Play control method, device and equipment based on IPTV (Internet Protocol Television) and medium

Abstract

The invention relates to the technical field of IPTV, and discloses an IPTV-based playing control method, device and equipment and a medium, and the method comprises the steps: carrying out the semantic extraction of a description text of a public opinion event and the professional knowledge of the public opinion event, and obtaining a semantic vector of the public opinion event; generating a popularity index of the description text of the public opinion event, when the popularity index is greater than a preset index, performing emotion recognition on the description text of the public opinion event by adopting an emotion recognition model to obtain an emotion tag, and integrating the name of the public opinion platform, the description text of the public opinion event, the emotion tag and the popularity index to obtain public opinion prompt content. And splicing the semantic vector of the public opinion prompt content, the semantic vector of the alarm event and the semantic vector of the work order event to obtain a fusion vector, processing the fusion vector through a large language model, and generating a play control strategy of the IPTV live broadcast content. According to the invention, the playing control efficiency of the IPTV live broadcast content can be improved.

Description

Technical Field

[0001] This application relates to the field of IPTV technology, and in particular to a playback control method, apparatus, device and medium based on IPTV. Background Technology

[0002] IPTV live content refers to media content that is pushed to end users in real-time streaming through the IPTV (Internet Protocol Television) technology system and is presented in sync with the program production and broadcast schedule.

[0003] However, the current process of obtaining playback control strategies for IPTV live content is cumbersome, hindering the improvement of acquisition efficiency. This is because existing technologies primarily rely on manual acquisition of these strategies. Manual acquisition consumes significant human and time resources, increasing the time required for obtaining playback control strategies and making it susceptible to human intervention. Therefore, it is detrimental to improving the efficiency of playback control strategy acquisition. Summary of the Invention

[0004] This application provides a playback control method, apparatus, device, and medium based on IPTV to solve the technical problem that the acquisition process of playback control strategies for existing IPTV live content is cumbersome and not conducive to improving the efficiency of acquisition.

[0005] In a first aspect, embodiments of this application provide an IPTV-based playback control method applied to electronic devices, the playback control method comprising: Obtain the description text of alarm events from the alarm system associated with IPTV live streaming content; obtain the description text of work order events from the work order system associated with IPTV live streaming content; obtain the description text of public opinion events from the public opinion platform associated with IPTV live streaming content. The description texts of alarm events, work order events, and public opinion events are matched with the knowledge graph to obtain the professional knowledge of alarm events, work order events, and public opinion events, respectively. Semantic extraction is performed on the descriptive text of alarm events and the professional knowledge of alarm events to obtain the semantic vector of alarm events. Semantic extraction is also performed on the descriptive text of work order events and the professional knowledge of work order events to obtain the semantic vector of work order events. Semantic extraction is also performed on the descriptive text of public opinion events and the professional knowledge of public opinion events to obtain the semantic vector of public opinion events. The semantic vectors of alarm events and work order events are processed by the cosine similarity algorithm to obtain the first similarity. The semantic vectors of work order events and public opinion events are processed by the cosine similarity algorithm to obtain the second similarity. The average of the first and second similarities is selected as the current similarity. The current similarity is the similarity between the semantic vectors of alarm events, work order events, and public opinion events. When the current similarity is greater than the preset similarity, a heat model is used to generate a heat index for the descriptive text of the public opinion event. When the heat index is greater than the preset index, a sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, sentiment tags, and heat index are integrated to obtain public opinion prompt content. The semantic vectors of the public opinion prompt content, the semantic vectors of the alarm event, and the semantic vectors of the work order event are concatenated to obtain a fusion vector. The fusion vector is processed by a large language model to generate a playback control strategy for IPTV live content.

[0006] In one possible implementation of the first aspect, the step of matching the description text of the alarm event, the description text of the work order event, and the description text of the public opinion event with the knowledge graph respectively to obtain the professional knowledge of the alarm event, the professional knowledge of the work order event, and the professional knowledge of the public opinion event, includes: Call the retrieval interface of the retrieval enhancement model and input the description text of the alarm event, the description text of the work order event, and the description text of the public opinion event into the retrieval interface; The retrieval interface matches the description text of alarm events, work order events, and public opinion events with the knowledge graph to obtain the professional knowledge of alarm events, work order events, and public opinion events, respectively.

[0007] In one possible implementation of the first aspect, the semantic extraction of the descriptive text of the alarm event and the professional knowledge of the alarm event to obtain a semantic vector of the alarm event, the semantic extraction of the descriptive text of the work order event and the professional knowledge of the work order event to obtain a semantic vector of the work order event, and the semantic extraction of the descriptive text of the public opinion event and the professional knowledge of the public opinion event to obtain a semantic vector of the public opinion event, includes: The description text of the alarm event and the professional knowledge of the alarm event are input into the semantic parsing unit. The semantic parsing unit performs semantic extraction on the description text of the alarm event and the professional knowledge of the alarm event to obtain the semantic vector of the alarm event. The description text of the work order event and the professional knowledge of the work order event are input into the semantic parsing unit. The semantic parsing unit performs semantic extraction on the description text of the work order event and the professional knowledge of the work order event to obtain the semantic vector of the work order event. The descriptive text of the public opinion event and the professional knowledge related to the public opinion event are input into the semantic parsing unit. The semantic parsing unit performs semantic extraction on the descriptive text and the professional knowledge related to the public opinion event to obtain the semantic vector of the public opinion event.

