Automatic processing method based on multi-dimensional fusion propagation digital intelligent management

Abstract

The invention discloses an automatic processing method based on multi-dimensional fusion propagation digital intelligent management. A super administrator creates tenants and sets IDs, a digital intelligent management and control system is constructed, and a tenant administrator adds channels and accounts needing to be monitored and implements a supervision process. A user logs in and authorizes in a code scanning or Token mode, then validity detection, classified storage, preliminary evaluation, video and audio file collection, data transmission, audio processing, voice-to-text conversion and the like are performed on an authorized account, and comprehensive storage of the data is ensured. Collected data are automatically stored after being subjected to duplication removal, recognition and other processing, and preparation is made for follow-up analysis. And performing positive and negative assessment by using the fusion propagation evaluation model, and performing archiving. A user can view personalized and multi-dimensional data reports and analysis results through multiple ways. The method can achieve the automatic examination of the contents of television and broadcast programs, can carry out the deep mining of program transmission evaluation indexes through big data analysis, and forms a transmission digital intelligent evaluation system.

Description

Technical Field

[0001] The present application relates to the field of radio, television and media technology, and specifically, to an automated processing method based on multi-dimensional integrated communication digital intelligence management. Background Art

[0002] Although existing technologies have made significant progress in the automated assessment of TV and radio program content and in-depth mining of program communication evaluation indicators, there are still some limitations that prevent these functions from being perfectly implemented in all cases:

[0003] First, in some regions or platforms, data collection may be limited by policies and regulations, technical compatibility, or lack of data openness. For example, some radio and television stations may not have fully implemented digital transformation, resulting in incomplete data records or difficulty in real-time acquisition. Second, even if the data is available, its quality and integrity may be a problem. For example, audience interaction data may be noisy due to factors such as network environment and user behavior differences, affecting the accuracy of analysis. In addition, for niche programs or time periods, interactive data is scarce and difficult to conduct in-depth analysis. Third, although big data analysis and AI technology are developing rapidly, they are still limited in understanding and parsing complex human emotions, cultural backgrounds, and language nuances. In particular, when dealing with contexts such as regional dialects, slang, humor, and sarcasm, existing technologies may find it difficult to achieve high-precision sentiment analysis and content understanding. Fourth, building and maintaining such a comprehensive evaluation system requires a lot of computing resources, senior talents, and continuous R&D investment. For small media organizations, this may be a considerable expense, limiting the application and popularization of technology. Fifth, when collecting and analyzing personal data for in-depth mining, it is easy to touch the red line of user privacy and data protection. There is a certain difficulty in balancing the needs of content assessment and protecting personal privacy, and relevant laws and regulations must be strictly observed.

[0004] Therefore, how to solve the above-mentioned technical problems and design a solution that can simultaneously complete the automated assessment of television and radio program content and the in-depth mining of program communication evaluation indicators has become one of the research directions of technical personnel in this field. Summary of the invention

[0005] The purpose of this application is to provide an automated processing method based on multi-dimensional integrated communication digital management in order to overcome the existing technical defects. Through services such as code scanning authorization, appointment collection, intelligent stripping, and voice-to-text conversion, it takes the fusion of multi-dimensional data such as video, graphics, and Internet communication as the core, initiates inventory assessment, and realizes diversified assessment management needs.

[0006] The purpose of this application is achieved through the following technical solutions:

[0007] In the first aspect, the present application proposes an automated processing method based on multi-dimensional fusion communication digital intelligence management, the method comprising:

[0008] Step S1: After the system super administrator enters the system backend management page to create the tenant name and ID to be assessed, a unified digital intelligence management and control service system is built;

[0009] The unified digital intelligence management and control service system covers the user's operation rights throughout the service life cycle, integrating core services including channel authorization, data collection, evaluation rules, and assessment models;

[0010] Step S2: After the tenant administrator enters the system backend management page, he opens the channel management interface, enters the channel and account to be monitored, and performs the account supervision process;

