is installed on the terminal. Users (i.e., broadcasters) can click the "Start Live Stream" button on the live streaming application's client to request to start a live stream.

[0078] Step S502: The terminal generates a live broadcast start request based on the trigger operation.

[0079] In step S503, the terminal sends a live broadcast start request to the system server.

[0080] In step S504, the system server responds to the live stream start request to begin the live stream.

[0081] Step S505: During the current live broadcast, the system server sends a data retrieval request to the platform server.

[0082] In this embodiment, if the currently recommended product during the live stream is a product on the platform where the live stream application is located, the system server and the platform server can be the same server, and the system server can directly obtain the live stream conversion data. If the currently recommended product during the live stream is not a product on the platform where the live stream application is located, that is, if the currently recommended product during the live stream is a product on another third-party shopping platform, the platform server is the server corresponding to that third-party shopping platform, and the system server sends a data acquisition request to that platform server to request the live stream conversion data corresponding to the purchase of products on that third-party shopping platform by users through watching the live stream during the current live stream.

[0083] Step S506: The system server obtains the live streaming conversion data sent by the platform server during the current live streaming process in real time.

[0084] In some embodiments, the currently recommended items during the current live stream are products from other third-party shopping platforms. The system server can establish a private channel with the platform server for transmitting data, send a data acquisition request to the platform server through the private channel, and acquire the live stream conversion data returned by the platform server during the current live stream.

[0085] Step S507: The system server obtains historical live streaming conversion data within a preset historical time period.

[0086] Here, historical live stream conversion data can be stored in a preset storage unit. After each live stream ends, the system server stores the live stream conversion data of that live stream in this storage unit, so that subsequent live streams can directly retrieve historical live stream conversion data from this storage unit. For subsequent live streams, the live stream conversion data of this live stream will also become historical live stream conversion data and be retrieved from the storage unit.

[0087] In some embodiments, the preset storage unit may be a cloud storage device, that is, historical live streaming conversion data may be stored in a cloud storage device.

[0088] In step S508, the system server determines at least one recommended product category based on the live stream conversion data and historical live stream conversion data.

[0089] Step S509: The terminal receives the user's live stream end operation.

[0090] In step S510, the terminal responds to the live stream end operation by generating a live stream end request.

[0091] In step S511, the terminal sends a live stream end request to the system server.

[0092] In step S512, the system server generates a live streaming recommendation interface based on at least one recommended product category.

[0093] In step S513, the system server responds to the live stream end request and ends the current live stream.

[0094] In step S514, the system server sends the live streaming recommendation interface to the terminal.

[0095] Step S515: The terminal displays the live streaming recommendation interface on the current screen.

[0096] In this embodiment, the entire live streaming process is achieved through interaction between the terminal, system server, and platform server. This includes the collection of live streaming conversion data during the live stream and the generation of a live streaming recommendation interface based on the collected data and historical conversion data. This allows for easier viewing of optimization suggestions and product recommendations at the end of the live stream, making the live streaming software more intelligent and convenient, and improving user experience. Simultaneously, the system server and platform server interact through a private channel to collect data, ensuring the accuracy of the collected data. This results in more accurate product categories recommended in the generated live streaming recommendation interface, providing more precise optimization suggestions and product recommendations for subsequent live streams.

[0097] In some embodiments, there is at least one currently recommended object during the current live broadcast. The currently recommended object can be any object that can be traded or purchased, such as goods or services. During the current live broadcast, the host can recommend the currently recommended object to the users watching the live broadcast.

[0098] based on Figure 3A , Figure 6 This is an optional flowchart illustrating the live streaming recommendation method provided in this application embodiment, such as... Figure 6 As shown, the real-time acquisition of live conversion data during the current live stream in step S301 above can be achieved through the following steps: Step S601: Real-time statistics are collected on the current pageviews, sales volume, and profit margin for each currently recommended object during the current live stream.

[0099] Here, the data at the current moment can be counted, or when the current moment is the end of the live stream, the data for the entire current live stream process can be counted to obtain the current pageviews, current sales, and current profit margin for each currently recommended object. In some embodiments, step S601 can be implemented through the following steps: Step S6011: During the current live broadcast, for each currently recommended object, the terminal establishes a private channel with the platform server of the platform to which the currently recommended object belongs.

[0100] Here, when a streamer recommends a specific item during a live broadcast, the terminal can request the system server to establish a private channel with the platform server of the platform to which the recommended item belongs. This private channel refers to the data transmission channel between the system server and the platform server. During the establishment of the private channel, the system server and the platform server can perform a handshake; if the handshake is successful, the private channel is established successfully. Through this private channel, the terminal can request the system server to send data to the platform server, and the platform server can also request data from the terminal. In other words, the system server and the platform server can engage in bidirectional data interaction, enabling two-way data transmission.

[0101] Step S6012: Periodically obtain the transaction data of the currently recommended object through a private channel.

[0102] Once the private channel between the system server and the platform server is successfully established, the terminal can request the system server to send a subscription request message to the platform server. The subscription request message is used to subscribe to functions on the platform server in order to obtain the data corresponding to those functions.