[0008] In one possible implementation of the first aspect, when the current similarity is greater than a preset similarity, a heat model is used to generate a heat index of the descriptive text of the public opinion event. When the heat index is greater than a preset index, a sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, the sentiment tags, and the heat index are integrated to obtain public opinion prompt content. The semantic vectors of the public opinion prompt content, the semantic vectors of the alarm event, and the semantic vectors of the work order event are concatenated to obtain a fusion vector. The fusion vector is processed by a large language model to generate a playback control strategy for IPTV live content, including: When the current similarity is greater than the preset similarity, a heat model is used to generate a heat index of the descriptive text of the public opinion event; When the popularity index is greater than the preset index, the sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, sentiment tags, and popularity index are integrated to obtain the public opinion prompt content. The description texts of alarm events, work order events, and public opinion events are categorized into an event cluster. Keywords are extracted from the event cluster, and these keywords are matched with multiple preset keywords. The preset question corresponding to the successfully matched preset keyword is selected as the target question of the event cluster. The semantic vectors of the target question, public opinion prompts, alarm events, and work order events are concatenated to obtain a fusion vector. The fusion vector is then input into a large language model, which processes the fusion vector to generate a playback control strategy for IPTV live content.

[0009] In one possible implementation of the first aspect, when the current similarity is greater than a preset similarity, a heat model is used to generate a heat index of the descriptive text of the public opinion event. When the heat index is greater than a preset index, a sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, the sentiment tags, and the heat index are integrated to obtain public opinion prompt content. The semantic vectors of the public opinion prompt content, the semantic vectors of the alarm event, and the semantic vectors of the work order event are concatenated to obtain a fusion vector. After processing the fusion vector through a large language model to generate a playback control strategy for IPTV live content, the playback control method includes: When the playback control strategy is an expansion strategy, an expansion instruction is sent to the content distribution system. The expansion instruction includes instructions to increase the distribution bandwidth of IPTV live content and instructions to expand the number of cache nodes for IPTV live content. When the playback control policy is set to remove content, a removal instruction is sent to the content storage system. The removal instruction includes instructions to delete IPTV live content and instructions to block access to IPTV live content.

[0010] In one possible implementation of the first aspect, the heat model is defined as follows: ; D represents the popularity index of the descriptive text of a public opinion event; the higher the popularity index of the descriptive text of a public opinion event, the higher the influence of the descriptive text of the public opinion event; the lower the popularity index of the descriptive text of a public opinion event, the lower the influence of the descriptive text of the public opinion event. F represents the number of reposts of the descriptive text of the public opinion event; C represents the number of comments on the descriptive text of the public opinion event; L represents the number of likes on the descriptive text of the public opinion event; λ represents the time decay factor, and λ represents the decay rate constant. t represents the difference between the generation time of the descriptive text of the public opinion event and the current time, in hours or minutes; W represents the weight parameter of the public opinion platform.

[0011] In one possible implementation of the first aspect, the first similarity is the similarity between the semantic vector of the alarm event and the semantic vector of the work order event; the second similarity is the similarity between the semantic vector of the work order event and the semantic vector of the public opinion event.

[0012] Secondly, embodiments of this application provide an IPTV-based playback control device, applied to electronic devices, comprising: The first acquisition module is used to acquire the description text of alarm events from the alarm system associated with IPTV live content, the description text of work order events from the work order system associated with IPTV live content, and the description text of public opinion events from the public opinion platform associated with IPTV live content. The matching module is used to match the description text of alarm events, work order events, and public opinion events with the knowledge graph to obtain the professional knowledge of alarm events, work order events, and public opinion events, respectively. The second acquisition module is used to perform semantic extraction on the description text of alarm events and the professional knowledge of alarm events to obtain the semantic vector of alarm events; to perform semantic extraction on the description text of work order events and the professional knowledge of work order events to obtain the semantic vector of work order events; and to perform semantic extraction on the description text of public opinion events and the professional knowledge of public opinion events to obtain the semantic vector of public opinion events. The third acquisition module is used to process the semantic vectors of alarm events and work order events using the cosine similarity algorithm to obtain the first similarity. It then processes the semantic vectors of work order events and public opinion events using the cosine similarity algorithm to obtain the second similarity. The average of the first and second similarities is selected as the current similarity. The current similarity is the similarity between the semantic vectors of alarm events, work order events, and public opinion events. The control module is used to generate a popularity index for the descriptive text of the public opinion event when the current similarity is greater than the preset similarity. When the popularity index is greater than the preset index, the module uses a sentiment recognition model to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The module integrates the name of the public opinion platform, the descriptive text of the public opinion event, the sentiment tags, and the popularity index to obtain public opinion prompt content. The module concatenates the semantic vectors of the public opinion prompt content, the semantic vectors of the alarm event, and the semantic vectors of the work order event to obtain a fusion vector. The module processes the fusion vector through a large language model to generate a playback control strategy for IPTV live content.