[0011] Step S3: After obtaining the permission, the user logs in to the Fusion Communication Digital Intelligence Management and Evaluation System, and executes the code scanning authorization service when the user is the account administrator. After the code scanning authorization service is completed, the Fusion Communication Digital Intelligence Management and Evaluation System adds the account to the monitoring queue for validity detection and performs initialization and classification storage, and generates evaluation points. In the code scanning authorization service, the new media channel data is authorized by scanning the QR code, and the traditional media channel data uses the token and site information for unified account authorization;

[0012] Step S4: assess the content dissemination evaluation of each channel media, and when the channel media is television, call the reservation collection and intelligent stripping system to obtain the audio and video files to be evaluated and the dissemination data;

[0013] Step S5: When the channel media is broadcast, the monitoring service is called to parse the audio file, and the speech-to-text service is initiated in real time, and the manuscript consistency is compared and then stored in the audio library;

[0014] Step S6: Automatically store the data from step S3 to step S5, using technologies including data deduplication, voice recognition, image recognition, text recognition, and consistency comparison to store the data;

[0015] Step S7: input the stored data into the integrated communication evaluation model, determine the comprehensive communication power assessment index after normalization processing through the entropy method, and then conduct positive assessment or negative assessment and store it in the positive assessment library or negative assessment library;

[0016] The integrated communication evaluation model includes content production index model, content quality index model, user interaction index model, user scale index model, topic operation index model, and provincial and central media forwarding index model.

[0017] Step S8: Users view data of different dimensions through multiple entrances including data portal, data cockpit, APP, H5 page, and mini program.

[0018] In a possible implementation, the account supervision process of step S2 includes:

[0019] Step S21, access and analyze the account registration information, crawl relevant web page content or call API interface, parse and extract core data, the core data includes registration number, registration subject name and website name;

[0020] Step S22: determining whether the registered account is officially registered based on the confirmation that the account has been registered;

[0021] Step S23: For the accounts that have been confirmed as officially registered, perform information capture and use intelligent recognition technology to analyze the activity to determine whether the account is a zombie account. If it is a zombie account, start the deregistration procedure. If the judgment result shows that the account is active and normal, activate the mechanism waiting for the user to scan the code for authorization.

[0022] In a possible implementation, the manuscript consistency comparison in step S5 includes:

[0023] Intelligent proofreading can identify and correct sensitive words, typos, misused idioms, improper punctuation, incorrect quantifiers, incorrect names of leaders and their job information, disordered order, grammatical flaws, English spelling, phonetic errors, and mixed number formats in the manuscript, and obtain the corrected manuscript;

[0024] An automated comparative analysis is performed on the corrected manuscripts for the same news content published on different media or platforms, followed by precise verification of the knowledge content, which includes but is not limited to: precise verification of important speeches, accuracy verification of personal names and positions, consistency check of geographical terms and institution names, application of professional terminology and proper nouns, compliance review of nouns related to laws and regulations, expression of basic common sense, and use of current political keywords.

[0025] In a possible implementation, the method further includes:

[0026] Construct a supervision service framework for the entire life cycle of the departments and accounts to be assessed, which integrates account location and status monitoring;

[0027] The framework includes an account location module, which is used to perform in-depth analysis and pattern mining on multi-source heterogeneous cross-platform data, identify the official attributes of accounts through unsupervised learning methods for influence assessment, and mine potential connections and influence paths between accounts.

[0028] The cross-platform account deduplication mechanism of the framework relies on a similarity measurement algorithm, comprehensively considers multi-dimensional features, accurately identifies and merges duplicate accounts, crawls through the Internet, sets a threshold for the account content release cycle, and warns of zombie accounts when the threshold is exceeded, cancels supervision and evaluation of the account, and enables account supervision if it is not a zombie account, and adds it to the assessment and evaluation queue.

[0029] In a possible implementation, the method further includes:

[0030] Automatically collect TV serial data and automatically extract key metadata by establishing a real-time API communication link with the TV serial system;

[0031] The integrity and consistency of metadata can be verified through the built-in verification mechanism, and the list of assessors can be matched to conduct accurate assessments of the staff involved in program production.

[0032] In a possible implementation, the method further includes:

[0033] Perform XML parsing on the received broadcast data stream, where the XML document complies with metadata standards, including PRSS and RDS;

[0034] The parsing process extracts key information, including but not limited to program ID, broadcast time, and channel ID;

[0035] The collection mechanism is implemented by integrating with cloud storage services and sending requests using the RESTful API interface.