[0103] In some embodiments, a terminal may request the system server to periodically obtain transaction data of the currently recommended object through a private channel. Transaction data includes, but is not limited to, at least one of the following: product pageviews, product selling price, sales volume, sales revenue, conversion rate, profit margin, etc. In other embodiments, the system server may also, in response to each data acquisition request received from the terminal, obtain the transaction data of the currently recommended object through a private channel.

[0104] Step S6013: Sum the transaction data of the current live broadcast process up to the current moment to obtain the total transaction data of the current recommended object.

[0105] Here, transaction data includes at least sales volume, browsing time, and pageviews.

[0106] Regarding sales volume, if a product is sold during the live stream, the sales revenue is positive; if a return occurs during the live stream, the sales revenue is negative, and so on. The sales revenue for the product during the current live stream up to the current moment or during the entire current live stream is recorded, and the current sales volume is calculated based on the sales revenue. If it is for the entire current live stream, the sales revenue obtained during the entire current live stream can be summed at the end of the current live stream to obtain the current sales volume in the total transaction data of the currently recommended object.

[0107] Regarding pageviews, if a user spends more than 0 time viewing a particular product within a given period, that viewing time is recorded; if a user spends no time viewing the product within a given period, that viewing time is recorded as 0, and so on, recording the total viewing time for the product throughout the current live stream. At the end of the current live stream, the total viewing time for the user throughout the entire live stream can be summed to obtain the total viewing time in the total transaction data of the currently recommended product.

[0108] Regarding pageviews, if a first number of users viewed a product within a certain period, that first number is recorded; if a second number of users viewed the product within another period, that second number is recorded, and so on, recording the total number of people who viewed the product during the entire current live stream. At the end of the current live stream, the total number of pageviews for the recommended product can be obtained by summing the total number of users who viewed the product during the entire live stream.

[0109] Step S6014: Based on the total transaction data, determine the current pageviews, current sales volume, and current profit margin of the currently recommended object.

[0110] Here, after determining the total transaction data, the current pageviews, current sales volume, and current profit margin of the currently recommended object are calculated based on the total transaction data.

[0111] Step S602: The current pageviews, current sales, and current profit margin of each currently recommended object are determined as live streaming conversion data.

[0112] Correspondingly, obtaining historical live streaming conversion data within a preset historical time period in step S301 above can be achieved through the following steps: Step S603: Calculate the historical pageviews, historical sales, and historical profit margins for each currently recommended object during the historical live streaming process within the preset historical time period.

[0113] Step S604: The historical pageviews, historical sales, and historical profit margins of each currently recommended object are determined as historical live streaming conversion data.

[0114] In this embodiment of the application, data for each current recommended object can be obtained from a preset storage unit, and then historical live streaming conversion data can be calculated.

[0115] In some embodiments, the recommended product categories may include at least one of the following: traffic-driving products, high-profit products, future star products, and high-exposure, low-conversion products.

[0116] In some embodiments, when the recommended product category includes lead generation products, step S302 above can be implemented through the following steps: Step S605: For each currently recommended object, at the current moment, determine the current conversion rate of the currently recommended object based on its current sales volume and current pageviews.

[0117] Here, the current conversion rate refers to the ratio between current sales volume and current page views.

[0118] Step S606: Determine the historical conversion rate of the current recommended object based on its historical sales volume and historical pageviews.

[0119] Here, historical conversion rate refers to the ratio between historical sales volume and historical pageviews.

[0120] Step S607: Sort all currently recommended objects in the current live broadcast process according to the current number of views, historical number of views, current conversion rate and historical conversion rate to form the first recommended object sequence.

[0121] Here, the current pageviews, historical pageviews, current conversion rate, and historical conversion rate can be combined to perform a comprehensive ranking of all currently recommended objects in the current live stream. The comprehensive ranking can be achieved by weighted summing of the current pageviews, historical pageviews, current conversion rate, and historical conversion rate of each currently recommended object in the current live stream, and then ranking all currently recommended objects based on the weighted summation result to form the first recommended object sequence.

[0122] Step S608: Select the first target recommendation object according to the first recommendation object sequence.

[0123] Here, if the first recommended object sequence is formed by sorting the weighted sum results from largest to smallest or from highest to lowest, then the first recommended object in the first recommended object sequence can be determined as the first target recommended object.

[0124] Step S609: The first target recommendation object is determined as the diversion product at the current moment.

[0125] In some embodiments, when the recommended product category includes high-profit products, step S302 above can be implemented through the following steps: Step S610: Sort all currently recommended objects in the current live broadcast process according to the current profit margin and historical profit margin of each currently recommended object to form a second recommended object sequence.

[0126] Here, a comprehensive ranking of all currently recommended objects in the current live broadcast process can be performed by combining the current profit margin and historical profit margin. This comprehensive ranking can be achieved by weighted summing of the current and historical profit margins of each currently recommended object in the current live broadcast process, and then ranking all currently recommended objects based on the weighted summation result to form a second recommended object sequence. It should be noted that the weights in the weighted summation in step S610 are different from those in step S607 mentioned above.

[0127] Step S611: Select the second target recommendation object according to the second recommendation object sequence.

[0128] Here, if the second recommendation object sequence is formed by sorting the weighted sum results from largest to smallest or from highest to lowest, then the first recommendation object in the second recommendation object sequence can be determined as the second target recommendation object.