[0013] Thirdly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the playback control method described in the first aspect above.

[0014] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the playback control method described in the first aspect above.

[0015] Fifthly, embodiments of this application provide a computer program product that, when run on an electronic device, causes the electronic device to execute the playback control method described in the first aspect.

[0016] The beneficial effects of this application's embodiments are twofold. Firstly, when the current similarity is greater than a preset similarity, a heat model is used to generate a heat index for the descriptive text of the public opinion event. When the heat index is greater than a preset index, a sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, sentiment tags, and heat index are integrated to obtain public opinion prompt content. The semantic vectors of the public opinion prompt content, the semantic vectors of the alarm event, and the semantic vectors of the work order event are concatenated to obtain a fusion vector. The fusion vector is processed through a large language model to generate a playback control strategy for IPTV live content. Since no manual acquisition is required, the acquisition time for the playback control strategy of IPTV live content is reduced, which is conducive to improving the acquisition efficiency of the playback control strategy of IPTV live content. Secondly, it is not affected by manual intervention, which is conducive to improving the reliability of the acquired playback control strategy of IPTV live content. Attached Figure Description

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

[0018] Figure 1 This is an application scenario diagram of the playback control method provided in the embodiments of this application; Figure 2 This is a flowchart illustrating the playback control method provided in an embodiment of this application; Figure 3 A flowchart illustrating the implementation of S205 provided in this application embodiment; Figure 4 A schematic block diagram of a playback control device provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.

[0020] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0021] The playback control method provided in this application can be applied to electronic devices such as servers, mobile phones, tablets, wearable devices, in-vehicle devices, laptops, netbooks, and personal digital assistants (PDAs). This application does not impose any restrictions on the specific type of electronic device.

[0022] Please see Figure 1 , Figure 1 The application scenario diagram of the playback control method provided in the embodiments of this application is described in detail below: The electronic devices are respectively connected to the alarm system associated with IPTV live streaming content, the work order system associated with IPTV live streaming content, and the public opinion platform associated with IPTV live streaming content; Electronic devices obtain description text of alarm events from the alarm system associated with IPTV live content, description text of work order events from the work order system associated with IPTV live content, and description text of public opinion events from the public opinion platform associated with IPTV live content.

[0023] In this embodiment of the application, after the electronic device establishes a connection with the alarm system, the work order system, and the public opinion platform, it can quickly obtain the description text of alarm events, the description text of work order events, and the description text of public opinion events.

[0024] Please see Figure 2 , Figure 2 This is a flowchart illustrating the playback control method provided in an embodiment of this application, which can be applied to electronic devices.

[0025] like Figure 2 As shown, the playback control method provided in this application includes the following steps, detailed below: S201. Obtain the description text of alarm events from the alarm system associated with IPTV live content, obtain the description text of work order events from the work order system associated with IPTV live content, and obtain the description text of public opinion events from the public opinion platform associated with IPTV live content. Among them, the IPTV live content association alarm system is a technical system in the IPTV operation platform used to monitor the playback quality of live content and trigger alarm prompts.

[0026] Among them, the IPTV live content associated work order system is a technical system in the IPTV operation platform used to receive, transfer, and process user feedback, fault reports, compliance complaints, and other requests related to IPTV live content.

[0027] Public opinion platforms associated with IPTV live streaming content include, but are not limited to, Weibo, WeChat, and video interaction platforms.

[0028] The descriptive text of the alarm event is used to provide an accurate description of the fault, enabling the large language model to quickly locate the problem type and severity.

[0029] The description text of the work order event is used to provide a description of the user's problem, enabling the large language model to understand the user's needs.

[0030] The descriptive text of public opinion events is used to supplement the contextual information of large language models, enabling them to perceive the scope of the issue's impact.

[0031] For ease of explanation, the following example is provided: For example, the description text of the alarm event: Server CPU utilization 89%, IPTV live broadcast content delay 5.1 seconds; For example, the description text of the work order event: When I was watching the live broadcast of the ball game on IPTV at 8 pm last night, the picture kept choppy and I couldn't watch it at all. It made me so angry. The replay was still like this today. For example, the description of a public opinion event: IPTV live broadcasts in Guangdong Province frequently buffered, even after switching to several channels.

[0032] S202, match the description text of alarm events, work order events, and public opinion events with the knowledge graph respectively to obtain the professional knowledge of alarm events, work order events, and public opinion events respectively. Specifically, the process of matching the description text of alarm events, work order events, and public opinion events with a knowledge graph to obtain professional knowledge of alarm events, work order events, and public opinion events, respectively, includes: Call the retrieval interface of the retrieval enhancement model and input the description text of the alarm event, the description text of the work order event, and the description text of the public opinion event into the retrieval interface; The retrieval interface matches the description text of alarm events, work order events, and public opinion events with the knowledge graph to obtain the professional knowledge of alarm events, work order events, and public opinion events, respectively.