[0036] The system schedules and configures recording tasks based on the parsed program information to ensure the complete recording of the target program. The recording process adopts an event-driven microservice architecture, in which each microservice is responsible for handling different types of recording requirements.

[0037] Based on the monitoring system, we introduced intelligent speech recognition and noise reduction technology optimized by deep neural network, which can maintain high-precision speech-to-text effect even in complex environments;

[0038] Synchronously analyze the speaker's emotional state through real-time sentiment analysis.

[0039] In a possible implementation, the fusion communication digital intelligence management and evaluation system includes a closed-door recycling mechanism, a data marking function, a data rollback and recovery mechanism, and an automatic cleanup function;

[0040] The closed-door recycling mechanism manages the timeliness and accuracy of assessment data. According to the preset assessment standards and real-time data monitoring, outdated or inaccurate data is automatically put into the recycling bin and not included in the assessment statistics.

[0041] The data marking function classifies and marks the recovered data in detail, recording the specific reasons, time and related evidence for the recovery;

[0042] The data rollback and recovery mechanism provides data recovery options, supports single data recovery and batch recovery configuration, and ensures that data can be restored accurately in special circumstances or when the system makes a misjudgment;

[0043] The automatic cleanup function intelligently manages data in the recycle bin, automatically deletes old data based on data tags and time thresholds, effectively controls resource usage, dynamically adjusts the cleanup strategy, and adaptively sets the cleanup cycle based on data type, size, and importance.

[0044] In one possible implementation, the unified digital management and control service system covers the user's main operating permissions throughout the entire service life cycle, integrates core services, and through centralized management, flexible configuration, intelligent analysis and other functions, from creating service instances to monitoring service operations, and then to data analysis and result output.

[0045] The above-mentioned main scheme of the present application and its further options can be freely combined to form multiple schemes, all of which are schemes that can be adopted and claimed for protection in the present application; and in the present application, (non-conflicting options) options and other options can also be freely combined. After understanding the scheme of the present application, those skilled in the art can understand that there are multiple combinations based on the prior art and common knowledge, all of which are technical schemes to be protected by the present application, and they are not exhaustively listed here.

[0046] The present application discloses an automated processing method based on multi-dimensional integrated communication digital intelligence management. After the system super administrator enters the system background management page to create the tenant name and ID to be assessed, a unified digital intelligence management and control service system is built. After the tenant administrator enters the system background management page, he opens the channel management interface, enters the channel and account to be monitored, and starts the account supervision process. The process starts with account supervision, through code scanning authorization, appointment collection, intelligent stripping, voice-to-text and other services, with multi-dimensional data integration such as video, graphics, and Internet communication as the core, to initiate warehousing assessment and realize diversified assessment management needs. On the basis of data assessment, the automated assessment of television and radio program content has been improved. It can realize the automated assessment of television and radio program content, and can also conduct in-depth mining of program transmission evaluation indicators through big data analysis to form a communication digital intelligence evaluation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0048] Figure 1 A flow chart of an automated processing method based on multi-dimensional fusion communication digital intelligence management proposed in an embodiment of the present application is shown.

[0049] Figure 2 A flow chart of the integrated communication digital intelligence management and evaluation system proposed in an embodiment of the present application is shown.

[0050] Figure 3 A schematic diagram of a unified digital intelligence management and control service system proposed in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0051] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0052] Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of this application.

[0053] The embodiment of this application proposes an automated processing method based on multi-dimensional fusion communication digital intelligence management. The system super administrator logs in to the default tenant background management, creates tenants to be assessed, configures the multi-dimensional intelligent control assessment service system, sets up tenant administrators, and logs in to the tenant management background and configures related processes. The process starts with account supervision, through code scanning authorization, appointment collection, intelligent stripping, voice-to-text and other services, with multi-dimensional data fusion such as video, graphics, and Internet communication as the core, initiates warehousing assessment, and realizes diversified assessment management needs, which will be described in detail below.