[0129] Step S612: The second target recommendation object is determined as the high-profit product at the current moment.

[0130] In some embodiments, when the recommended product category includes future star products, step S302 above can be implemented through the following steps: Step S613: For each currently recommended object, at the current moment, determine the current conversion rate of the currently recommended object based on its current sales volume and current pageviews.

[0131] Here, the current conversion rate refers to the ratio between current sales volume and current page views.

[0132] Step S614: Determine the historical conversion rate of the current recommended object based on its historical sales volume and historical pageviews.

[0133] Here, historical conversion rate refers to the ratio between historical sales volume and historical pageviews.

[0134] Step S615: Based on the current conversion rate and historical conversion rate, sort all currently recommended objects in the current live broadcast process to form a third recommended object sequence.

[0135] Here, the current conversion rate and historical conversion rate can be combined to perform a comprehensive ranking of all currently recommended objects during the current live stream. This comprehensive ranking can be achieved by weighted summing of the current and historical conversion rates for each currently recommended object during the current live stream, and then ranking all currently recommended objects based on the weighted summation result to form a third recommended object sequence. It should be noted that the weights in the weighted summation in step S615 are different from the weights in the weighted summation in step S610 and step S607.

[0136] Step S616: Select the third target recommendation object based on the third recommendation object sequence.

[0137] Here, if the third recommendation object sequence is formed by sorting the weighted sum results from largest to smallest or from highest to lowest, then the first recommendation object in the third recommendation object sequence can be determined as the third target recommendation object.

[0138] Step S617: The third target recommendation object is determined as the future star product at the current moment.

[0139] In some embodiments, when the recommended category includes high-exposure, low-conversion varieties, step S302 above can be achieved through the following steps: Step S618: For each current recommended object, obtain the recommendation duration for the current recommended object during the current live broadcast.

[0140] Step S619: For each currently recommended object, at the current moment, determine the current conversion rate of the currently recommended object based on its current sales volume and current pageviews.

[0141] Here, the current conversion rate refers to the ratio between current sales volume and current page views.

[0142] Step S620: Sort all currently recommended objects in the current live broadcast process according to the recommendation duration and the current conversion rate to form the fourth recommendation object sequence.

[0143] Here, the recommendation duration and current conversion rate can be combined to comprehensively rank all currently recommended objects during the current live stream. This comprehensive ranking can be achieved by weighted summing of the recommendation duration and current conversion rate for each currently recommended object during the current live stream, and then ranking all currently recommended objects based on the weighted summation result, forming a fourth recommendation object sequence. It should be noted that the weights in the weighted summation in step S620 are different from those in steps S615, S610, and S607.

[0144] Step S621: Select the fourth target recommendation object based on the fourth recommendation object sequence.

[0145] Step S622: The fourth target recommendation object is determined as the high-exposure, low-transfer product at the current moment.

[0146] Figure 7 This is an optional flowchart illustrating the live streaming recommendation method provided in this application embodiment, such as... Figure 7 As shown, the method includes the following steps: Step S701: In response to the trigger operation of requesting to start live streaming, start live streaming, obtain live streaming conversion data during the current live streaming process, and obtain historical live streaming conversion data within a preset historical time period.

[0147] Step S702: Based on the live stream conversion data and historical live stream conversion data, determine at least one recommended product category.

[0148] Step S703: Generate annotation information and corresponding identifiers for each recommended product category.

[0149] Here, the annotation information for each recommended category refers to the explanation of that category. This annotation information allows users to more clearly understand the product type and information corresponding to each recommended category when viewing the recommended categories displayed on the live stream recommendation interface. The corresponding icon is an operation icon for clicking to view the annotation information. When a user clicks the icon corresponding to the annotation information, the annotation information for each recommended category will be displayed on the current interface.

[0150] Step S704: Determine the description information of each recommended category, the suggested recommended object corresponding to each recommended category, and the Uniform Resource Locator (URL) of each suggested recommended object.

[0151] Here, the suggested recommended objects may or may not include the currently recommended objects during this live stream. Suggested recommended objects can be other objects corresponding to the recommended product category and having the same function, attributes, or similar performance as a currently recommended object during this live stream. Suggested recommended objects can also be objects searched or filtered from third-party shopping platforms, or objects searched or filtered from all products corresponding to the live stream application.

[0152] Step S705: Add the description information, suggested recommended objects, and URL to the live streaming recommendation interface.

[0153] Step S706: Add the live stream conversion data corresponding to each recommended product category to the live stream recommendation interface.

[0154] In this embodiment of the application, the collected live streaming conversion data can also be displayed on the live streaming recommendation interface to provide users with more comprehensive data on the live streaming session, making it easier for users to conduct data analysis and determine a more suitable product selection plan.

[0155] In step S707, in response to the live stream end operation, the current live stream ends, and the live stream recommendation interface and the live stream conversion data corresponding to each recommended category are displayed.

[0156] In step S708, in response to the viewing operation of the identifier corresponding to the annotation information, the annotation information of each recommended category is displayed on the live recommendation interface.