[0033] For illustrative purposes, the retrieval interface matches the description text of alarm events, work order events, and public opinion events with the knowledge graph, respectively, to obtain the professional knowledge of alarm events, work order events, and public opinion events, including: Import the knowledge graph of the broadcast control system, and the retrieval interface matches the description text of alarm events, work order events, and public opinion events with the knowledge graph of the broadcast control system to obtain the professional knowledge of alarm events, work order events, and public opinion events, respectively.

[0034] S203, Semantic extraction is performed on the description text of alarm events and the professional knowledge of alarm events to obtain the semantic vector of alarm events; semantic extraction is performed on the description text of work order events and the professional knowledge of work order events to obtain the semantic vector of work order events; semantic extraction is performed on the description text of public opinion events and the professional knowledge of public opinion events to obtain the semantic vector of public opinion events. The process of semantically extracting the descriptive text and professional knowledge of alarm events to obtain semantic vectors for alarm events, semantically extracting the descriptive text and professional knowledge of work order events to obtain semantic vectors for work order events, and semantically extracting the descriptive text and professional knowledge of public opinion events to obtain semantic vectors for public opinion events includes: The description text of the alarm event and the professional knowledge of the alarm event are input into the semantic parsing unit. The semantic parsing unit performs semantic extraction on the description text of the alarm event and the professional knowledge of the alarm event to obtain the semantic vector of the alarm event. The description text of the work order event and the professional knowledge of the work order event are input into the semantic parsing unit. The semantic parsing unit performs semantic extraction on the description text of the work order event and the professional knowledge of the work order event to obtain the semantic vector of the work order event. The descriptive text of the public opinion event and the professional knowledge related to the public opinion event are input into the semantic parsing unit. The semantic parsing unit performs semantic extraction on the descriptive text and the professional knowledge related to the public opinion event to obtain the semantic vector of the public opinion event.

[0035] S204. The semantic vectors of alarm events and work order events are processed by the cosine similarity algorithm to obtain the first similarity. The semantic vectors of work order events and public opinion events are processed by the cosine similarity algorithm to obtain the second similarity. The average of the first and second similarities is selected as the current similarity. The current similarity is the similarity between the semantic vectors of alarm events, work order events, and public opinion events. The first similarity is the similarity between the semantic vector of the alarm event and the semantic vector of the work order event; the second similarity is the similarity between the semantic vector of the work order event and the semantic vector of the public opinion event.

[0036] S205: When the current similarity is greater than the preset similarity, a heat model is used to generate a heat index of the descriptive text of the public opinion event. When the heat index is greater than the preset index, a sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, sentiment tags, and heat index are integrated to obtain public opinion prompt content. The semantic vector of the public opinion prompt content, the semantic vector of the alarm event, and the semantic vector of the work order event are concatenated to obtain a fusion vector. The fusion vector is processed by a large language model to generate a playback control strategy for IPTV live content.

[0037] The popularity model is defined as follows: ; D represents the popularity index of the descriptive text of a public opinion event; the higher the popularity index of the descriptive text of a public opinion event, the higher the influence of the descriptive text of the public opinion event; the lower the popularity index of the descriptive text of a public opinion event, the lower the influence of the descriptive text of the public opinion event. F represents the number of reposts of the descriptive text of the public opinion event; C represents the number of comments on the descriptive text of the public opinion event; L represents the number of likes on the descriptive text of the public opinion event; λ represents the time decay factor, and λ represents the decay rate constant. t represents the difference between the generation time of the descriptive text of the public opinion event and the current time, in hours or minutes; W represents the weight parameter of the public opinion platform.

[0038] The process involves concatenating the semantic vectors of public opinion alerts, alarm events, and work order events to obtain a fused vector. This fused vector is then processed by a large language model to generate playback control strategies for IPTV live content. Therefore, the large language model possesses a deep understanding of public opinion alerts, alarm events, and work order events. By processing the fused vector through the large language model, the playback control strategies for IPTV live content are generated. Furthermore, by deeply analyzing the emotional tags in public opinion alerts, the degree of system anomalies reflected in alarms, and the problem-solving progress in work orders, the large language model can accurately locate the core elements and key characteristics of problems. Based on this, the playback control strategies for IPTV live content generated by the large language model can accurately adapt to the actual situation, avoiding decision-making errors caused by incomplete information or misunderstandings, and ensuring the scientific validity and effectiveness of the playback control strategies.

[0039] Specifically, when the current similarity is greater than a preset similarity, a heat model is used to generate a heat index for the descriptive text of the public opinion event. When the heat index is greater than a preset index, a sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, the sentiment tags, and the heat index are integrated to obtain public opinion alert content. The semantic vectors of the public opinion alert content, the alarm event, and the work order event are concatenated to obtain a fusion vector. After processing the fusion vector through a large language model to generate the playback control strategy for IPTV live content, the playback control method includes: When the playback control strategy is an expansion strategy, an expansion instruction is sent to the content distribution system. The expansion instruction includes instructions to increase the distribution bandwidth of IPTV live content and instructions to expand the number of cache nodes for IPTV live content. When the playback control policy is set to remove content, a removal instruction is sent to the content storage system. The removal instruction includes instructions to delete IPTV live content and instructions to block access to IPTV live content.