[0054] Please refer to Figure 1 , Figure 1 A flow chart of an automated processing method based on multi-dimensional fusion communication digital intelligence management proposed in an embodiment of the present application is shown, comprising the following steps:

[0055] Step S1: After the system super administrator enters the system backend management page to create the tenant name and ID to be assessed, a unified digital intelligence management and control service system is built;

[0056] The unified digital management and control service system covers the user's operating rights throughout the entire service life cycle, and integrates core services including channel authorization, data collection, evaluation rules, and assessment models.

[0057] The administrator first creates a new tenant to be assessed, defines the tenant's name and unique ID, and builds a comprehensive service management system that covers all key operation permissions of users during the service cycle to ensure that each link has corresponding permission control. In the core components of the service system, set access authentication methods for different channels (such as social media, traditional media) to ensure the security and legitimacy of data sources. Configure multiple data collection methods to meet the data acquisition requirements of different media types (video, audio, text, etc.). Design and implement flexible and changeable assessment rules, and adjust the scoring standards and weights according to different business scenarios. Integrate evaluation models such as content consistency and communication effect to analyze data quality and communication effectiveness in an automated manner.

[0058] The service management system provides a centralized management panel, allowing administrators to easily monitor the status of service instances, from creation, configuration to operation. Using intelligent analysis tools, the collected data is deeply mined, and valuable analysis reports are output to guide business decisions.

[0059] The system supports highly customized configurations, and users can adjust service parameters and workflows according to actual needs. Through automated workflow design, seamless integration of data processing, analysis, and feedback is achieved.

[0060] Step S2: After the tenant administrator enters the system backend management page, he / she opens the channel management interface, enters the channel and account to be monitored, and starts the account supervision process.

[0061] The account supervision process of step S2 includes:

[0062] Step S21, access and analyze the account registration information, crawl relevant web page content or call API interface, parse and extract core data, the core data includes registration number, registration subject name and website name;

[0063] Step S22: determining whether the registered account is officially registered based on the confirmation that the account has been registered;

[0064] Step S23: For the accounts that have been confirmed as officially registered, perform information capture and use intelligent recognition technology to analyze the activity to determine whether the account is a zombie account. If it is a zombie account, start the deregistration procedure. If the judgment result shows that the account is active and normal, activate the mechanism waiting for the user to scan the code for authorization.

[0065] When the tenant administrator enters the specified channel and account in the channel management interface, the system will automatically initiate a request to the relevant platform or public database to capture the record information of the account. This step is to verify whether the account has completed the online record in accordance with national regulations. During the process, the core information such as the record number, the name of the record subject and the website (or account) name will be extracted from the data returned by the web page or API.

[0066] The extracted filing information, especially the account name, is compared with the official account of the subject on major social media platforms (such as Weibo, Douyin, etc.). The purpose is to ensure that the account name used by the same subject on multiple platforms remains consistent and meets the regulatory requirements of government new media. The system can assist in identifying and recommending a standard official name for the administrator to review and confirm.

[0067] For accounts that have been confirmed and registered with the same name, the system will further perform Internet data crawling and use intelligent algorithms to analyze the account's activities, such as posting frequency, interaction volume and other indicators. Intelligent recognition technology can effectively identify "zombie accounts" with extremely low activity or no effective interaction. Once confirmed as a zombie account, the system will automatically generate a notification to prompt the affiliated unit to consider canceling the account. After passing the verification process and confirming that the account is not in a zombie state, the system enters the user authorization stage, waiting for the actual operator of the account to scan the code to grant the system the necessary data access rights.

[0068] Step S3: After obtaining the permission, the user logs in to the Fusion Communication Digital Intelligence Management Evaluation System. If the user is an account administrator, the scanning authorization service is investigated. In the scanning authorization service, the new media channel data is bound and authorized by scanning the QR code, and the traditional media channel data uses the token and site information for unified account authorization. After the scanning authorization service is completed, the Fusion Communication Digital Intelligence Management Evaluation System adds the account to the monitoring queue for validity detection, and at the same time classifies it into the database, and generates an evaluation score through the Fusion Communication evaluation model;

[0069] Figure 2The flowchart of the Fusion Communication Digital Intelligence Management and Evaluation System proposed in the embodiment of the present application is shown, and a user with permission logs in to the Fusion Communication Digital Intelligence Management and Evaluation System. It is assumed here that the user has obtained the corresponding access rights and can enter the advanced function area of ​​the system. For new media channels (such as WeChat public accounts, Weibo, Douyin, etc.), the system provides a scan code authorization service. The account administrator needs to scan the QR code with a mobile phone to bind the new media account of the individual or organization to the system. This action is equivalent to authorizing the system to access part or all of the account's data.