[0157] In this embodiment, the live stream recommendation interface not only displays the recommended product categories intelligently determined by the system, but also annotation information for each recommended product category and the collected live stream conversion data. This allows for a more comprehensive presentation of relevant data from the live stream to the user, enabling them to conduct independent analysis based on the data displayed on the live stream recommendation interface after the live stream ends, building upon the product selection suggestions provided by the system to determine a better product selection strategy. This, in turn, maximizes the conversion rate of products in subsequent live streams.

[0158] The following will describe an exemplary application of the embodiments of this application in a real-world application scenario.

[0159] This application provides a live streaming recommendation method. After the live stream ends, the system intelligently generates recommended product categories (such as lead-generation products, high-profit products, and future star products) for the live stream showcase (i.e., the live stream recommendation interface). The product selection and recommendation categories in the live stream showcase are generated based on the core key conversion data of the live stream's product SKUs (e.g., page views, conversion rate, profit margin, selling price, sales revenue, and sales volume). This application not only allows broadcasters and merchants to easily review the live stream sales data but also provides real-time and convenient access to optimization suggestions for the next broadcast and product recommendations for the showcase. The system's intelligent recommendations and data analysis can help broadcasters and merchants accurately improve sales conversion rates.

[0160] From a product perspective, the embodiments of this application are mainly applied to e-commerce live streaming scenarios. They can help live streamers and merchants conduct timely reviews and conveniently view optimization suggestions for the next live stream and product recommendations in the product showcase.

[0161] Figure 8 This is a diagram of the live-streaming e-commerce interface provided in this application embodiment. In the lower left corner of the e-commerce interface 801, there is a product purchase entry 802 and a link 803 for the product currently being explained. Figure 9This is an interface diagram showing the end of a live stream according to an embodiment of this application. When the live stream ends, page 901 displays regular live stream data 902, such as number of viewers, live stream duration, and transaction amount, and also generates a showcase recommendation category 903 for this live stream. The showcase recommendation category data is generated based on the core key conversion data of the product SKUs in this live stream (e.g., page views, conversion rate, profit margin, selling price, sales amount, and sales volume). Figure 9 There are four product categories featured in this live stream: traffic-driving products, high-profit products, future star products, and high-exposure, low-conversion products. One type of product—high-exposure, low-conversion—will not appear in the product showcase.

[0162] Figure 10 This is an interface diagram of the product category description in the showcase provided in the embodiments of this application, wherein, Figure 10 Is it a click? Figure 9 The interface that appears after question mark 904 following category 903 in the featured product showcase. Figure 10 The product showcase's recommended product category descriptions in the interface will explain and describe each recommended category. Because there are many categories of retail goods, each category should have a clear positioning, and different positioning categories should have different operational strategies. Based on the product's performance in three dimensions—exposure, conversion rate, and profit margin—the recommended product categories can be divided into four types: 1) Traffic-driving product categories (i.e., traffic-driving products): These are product categories with high exposure and high conversion rates. Traffic-driving products represent products with large purchase volumes and high market demand. Regardless of their profit margin, these products can bring cash flow and traffic to the company.

[0163] 2) Future Star Category (i.e. Future Star Product): This refers to product categories with very low exposure but extremely high conversion rates. These products are worth trying to increase exposure and are potential stars among products.

[0164] 3) High-profit categories (i.e., high-margin products): These are product categories with low exposure and low conversion rates. If the profit margin of these products is high, you can try to increase their exposure. If the profit margin is not high, you can consider removing them from the shelves.

[0165] 4) High exposure, low conversion rate product category (i.e., high exposure, low conversion rate product): refers to product categories with high exposure but low conversion rate. These products should be abandoned regardless of profit margin, and the space should be reserved for products with better conversion rates.

[0166] Figure 11 This is an interface diagram of the product detail page for recommended product categories provided in an embodiment of this application, wherein, Figure 11 Is it a click? Figure 9 The interface that appears when you click "View More 905" will then redirect you to... Figure 11The showcase recommendation category details page interface displays the star product 1102 of this live stream at the top; below the showcase recommendation category details page 1101, based on the performance of this live stream, the showcase recommendation category 1103 for the next live stream will be generated, and corresponding recommendation text 1104 will be provided for the host to review.

[0167] The key technologies involved in this application include intelligent product showcase generation and big data. After the live stream ends, the system intelligently generates recommended product categories for the live stream showcase. This not only facilitates review of the live stream sales data for both the host and the merchant, but also allows for real-time and convenient viewing of optimization suggestions and product recommendations for the next live stream. The system's intelligent recommendations and data analysis can help hosts and merchants accurately improve sales conversion rates.

[0168] Figure 12 This is a schematic diagram of the control process of the live streaming recommendation method provided in the embodiments of this application, such as... Figure 12 As shown, it includes the following steps: Step S121: The host starts the live stream, and users watch the live stream and purchase products.

[0169] When a streamer starts a live broadcast and selects a product to recommend, the backend server (i.e., the system server) returns the data of the current product to the client. The client displays the product recommended and explained by the streamer in a bubble in the lower left corner. Users watch the live broadcast and, if they find a product they like, they can click to buy it.

[0170] In step S122, the backend server records the transaction data of the goods in real time.