[0040] When the playback control strategy is an expansion strategy, an expansion command is sent to the content distribution system. This command means faster data transmission, allowing IPTV live content to be transmitted from the source to the user terminal more quickly, reducing waiting time during transmission. This not only shortens the interval between the user clicking to play and actually watching, but also ensures the real-time nature of IPTV live content, allowing users to obtain information almost simultaneously with the scene. This advantage is even more pronounced for content with extremely high timeliness requirements, such as news broadcasts and sports events. For clarity, an example is provided below: For example, the description text of the alarm event: CDN node load in Beijing exceeded the limit, IPTV live broadcast content was delayed by 4.2 seconds; For example, the description text of a work order event: When I watch IPTV live content, the picture is choppy and unwatchable; For example, the preset index is 200; the description text of the public opinion event is posted on Weibo, and the description text of the public opinion event is: "The IPTV live broadcast in Shenzhen is stuck like a PPT, what's going on?"; the popularity index of the description text of the public opinion event is 326, and the sentiment label is negative. At this time, the public opinion alert content is as follows: The online public opinion was published on Weibo; the description of the public opinion event is: In Shenzhen, the IPTV live broadcast content is stuck like a PPT, what's going on; the sentiment tag is negative; the popularity index is 326; The semantic vectors of public opinion alerts, alarm events, and work order events are concatenated to obtain a fusion vector. This fusion vector is then processed using a large language model to generate an expansion strategy for IPTV live streaming content. Expansion instructions are then sent to the content distribution system. These instructions include increasing the distribution bandwidth of IPTV live streaming content and expanding the number of cache nodes for IPTV live streaming content, thus ensuring the real-time performance of IPTV live streaming content.

[0041] When the playback control policy is set to remove content, a removal instruction is sent to the content storage system. This instruction includes deleting IPTV live content and blocking access to it. Timely removal instructions ensure the platform can swiftly and thoroughly eliminate content that violates laws, regulations, industry standards, or platform rules, avoiding legal penalties and regulatory warnings for continued dissemination of illegal information and maintaining a legitimate platform environment.

[0042] For ease of explanation, the following example is provided: For example, the description text of the alarm event: Serious violations of regulations were detected in the IPTV live broadcast content, triggering an automatic system alarm to require the live broadcast to be taken down; For example, the description of the work order incident: Multiple viewers complained that the live broadcast content contained false advertising and did not match the facts, and strongly demanded that it be taken down; For example, the preset index is 200; the descriptive text of the public opinion event is published on the video's interactive platform, and the descriptive text of the public opinion event is: This anchor's remarks are too bad, not taking them down is disrespectful to the audience; the popularity index of the descriptive text of the public opinion event is 230, and the sentiment label is negative; At this time, the public opinion alert message is as follows: The online public opinion was posted in the comment section of the video platform; the description of the public opinion event was: This streamer's remarks are too rude, not taking them down is disrespectful to the audience; the emotional tag is negative; the popularity index is 326; The semantic vectors of public opinion alerts, alarm events, and work order events are concatenated to obtain a fusion vector. This fusion vector is then processed using a large language model to generate a strategy for removing IPTV live content. A removal command is then sent to the content storage system, including commands to delete IPTV live content and to block access to IPTV live content. This ensures that the platform can quickly and thoroughly remove illegal content, avoiding the risks of legal penalties and regulatory warnings due to the continued dissemination of illegal information, and maintaining the platform's legitimate environment.

[0043] The beneficial effects of this application's embodiments are twofold. Firstly, when the current similarity is greater than a preset similarity, a heat model is used to generate a heat index for the descriptive text of the public opinion event. When the heat index is greater than a preset index, a sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, sentiment tags, and heat index are integrated to obtain public opinion prompt content. The semantic vectors of the public opinion prompt content, the semantic vectors of the alarm event, and the semantic vectors of the work order event are concatenated to obtain a fusion vector. The fusion vector is processed through a large language model to generate a playback control strategy for IPTV live content. Since no manual acquisition is required, the acquisition time for the playback control strategy of IPTV live content is reduced, which is conducive to improving the acquisition efficiency of the playback control strategy of IPTV live content. Secondly, it is not affected by manual intervention, which is conducive to improving the reliability of the acquired playback control strategy of IPTV live content.