[0070] Unlike new media channels, traditional media channels (such as news websites and online platforms of radio and television stations) usually use Token (access token) and site information for authorization. The administrator needs to enter the Token and the corresponding site information in the system. The system communicates with the media channel server through these credentials to achieve data docking and permission granting.

[0071] The system has a built-in mechanism to detect the validity of authorization. Once it is found that the authorization is about to expire or has expired, the system will promptly remind you to re-authorize to ensure data continuity and integrity and avoid data interruptions caused by authorization issues. After successful authorization, the accounts of new media and traditional media channels will be automatically added to the monitoring queue. The system begins to regularly capture content data on these accounts, including videos, texts, reading counts, likes, etc., and store them in the database according to content type.

[0072] Based on the collected quantitative data, the system runs a pre-built integrated communication evaluation model. This model can comprehensively consider multiple dimensions (such as content quality, audience participation, and communication breadth) to generate an evaluation score for each account or content. This is not only applicable to the evaluation of the communication effect of the official account itself, but also to monitoring the reprint performance of its content in other media channels.

[0073] Step S4: assess the content dissemination evaluation of the media reprinted in each channel. If the channel is determined to be television, call the reservation collection and intelligent stripping system to obtain the audio and video files to be evaluated and the dissemination data.

[0074] Authorized users log in to the integrated communication digital management system of their tenants. The assessment point that users are concerned about is the reprinting of Zhejiang Radio and Television's original video content in central media, which is regarded as an important plus point for communication effect. For TV programs, the system will automatically trigger the reservation collection and intelligent stripping services. This means that the system not only needs to arrange the collection of specific TV program content, but also needs to process the collected audio and video files through intelligent technology, such as cutting them into manageable small clips and automatically transcoding them into a format suitable for analysis.

[0075] For the central media videos that need to be monitored, the system will transcode them into low-bitrate versions and store them in the analysis and comparison library. In addition, the system will compare the low-bitrate videos with the original material library of the "China Blue Cloud" content center. If the comparison results show that the content consistency between the two is higher than 66%, the video will be deemed to be a reprint of Zhejiang Radio and Television's original content. At this time, the system will record the original broadcast time, name and other metadata of the reprinted video into the metadata of the corresponding original material in the content center to track and record the reprint situation.

[0076] The system will push low-code videos identified as reprints to the assessment database, where they will be included in the integrated communication evaluation model to further analyze communication efficiency, coverage, etc., thereby forming a comprehensive evaluation score.

[0077] Step S5: If it is determined that the channel is broadcast, the monitoring service is called to parse the audio file, and the speech-to-text service is initiated in real time. After the manuscript consistency comparison is performed, it is stored in the audio library.

[0078] Authorized users log in to the Fusion Communication Digital Intelligence Management and Evaluation System and operate as an account administrator. The system starts monitoring services for broadcast programs and continuously captures audio streams. Through advanced speech-to-text technology, audio content is converted into text format in real time and saved in the audio library. This step allows the audio content to be further analyzed and processed.

[0079] For example, if a China Blue Cloud tenant needs to assess the broadcast reprint of its original radio program in the central media, the system will extract relevant manuscripts from the content library and conduct a detailed comparison with the central media broadcast manuscripts obtained through the monitoring service. The comparison process uses the semantic understanding ability of the big model to not only determine the consistency of the content, but also identify the specific differences.

[0080] The manuscript consistency comparison in step S5 includes:

[0081] Intelligent proofreading can identify and correct sensitive words, typos, misused idioms, improper punctuation, incorrect quantifiers, incorrect names of leaders and their job information, disordered order, grammatical flaws, English spelling, phonetic errors, and mixed number formats in the manuscript, and obtain the corrected manuscript;

[0082] The corrected manuscripts are automatically compared and analyzed for the same news content published on different media or platforms, followed by precise verification of the knowledge content. The precise verification includes but is not limited to: precise verification of speeches, accuracy verification of personal names and positions, consistency check of geographical terms and institution names, application of professional terminology and proper nouns, compliance review of nouns related to laws and regulations, expression of basic common sense, and use of current affairs keywords.