[0171] During the live stream, the client requests data from the backend server once per second. The backend server then requests transaction data from the third-party e-commerce platform or coordinates with its own server and sends it back to the client. Because data from the third-party e-commerce platform is involved, a WebSocket request method can be used. In the WebSocket method, the client and the third-party e-commerce platform's server first perform a handshake; a successful handshake establishes a connection, which can be understood as establishing a private channel. WebSockets are full-duplex; the client can send messages to the server, and the server can also send messages to the client. The requested transaction data includes product data IDs, product views, product selling price, sales volume, sales revenue, conversion rate, and profit margin for this live stream.

[0172] After establishing a connection, the client sends several messages to the backend server, informing it which data it wants to subscribe to. The backend server then continuously pushes new data to the client. During this data push, the backend server may send a ping request to the client; in response, the client must return a pong message to the backend server to check the connection's integrity (i.e., perform a heartbeat check).

[0173] The steps for requesting data using WebSocket are: establish a connection -> send messages to subscribe to data -> continuously receive data -> periodically respond with heartbeat checks.

[0174] Once the backend server detects data generation, it immediately appends the new data to the same target object (Object, referring to a polymorphic form where method parameters are uncertain, meaning the method can accept multiple parameters). The client only needs to periodically retrieve the length of the target object and compare it with the previously read length. If new data is found to be readable, i.e., the length of the target object has changed, then a read operation is triggered to retrieve the newly uploaded data.

[0175] Step S123: After the live stream ends, the backend server determines the recommended product categories for the product showcase based on the transaction data and sends the recommended product categories back to the client to generate the featured products recommended by the live stream stars.

[0176] At the end of the live stream, the client will receive data from the backend server on the featured products recommended during the live stream. This data includes product ID, product image, product views, product price, sales volume, sales revenue, conversion rate, and profit margin. There are four recommended product categories: traffic-driving products, high-profit products, future star products, and high-exposure, low-conversion products. High-exposure, low-conversion products will not appear on the live stream recommendation page. The backend server generates the recommended product categories based on all product data from the live stream.

[0177] Here, "leading products" refers to products whose sales data for all items promoted in this live stream are ranked based on a combination of page views and conversion rates. The product with the highest ranking is then identified as the leading product. Conversion rate = sales volume / page views, where " / " means dividing by.

[0178] High-profit items: This refers to the items whose profit margins are ranked based on the data of all the products sold in this live stream, and the items with the highest ranking are identified as high-profit items.

[0179] Future Star Products: This refers to the products that are ranked based on their conversion rates during the live stream, and the products with the highest rankings are identified as future star products.

[0180] High-exposure, low-conversion products: This refers to ranking all the products promoted in this live stream based on a combination of page views and conversion rates, and then identifying the products with the lowest ranking as high-exposure, low-conversion products.

[0181] In step S124, the backend server intelligently generates recommended product categories for the next live stream based on big data and the sales data of the current session.

[0182] The backend server analyzes product data and historical sales data to generate recommended product categories. When the server detects which product is driving traffic to the current live stream, it searches the entire network for products with similar attributes and recommends them. These attributes include: product category, selling price, profit margin, and sales volume. The backend server generates different recommendation keywords for the next live stream's product showcase based on different categories. When the host clicks "view more," the client requests data from the backend server, which returns the corresponding recommended categories to the host, helping them to better understand the sales performance of the current live stream and provide product selection references for the next live stream.

[0183] In this embodiment, the system intelligently generates recommended product categories (i.e., traffic-driving products, high-profit products, and future star products) for the live stream showcase after the live stream ends. The recommended product categories in the showcase are generated based on the core key conversion data of the live stream's product SKUs (i.e., page views, conversion rate, profit margin, selling price, sales revenue, and sales volume). This embodiment not only facilitates review of the live stream sales data for both the streamer and the merchant, but also allows for convenient real-time viewing of optimization suggestions and product recommendations for the next live stream. The system's intelligent recommendations and data analysis can help streamers and merchants accurately improve sales conversion rates.

[0184] The following continues to describe the exemplary structure of the live streaming recommendation device 354 provided in the embodiments of this application as a software module. In some embodiments, such as Figure 2 As shown, the software module stored in the live streaming recommendation device 354 in the memory 350 can be the live streaming recommendation device in the live streaming recommendation equipment, the device including: The response module 3541 is used to respond to a trigger operation requesting the start of live streaming to start the live stream, and to acquire live stream conversion data in real time during the current live stream process, as well as historical live stream conversion data within a preset historical time period; the determination module 3542 is used to determine at least one recommended category during the current live stream process, according to a preset determined frequency, based on the live stream conversion data at the current moment and the historical live stream conversion data; the generation module 3543 is used to generate live stream recommendation information based on the at least one recommended category; and the display module 3544 is used to display the live stream recommendation information on the live stream interface.

[0185] In some embodiments, the determining module is further configured to: periodically determine at least one recommended category based on the live streaming conversion data at the current moment and the historical live streaming conversion data; or, perform voice analysis on the voice information collected in real time to obtain voice analysis results; and when the voice analysis results indicate that live streaming recommendations are needed, determine at least one recommended category based on the live streaming conversion data at the current moment and the historical live streaming conversion data.