[0044] Please see Figure 3 , Figure 3 The implementation flowchart of S205 provided in the embodiments of this application is described in detail below: S301, When the current similarity is greater than the preset similarity, the heat model is used to generate the heat index of the descriptive text of the public opinion event; S302, When the popularity index is greater than the preset index, the sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, sentiment tags, and popularity index are integrated to obtain public opinion prompt content. S303: The description texts of alarm events, work order events, and public opinion events are categorized into an event cluster. Keywords in the event cluster are extracted and matched with multiple preset keywords. The preset question corresponding to the successfully matched preset keyword is selected as the target question of the event cluster. The semantic vectors of the target question, public opinion prompts, alarm events, and work order events are concatenated to obtain a fusion vector. The fusion vector is input into a large language model, which processes the fusion vector to generate a playback control strategy for IPTV live content.

[0045] In this embodiment, the large language model, with its powerful data processing capabilities, can perform real-time in-depth analysis of the semantic vectors of the target issue, the semantic vectors of public opinion alerts, the semantic vectors of alarm events, and the semantic vectors of work order events, rapidly generating highly targeted playback control strategies. This process significantly shortens the time cycle from target issue discovery to playback control strategy formulation, enabling the system to respond to target issues promptly.

[0046] For the playback control method described in the above embodiments, please refer to [link / reference]. Figure 4 , Figure 4 This is a schematic block diagram of a playback control device provided in an embodiment of this application. Figure 4 The playback control device 400 shown can be applied to, for example... Figure 1 The application scenario diagram shows electronic devices. The following section uses electronic devices as an example to illustrate this. Figure 4 The playback control device 400 shown will be described in detail. The playback control device 400 may include a first acquisition module 401, a matching module 402, a second acquisition module 403, a third acquisition module 404, and a control module 405.

[0047] The first acquisition module 401 is used to acquire the description text of alarm events from the alarm system associated with IPTV live content, the description text of work order events from the work order system associated with IPTV live content, and the description text of public opinion events from the public opinion platform associated with IPTV live content. The matching module 402 is used to match the description text of alarm events, the description text of work order events, and the description text of public opinion events with the knowledge graph respectively, so as to obtain the professional knowledge of alarm events, the professional knowledge of work order events, and the professional knowledge of public opinion events respectively. The second acquisition module 403 is used to perform semantic extraction on the description text of alarm events and the professional knowledge of alarm events to obtain the semantic vector of alarm events; to perform semantic extraction on the description text of work order events and the professional knowledge of work order events to obtain the semantic vector of work order events; and to perform semantic extraction on the description text of public opinion events and the professional knowledge of public opinion events to obtain the semantic vector of public opinion events. The third acquisition module 404 is used to process the semantic vectors of alarm events and work order events using a cosine similarity algorithm to obtain a first similarity, process the semantic vectors of work order events and public opinion events using a cosine similarity algorithm to obtain a second similarity, and select the average of the first and second similarities as the current similarity. The current similarity is the similarity between the semantic vectors of alarm events, work order events, and public opinion events. The control module 405 is used to generate a heat index of the description text of the public opinion event using a heat model when the current similarity is greater than a preset similarity. When the heat index is greater than a preset index, it uses a sentiment recognition model to perform sentiment recognition on the description text of the public opinion event to obtain sentiment tags. It integrates the name of the public opinion platform, the description text of the public opinion event, sentiment tags, and heat index to obtain public opinion prompt content. It concatenates the semantic vector of the public opinion prompt content, the semantic vector of the alarm event, and the semantic vector of the work order event to obtain a fusion vector. It processes the fusion vector through a large language model to generate a playback control strategy for IPTV live content.

[0048] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0049] The beneficial effects of this application's embodiments are twofold. Firstly, when the current similarity is greater than a preset similarity, a heat model is used to generate a heat index for the descriptive text of the public opinion event. When the heat index is greater than a preset index, a sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, sentiment tags, and heat index are integrated to obtain public opinion prompt content. The semantic vectors of the public opinion prompt content, the semantic vectors of the alarm event, and the semantic vectors of the work order event are concatenated to obtain a fusion vector. The fusion vector is processed through a large language model to generate a playback control strategy for IPTV live content. Since no manual acquisition is required, the acquisition time for the playback control strategy of IPTV live content is reduced, which is conducive to improving the acquisition efficiency of the playback control strategy of IPTV live content. Secondly, it is not affected by manual intervention, which is conducive to improving the reliability of the acquired playback control strategy of IPTV live content.

[0050] Please see Figure 5 , Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0051] like Figure 5 As shown, Figure 5 The electronic device 2 includes: at least one processor 20, a memory 21, and a computer program 22 stored in the memory 21 and executable on the at least one processor 20, wherein the processor 20 executes the computer program 22 to implement the steps in any of the above method embodiments.

[0052] The electronic device 2 may include, but is not limited to, a processor 20 and a memory 21. Those skilled in the art will understand that... Figure 5 This is merely an example of electronic device 2 and does not constitute a limitation on electronic device 2. It may include more or fewer components than shown in the figure, or combine certain components, or different components. For example, it may also include input / output devices, network access devices, etc.