[0083] During the manuscript consistency comparison process, the system performs intelligent proofreading, which is a deep text review function that can automatically detect and correct various types of errors, including but not limited to sensitive words, typos, punctuation errors, grammatical errors, and misuse of proper nouns. In addition, for knowledge errors, the system will also proofread important speech quotations, names and positions, geographical terms, institutional terms, current affairs vocabulary, etc. to ensure the accuracy and compliance of the content. Once the system confirms that the content of the two manuscripts is highly consistent, the metadata such as the broadcast time and name of the original material involved will be recorded in the metadata of the relevant manuscripts in the content library, and the manuscripts that have passed the review will be pushed to the assessment library as a basis for evaluation and bonus points.

[0084] The following are examples of large model prompt words:

[0085] prompt_content={

[0086] "Instructions": "Please compare the text contents in the following "document data" and point out their consistency and differences in terms of time, place, characters, etc.",

[0087] "Document Data": Text 1: Title 1 Text 2: Title 2 Text 3: Title 3

[0088] "Output format":"Please list the comparison results of each dimension separately. The specific output is as follows: Dimension 1: Same or Different, Specific Difference: Fill in None or Difference Content, Dimension 2: Same or Different, Specific Difference: Fill in None or Difference Content"}.

[0089] Step S6: Automatically store the data from step S3 to step S5 in the database by using technologies including data deduplication, speech recognition, image recognition, text recognition, and consistency comparison.

[0090] The data of step S3, step S4, and step S5 are automatically stored in the database. Through data deduplication, voice recognition, image recognition, text recognition, consistency comparison and other technologies, the data of channels, columns, authors, manuscripts, serial lists, videos, audios, reading numbers, likes numbers, reposts numbers, and comments numbers are stored in the database.

[0091] Step S7: input the stored data into the integrated communication evaluation model, determine the comprehensive communication ability assessment index after normalization through the entropy method, then conduct positive assessment or negative assessment and store it in the positive assessment library or negative assessment library.

[0092] The integrated communication evaluation model includes content production index, content quality index, user interaction index, user scale index, topic operation index, and provincial and central media forwarding index model.

[0093] The comprehensive evaluation model includes content production index, content quality index, user interaction index, user scale index, topic operation index, and provincial and central media forwarding index. Each index is standardized through the entropy method to ensure the comparability of indicators in different dimensions and eliminate the impact of dimensions. Finally, the communication power assessment index is calculated comprehensively. This index reflects the overall effectiveness of content dissemination and will be stored in the positive assessment library as a basis for positive incentives.

[0094] The negative assessment database is used to record data that does not meet the dissemination standards or performs poorly. During the entire assessment cycle, the system implements strict data management measures, including closed-door recycling (i.e. closing the data entry after the assessment cycle to prevent data tampering), data marking (labeling the data with status, such as abnormal, pending review, etc.), data rollback (erroneous data can be traced and restored to the previous state), automatic cleaning (regularly deleting no longer needed or expired data), etc., to ensure the accuracy of the data and the fairness of the assessment.

[0095] Step S8: Users can view data of different dimensions through multiple entrances such as data portal, data cockpit, APP, H5 page, mini program, etc.

[0096] The method further includes:

[0097] Construct a supervision service framework for the entire life cycle of the departments and accounts to be assessed, which integrates account location and status monitoring;

[0098] The framework includes an account location module, which is used to perform in-depth analysis and pattern mining on multi-source heterogeneous cross-platform data, identify official account attributes through influence assessment unsupervised learning methods, and mine potential connections and influence paths between accounts.

[0099] The cross-platform account deduplication mechanism of the framework relies on a similarity measurement algorithm, comprehensively considers multi-dimensional features, accurately identifies and merges duplicate accounts, crawls through the Internet, sets a threshold for the account content release cycle, and warns of zombie accounts when the threshold is exceeded, cancels supervision and evaluation of the account, and enables account supervision if it is not a zombie account, and adds it to the assessment and evaluation queue.