[0186] In some embodiments, the device further includes: a text analysis module, configured to perform text analysis on each comment message acquired within a specific time period during the live broadcast to obtain a text analysis result; a recommendation object type determination module, configured to determine the recommendation object type with the highest interaction frequency within the specific time period based on the text analysis result of each comment message; a link acquisition module, configured to acquire a link of a to-be-recommended object corresponding to the recommendation object type from a preset recommendation object library; and a first link display module, configured to display the link of the to-be-recommended object on the live broadcast interface for a first preset duration.

[0187] In some embodiments, a product search box is also displayed on the live streaming interface; the device further includes: a receiving module, configured to receive search information input through the product search box; and a second link display module, configured to display the link of the matched recommended object on the live streaming interface for a second preset duration when a corresponding recommended object is matched in a preset recommended object library based on the search information, or when a corresponding recommended object is matched among all currently recommended objects in the current live streaming process based on the search information.

[0188] In some embodiments, the apparatus further includes: a current recommended object determination module, configured to determine at least one current recommended object in the current live streaming process; the response module is further configured to: statistically analyze in real time the current pageviews, current sales, and current profit margin for each current recommended object in the current live streaming process at the current moment; determine the current pageviews, current sales, and current profit margin for each current recommended object as the live streaming conversion data at the current moment; statistically analyze the historical pageviews, historical sales, and historical profit margin for each current recommended object in historical live streaming processes within the preset historical time period; and determine the historical pageviews, historical sales, and historical profit margin for each current recommended object as the historical live streaming conversion data.

[0189] In some embodiments, the recommended product category includes traffic-driving products; the determining module is further configured to: for each current recommended object, at the current time, determine the current conversion rate of the current recommended object based on the current sales volume and the current pageviews of the current recommended object; determine the historical conversion rate of the current recommended object based on the historical sales volume and the historical pageviews of the current recommended object; sort all current recommended objects in the current live broadcast process based on the current pageviews, the historical pageviews, the current conversion rate, and the historical conversion rate to form a first recommended object sequence; select a first target recommended object based on the first recommended object sequence; and determine the first target recommended object as the traffic-driving product at the current time.

[0190] In some embodiments, the recommended product category includes high-profit products; the determining module is further configured to: sort all current recommended products in the current live broadcast process according to the current profit rate and the historical profit rate of each current recommended product to form a second recommended product sequence; select a second target recommended product according to the second recommended product sequence; and determine the second target recommended product as the high-profit product at the current moment.

[0191] In some embodiments, the recommended product category includes future star products; the determining module is further configured to: for each current recommended object, at the current moment, determine the current conversion rate of the current recommended object based on the current sales volume and the current pageviews of the current recommended object; determine the historical conversion rate of the current recommended object based on the historical sales volume and the historical pageviews of the current recommended object; sort all current recommended objects in the current live broadcast process based on the current conversion rate and the historical conversion rate to form a third recommended object sequence; select a third target recommended object based on the third recommended object sequence; and determine the third target recommended object as the future star product at the current moment.

[0192] In some embodiments, the recommended product category includes high-exposure, low-conversion products; the determining module is further configured to: for each current recommended object, obtain the recommendation duration for the current recommended object during the current live stream; for each current recommended object, at the current moment, determine the current conversion rate of the current recommended object based on the current sales volume and the current pageviews of the current recommended object; sort all current recommended objects during the current live stream based on the recommendation duration and the current conversion rate to form a fourth recommended object sequence; select a fourth target recommended object based on the fourth recommended object sequence; and determine the fourth target recommended object as the high-exposure, low-conversion product at the current moment.

[0193] In some embodiments, the apparatus further includes: a first generation module, configured to generate annotation information and an identifier corresponding to the annotation information for each of the recommended product categories in response to a live broadcast end operation; a second generation module, configured to generate a live broadcast recommendation interface based on the at least one recommended product category; an interface display module, configured to display the live broadcast recommendation interface; and an annotation information display module, configured to display the annotation information for each of the recommended product categories on the live broadcast recommendation interface in response to a viewing operation input on the live broadcast recommendation interface for the identifier corresponding to the annotation information.

[0194] In some embodiments, the apparatus further includes: a current recommended object acquisition module, configured to acquire at least one current recommended object uploaded during the current live stream before responding to the trigger operation; a historical live stream conversion data acquisition module, configured to acquire the historical live stream conversion data corresponding to each current recommended object; a third generation module, configured to generate current live stream optimization suggestions based on the historical live stream conversion data corresponding to each current recommended object; and a suggestion information display module, configured to display the suggestion information corresponding to the current live stream optimization suggestions on the interface before starting the live stream.

[0195] It should be noted that the description of the apparatus in this application embodiment is similar to the description of the method embodiment described above, and has similar beneficial effects as the method embodiment; therefore, it will not be repeated. For technical details not disclosed in this apparatus embodiment, please refer to the description of the method embodiment of this application for understanding.

[0196] This application provides a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method described in this application.

[0197] This application provides a storage medium storing executable instructions. When these executable instructions are executed by a processor, they cause the processor to perform the method provided in this application, for example... Figure 3A The method shown.