[0053] The processor 20 is used to run a computer program 22 stored in the memory 21, and performs the following steps when executing the computer program 22: Obtain the description text of alarm events from the alarm system associated with IPTV live streaming content; obtain the description text of work order events from the work order system associated with IPTV live streaming content; obtain the description text of public opinion events from the public opinion platform associated with IPTV live streaming content. The description texts of alarm events, work order events, and public opinion events are matched with the knowledge graph to obtain the professional knowledge of alarm events, work order events, and public opinion events, respectively. Semantic extraction is performed on the descriptive text of alarm events and the professional knowledge of alarm events to obtain the semantic vector of alarm events. Semantic extraction is also performed on the descriptive text of work order events and the professional knowledge of work order events to obtain the semantic vector of work order events. Semantic extraction is also performed on the descriptive text of public opinion events and the professional knowledge of public opinion events to obtain the semantic vector of public opinion events. The semantic vectors of alarm events and work order events are processed by the cosine similarity algorithm to obtain the first similarity. The semantic vectors of work order events and public opinion events are processed by the cosine similarity algorithm to obtain the second similarity. The average of the first and second similarities is selected as the current similarity. The current similarity is the similarity between the semantic vectors of alarm events, work order events, and public opinion events. When the current similarity is greater than the preset similarity, a heat model is used to generate a heat index for the descriptive text of the public opinion event. When the heat index is greater than the preset index, a sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, sentiment tags, and heat index are integrated to obtain public opinion prompt content. The semantic vectors of the public opinion prompt content, the semantic vectors of the alarm event, and the semantic vectors of the work order event are concatenated to obtain a fusion vector. The fusion vector is processed by a large language model to generate a playback control strategy for IPTV live content.

[0054] The processor 20 may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors, field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0055] In some embodiments, the memory 21 may be an internal storage unit of the electronic device 2, such as a hard disk or memory of the electronic device 2.

[0056] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0057] This application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.

[0058] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A playback control method based on IPTV, characterized in that, The playback control method, applied to electronic devices, includes: Obtain the description text of alarm events from the alarm system associated with IPTV live streaming content; obtain the description text of work order events from the work order system associated with IPTV live streaming content; obtain the description text of public opinion events from the public opinion platform associated with IPTV live streaming content. The description texts of alarm events, work order events, and public opinion events are matched with the knowledge graph to obtain the professional knowledge of alarm events, work order events, and public opinion events, respectively. Semantic extraction is performed on the descriptive text of alarm events and the professional knowledge of alarm events to obtain the semantic vector of alarm events. Semantic extraction is also performed on the descriptive text of work order events and the professional knowledge of work order events to obtain the semantic vector of work order events. Semantic extraction is also performed on the descriptive text of public opinion events and the professional knowledge of public opinion events to obtain the semantic vector of public opinion events. The semantic vectors of alarm events and work order events are processed by the cosine similarity algorithm to obtain the first similarity. The semantic vectors of work order events and public opinion events are processed by the cosine similarity algorithm to obtain the second similarity. The average of the first and second similarities is selected as the current similarity. The current similarity is the similarity between the semantic vectors of alarm events, work order events, and public opinion events. When the current similarity is greater than the preset similarity, a heat model is used to generate a heat index for the descriptive text of the public opinion event. When the heat index is greater than the preset index, a sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, sentiment tags, and heat index are integrated to obtain public opinion prompt content. The semantic vectors of the public opinion prompt content, the semantic vectors of the alarm event, and the semantic vectors of the work order event are concatenated to obtain a fusion vector. The fusion vector is processed by a large language model to generate a playback control strategy for IPTV live content.

2. The playback control method according to claim 1, characterized in that, The process involves matching the description text of alarm events, work order events, and public opinion events with the knowledge graph to obtain professional knowledge related to alarm events, work order events, and public opinion events, respectively. This includes: Call the retrieval interface of the retrieval enhancement model and input the description text of the alarm event, the description text of the work order event, and the description text of the public opinion event into the retrieval interface; The retrieval interface matches the description text of alarm events, work order events, and public opinion events with the knowledge graph to obtain the professional knowledge of alarm events, work order events, and public opinion events, respectively.

3. The playback control method according to claim 1, characterized in that, The process of semantically extracting the descriptive text and professional knowledge of alarm events to obtain semantic vectors for alarm events, semantically extracting the descriptive text and professional knowledge of work order events to obtain semantic vectors for work order events, and semantically extracting the descriptive text and professional knowledge of public opinion events to obtain semantic vectors for public opinion events includes: The description text of the alarm event and the professional knowledge of the alarm event are input into the semantic parsing unit. The semantic parsing unit performs semantic extraction on the description text of the alarm event and the professional knowledge of the alarm event to obtain the semantic vector of the alarm event. The description text of the work order event and the professional knowledge of the work order event are input into the semantic parsing unit. The semantic parsing unit performs semantic extraction on the description text of the work order event and the professional knowledge of the work order event to obtain the semantic vector of the work order event. The descriptive text of the public opinion event and the professional knowledge related to the public opinion event are input into the semantic parsing unit. The semantic parsing unit performs semantic extraction on the descriptive text and the professional knowledge related to the public opinion event to obtain the semantic vector of the public opinion event.