[0198] In some embodiments, the storage medium may be a computer-readable storage medium, such as a ferromagnetic random access memory (FRAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), flash memory, magnetic surface memory, optical disc, or a compact disk-read-only memory (CD-ROM); or it may be a device that includes one or any combination of the above-mentioned memories.

[0199] In some embodiments, executable instructions may take the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a standalone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0200] As an example, executable instructions may, but do not necessarily, correspond to files in a file system. They may be stored as part of a file that holds other programs or data, for example, in one or more scripts within a Hyper Text Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple co-located files (e.g., files storing one or more modules, subroutines, or code sections). As an example, executable instructions may be deployed to execute on a single computing device, or on multiple computing devices located in one location, or on multiple computing devices distributed across multiple locations and interconnected via a communication network.

[0201] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, and improvements made within the spirit and scope of this application are included within the scope of protection of this application.

Claims

1. A live streaming recommendation method, characterized in that, Applied to a terminal, the method includes: Responding to the trigger operation of the request to start live streaming to start live streaming, and obtaining live streaming conversion data in real time during the current live streaming process, as well as obtaining historical live streaming conversion data within a preset historical time period; During the current live stream, at a predetermined frequency, based on the current live stream conversion data and the historical live stream conversion data, at least one recommended category is determined; live stream recommendation information is generated based on the at least one recommended category; and the live stream recommendation information is displayed on the live stream interface. The recommended product categories include at least high-exposure and high-conversion lead generation products, low-exposure and extremely high-conversion high-profit products, and low-exposure and low-conversion future star products. The real-time acquisition of live conversion data during the current live stream includes: real-time statistics of the current page views, current sales volume, and current profit margin for each currently recommended object during the current live stream at the current moment; The acquisition of historical live streaming conversion data within a preset historical time period includes: statistically analyzing the historical pageviews, historical sales, and historical profit margins for each currently recommended object during the historical live streaming process within the preset historical time period. The step of determining at least one recommended category based on the current live stream conversion data and the historical live stream conversion data includes: For each currently recommended object, at the current moment, the current conversion rate of the currently recommended object is determined based on the current sales volume and the current pageviews of the currently recommended object; and the historical conversion rate of the currently recommended object is determined based on the historical sales volume and the historical pageviews of the currently recommended object. Based on the current pageviews, the historical pageviews, the current conversion rate, and the historical conversion rate, all currently recommended objects in the current live stream are sorted to form a first recommended object sequence, and the first target recommended object in the first recommended object sequence is determined as the traffic-driving product; The current profit margin and historical profit margin of each of the current recommended objects are sorted to form a second recommended object sequence, and the second target recommended object in the second recommended object sequence is determined as the high-profit product. Based on the current conversion rate and the historical conversion rate, all currently recommended objects in the current live broadcast process are sorted to form a third recommended object sequence, and the third target recommended object in the third recommended object sequence is determined as the future star product.

2. The method according to claim 1, characterized in that, The step of determining at least one recommended category based on the current live stream conversion data and the historical live stream conversion data at a preset frequency includes: Periodically determine the at least one recommended product category based on the current live stream conversion data and the historical live stream conversion data; or, The voice information collected in real time is analyzed to obtain the voice analysis results; and when the voice analysis results indicate that live streaming recommendations are needed, at least one recommended category is determined based on the live streaming conversion data at the current moment and the historical live streaming conversion data.

3. The method according to claim 1, characterized in that, The method further includes: During the live stream, text analysis is performed on each comment within a specific time period to obtain the text analysis results; Based on the text analysis results of each comment, the type of recommended object with the highest interaction frequency within the specific time period is determined; Obtain the link of the object to be recommended corresponding to the recommended object type from the preset recommendation object library; The link of the object to be recommended is displayed on the live stream interface for a first preset duration.

4. The method according to claim 1, characterized in that, The live stream interface also displays a product search box; the method further includes: Receive search information entered through the product search box; When a corresponding recommended object is matched in the preset recommendation object library based on the search information, or when a corresponding recommended object is matched in all the current recommended objects in the current live broadcast process based on the search information, the link of the matched recommended object is displayed on the live broadcast interface for a second preset duration.

5. The method according to claim 1, characterized in that, The method further includes: determining at least one currently recommended object during the current live broadcast process; The real-time statistics, at the current moment, after calculating the current page views, current sales volume, and current profit margin for each currently recommended object during the current live stream, further include: The current pageviews, current sales, and current profit margin of each currently recommended object are determined as the live streaming conversion data at the current moment; The method further includes, after calculating the historical views, sales volume, and profit margin of each currently recommended object during the historical live streaming process within the preset historical time period, the method also includes: The historical pageviews, historical sales, and historical profit margin of each currently recommended object are determined as the historical live streaming conversion data.

6. The method according to claim 5, characterized in that, The recommended categories include high-exposure, low-conversion products; The step of determining at least one recommended category based on the current live stream conversion data and the historical live stream conversion data includes: For each currently recommended object, obtain the recommendation duration for the current recommended object during the current live stream; For each of the currently recommended objects, at the current moment, the current conversion rate of the currently recommended object is determined based on the current sales volume and the current pageviews of the currently recommended object; Based on the recommended duration and the current conversion rate, all currently recommended objects in the current live broadcast process are sorted to form a fourth recommended object sequence; Select the fourth target recommendation object based on the fourth recommendation object sequence; The fourth target recommendation object is determined to be the high-exposure, low-transfer product at the current moment.