4. The playback control method according to claim 1, characterized in that, When the current similarity is greater than a preset similarity, a heat model is used to generate a heat index for the descriptive text of the public opinion event. When the heat index is greater than a preset index, a sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, the sentiment tags, and the heat index are integrated to obtain public opinion alert content. The semantic vectors of the public opinion alert content, the alarm event, and the work order event are concatenated to obtain a fusion vector. The fusion vector is processed by a large language model to generate a playback control strategy for IPTV live content, including: When the current similarity is greater than the preset similarity, a heat model is used to generate a heat index of the descriptive text of the public opinion event; When the popularity index is greater than the preset index, the sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, sentiment tags, and popularity index are integrated to obtain the public opinion prompt content. The description texts of alarm events, work order events, and public opinion events are categorized into an event cluster. Keywords are extracted from the event cluster, and these keywords are matched with multiple preset keywords. The preset question corresponding to the successfully matched preset keyword is selected as the target question of the event cluster. The semantic vectors of the target question, public opinion prompts, alarm events, and work order events are concatenated to obtain a fusion vector. The fusion vector is then input into a large language model, which processes the fusion vector to generate a playback control strategy for IPTV live content.

5. The playback control method according to claim 1, characterized in that, When the current similarity is greater than a preset similarity, a heat model is used to generate a heat index for the descriptive text of the public opinion event. When the heat index is greater than a preset index, a sentiment recognition model is used to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The name of the public opinion platform, the descriptive text of the public opinion event, the sentiment tags, and the heat index are integrated to obtain public opinion alert content. The semantic vectors of the public opinion alert content, the alarm event, and the work order event are concatenated to obtain a fusion vector. After processing the fusion vector through a large language model to generate a playback control strategy for IPTV live content, the playback control method includes: When the playback control strategy is an expansion strategy, an expansion instruction is sent to the content distribution system. The expansion instruction includes instructions to increase the distribution bandwidth of IPTV live content and instructions to expand the number of cache nodes for IPTV live content. When the playback control policy is set to remove content, a removal instruction is sent to the content storage system. The removal instruction includes instructions to delete IPTV live content and instructions to block access to IPTV live content.

6. The playback control method according to claim 1, characterized in that, The popularity model is defined as follows: ； D represents the popularity index of the descriptive text of a public opinion event; the higher the popularity index of the descriptive text of a public opinion event, the higher the influence of the descriptive text of the public opinion event; the lower the popularity index of the descriptive text of a public opinion event, the lower the influence of the descriptive text of the public opinion event. F represents the number of reposts of the descriptive text of the public opinion event; C represents the number of comments on the descriptive text of the public opinion event; L represents the number of likes on the descriptive text of the public opinion event; λ represents the time decay factor, and λ represents the decay rate constant. t represents the difference between the generation time of the descriptive text of the public opinion event and the current time, in hours or minutes; W represents the weight parameter of the public opinion platform.

7. The playback control method according to claim 1, characterized in that, The first similarity is the similarity between the semantic vector of the alarm event and the semantic vector of the work order event; the second similarity is the similarity between the semantic vector of the work order event and the semantic vector of the public opinion event.

8. A playback control device based on IPTV, characterized in that, Applied to electronic devices, including: The first acquisition module is used to acquire the description text of alarm events from the alarm system associated with IPTV live content, the description text of work order events from the work order system associated with IPTV live content, and the description text of public opinion events from the public opinion platform associated with IPTV live content. The matching module is used to match the description text of alarm events, work order events, and public opinion events with the knowledge graph to obtain the professional knowledge of alarm events, work order events, and public opinion events, respectively. The second acquisition module is used to perform semantic extraction on the description text of alarm events and the professional knowledge of alarm events to obtain the semantic vector of alarm events; to perform semantic extraction on the description text of work order events and the professional knowledge of work order events to obtain the semantic vector of work order events; and to perform semantic extraction on the description text of public opinion events and the professional knowledge of public opinion events to obtain the semantic vector of public opinion events. The third acquisition module is used to process the semantic vectors of alarm events and work order events using the cosine similarity algorithm to obtain the first similarity. It then processes the semantic vectors of work order events and public opinion events using the cosine similarity algorithm to obtain the second similarity. The average of the first and second similarities is selected as the current similarity. The current similarity is the similarity between the semantic vectors of alarm events, work order events, and public opinion events. The control module is used to generate a popularity index for the descriptive text of the public opinion event when the current similarity is greater than the preset similarity. When the popularity index is greater than the preset index, the module uses a sentiment recognition model to perform sentiment recognition on the descriptive text of the public opinion event to obtain sentiment tags. The module integrates the name of the public opinion platform, the descriptive text of the public opinion event, the sentiment tags, and the popularity index to obtain public opinion prompt content. The module concatenates the semantic vectors of the public opinion prompt content, the semantic vectors of the alarm event, and the semantic vectors of the work order event to obtain a fusion vector. The module processes the fusion vector through a large language model to generate a playback control strategy for IPTV live content.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the playback control method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the playback control method as described in any one of claims 1 to 7.

Images