7. The method according to any one of claims 1 to 6, characterized in that, The method further includes: Before responding to the triggering operation, obtain at least one currently recommended object from the current live streaming process; Obtain the historical live stream conversion data corresponding to each of the current recommended objects; Based on the historical live stream conversion data corresponding to each of the current recommended objects, generate current live stream optimization suggestions.

8. The method according to any one of claims 1 to 6, characterized in that, The method further includes: In response to the end of the live stream operation, generate annotation information for each of the recommended product categories and an identifier corresponding to the annotation information; Generate a live streaming recommendation interface based on at least one recommended product category; Display the live stream recommendation interface; When a view operation corresponding to the annotation information is received on the live streaming recommendation interface, the annotation information for each of the recommended categories is displayed on the live streaming recommendation interface in response to the view operation.

9. A live streaming recommendation device, characterized in that, The device includes: The response module is used to respond to the trigger operation of the request to start live streaming to start live streaming, and to obtain the live streaming conversion data in real time during the current live streaming process, as well as the historical live streaming conversion data within a preset historical time period. The determination module is used to determine at least one recommended category during the current live broadcast, based on the live broadcast conversion data at the current moment and the historical live broadcast conversion data, according to a preset determined frequency. The generation module is used to generate live streaming recommendation information based on the at least one recommended category; The display module is used to display the live streaming recommendation information on the live streaming interface; The recommended product categories include at least high-exposure and high-conversion lead generation products, low-exposure and extremely high-conversion high-profit products, and low-exposure and low-conversion future star products. The response module is also used to perform real-time statistics on the current pageviews, sales volume, and profit margin for each currently recommended object during the current live broadcast at the current moment. The response module is also used to count the historical views, historical sales and historical profit margins of each currently recommended object during the historical live broadcast process within the preset historical time period. The determining module is further configured to: For each currently recommended object, at the current moment, the current conversion rate of the currently recommended object is determined based on the current sales volume and the current pageviews of the currently recommended object; and the historical conversion rate of the currently recommended object is determined based on the historical sales volume and the historical pageviews of the currently recommended object. Based on the current pageviews, the historical pageviews, the current conversion rate, and the historical conversion rate, all currently recommended objects in the current live stream are sorted to form a first recommended object sequence, and the first target recommended object in the first recommended object sequence is determined as the traffic-driving product; The current profit margin and historical profit margin of each of the current recommended objects are sorted to form a second recommended object sequence, and the second target recommended object in the second recommended object sequence is determined as the high-profit product. Based on the current conversion rate and the historical conversion rate, all currently recommended objects in the current live broadcast process are sorted to form a third recommended object sequence, and the third target recommended object in the third recommended object sequence is determined as the future star product.

10. The apparatus according to claim 9, characterized in that, The determining module is also used for: Periodically determine the at least one recommended product category based on the current live stream conversion data and the historical live stream conversion data; or, The voice information collected in real time is analyzed to obtain the voice analysis results; and when the voice analysis results indicate that live streaming recommendations are needed, at least one recommended category is determined based on the live streaming conversion data at the current moment and the historical live streaming conversion data.

11. The apparatus according to claim 9, characterized in that, The device further includes: The text analysis module is used to perform text analysis on each comment message acquired within a specific time period during the live broadcast, and to obtain the text analysis results. The recommendation object type determination module is used to determine the recommendation object type with the highest interaction frequency within the specific time period based on the text analysis results of each comment. The link acquisition module is used to acquire links to objects to be recommended that correspond to the recommended object type from a preset recommendation object library; The first link display module is used to display the link of the object to be recommended on the live broadcast interface for a first preset duration.

12. The apparatus according to claim 9, characterized in that, The live streaming interface also displays a product search box; the device further includes: A receiving module is used to receive search information entered through the product search box; The second link display module is used to display the link of the matched recommended object on the live broadcast interface for a second preset duration when a corresponding recommended object is matched in the preset recommended object library based on the search information, or when a corresponding recommended object is matched in all the current recommended objects in the current live broadcast process based on the search information.

13. The apparatus according to claim 9, characterized in that, The device further includes: The current recommendation object determination module is used to determine at least one current recommendation object in the current live broadcast process; The response module is further configured to determine the current pageviews, current sales volume, and current profit margin of each of the currently recommended objects as the live streaming conversion data at the current moment; The response module is further configured to determine the historical pageviews, historical sales, and historical profit margin of each currently recommended object as the historical live streaming conversion data.

14. A live streaming recommendation device, characterized in that, include: Memory, used to store executable instructions; A processor, when executing executable instructions stored in the memory, implements the live streaming recommendation method according to any one of claims 1 to 8.

15. A computer-readable storage medium, characterized in that, It stores executable instructions for implementing the live streaming recommendation method according to any one of claims 1 to 8 when executed by a processor.

16. A computer program product, characterized in that, The computer program product includes computer instructions that, when executed by a processor, implement the live streaming recommendation method according to any one of claims 1 to 8.

Images