[0100] The system processes diverse data from different platforms and uses deep analysis technology to mine hidden information in the data. It uses machine learning algorithms to automatically identify the official attributes of accounts without manual pre-labeling, and identifies account influence and official certification possibilities through pattern mining. It analyzes the interactions between accounts, reveals the underlying network structure, and eliminates "zombie accounts" (accounts with no actual activities or malicious manipulation) and fake official media accounts to ensure the purity of assessment and evaluation.

[0101] By comprehensively evaluating multi-dimensional features such as account name, published content, and interactive behavior, algorithms are used to identify duplicate or highly similar accounts to achieve accurate deduplication and merging. Dynamic thresholds are set to monitor account activity. If the publishing frequency is lower than the preset threshold, the system will issue an early warning, identify it as a possible zombie account, and stop monitoring and evaluating it. On the contrary, active accounts will enter the next step of the supervision process.

[0102] Once an account is confirmed to be active and legal, the system will include it in the formal supervision system and arrange for it to be included in the assessment and evaluation sequence. Accounts that meet the requirements will be queued according to established rules and prepared to undergo comprehensive assessments in terms of content quality, interactivity, influence, etc., in order to determine the work performance evaluation of the department or individual.

[0103] The method further includes:

[0104] Automatically collect TV serial data and automatically extract key metadata by establishing a real-time API communication link with the TV serial system;

[0105] The integrity and consistency of metadata can be verified through the built-in verification mechanism, and the list of assessors can be matched to conduct accurate assessments of the staff involved in program production.

[0106] The API communication link established with the TV serial single system ensures the real-time data collection. It automatically extracts metadata such as program name, script, broadcast time, and reporter name, and then the built-in verification mechanism checks the integrity and consistency of the metadata. After that, the system can accurately match the list of staff involved in the program production to achieve personalized and accurate performance evaluation.

[0107] For videos produced independently by China Blue Cloud, the original materials are directly obtained through the program code. RNN is used to process video time series data, understand the continuity of video content, accurately locate key information fragments, and perform intelligent strip storage. The corresponding series order is then captured in real time through the Internet. After matching the program, the video is added to the reservation collection queue and the collection and stripping process is automatically executed. Using deep learning NLP technology, the audio or subtitles in the video are converted into text, and semantic understanding is performed to automatically identify information such as themes, characters, and locations, enhancing the searchability and analytical value of the content. The identified video clips are classified through clustering algorithms to form a structured video content index. The processed video clips and metadata are integrated into the fusion communication digital intelligence management and evaluation system, which relies on a distributed storage architecture and supports high-concurrency operations.

[0108] The method further includes:

[0109] Perform XML parsing on the received broadcast data stream, where the XML document complies with specific metadata standards, including PRSS and RDS;

[0110] The parsing process extracts key information, including but not limited to program ID, broadcast time, and channel ID;

[0111] The collection mechanism is implemented by integrating with cloud storage services and sending requests using the RESTful API interface.

[0112] The system schedules and configures recording tasks based on the parsed program information to ensure the complete recording of the target program. The recording process adopts an event-driven microservice architecture, in which each microservice is responsible for handling different types of recording requirements.

[0113] Based on the monitoring system, we introduced intelligent speech recognition and noise reduction technology optimized by deep neural network, which can maintain high-precision speech-to-text effect even in complex environments;

[0114] Synchronously analyze the speaker's emotional state through real-time sentiment analysis.

[0115] The system continuously monitors broadcast programs to ensure comprehensive coverage of content. The received audio data is accompanied by metadata in XML format, which complies with industry standards such as PRSS and RDS. Through XML parsing, the system automatically extracts key information such as program ID, broadcast time, channel ID, etc., which is used as the basis for arranging collection tasks. Using RESTful API to connect with cloud storage services, the system intelligently schedules recording tasks based on the parsed program information. Whether it is live streaming or pre-recorded content, it can automatically configure and complete the collection.

[0116] Based on the event-driven microservice architecture design, the system can flexibly respond to different recording needs, such as live recording or pre-recorded content processing, and each microservice focuses on a specific task. The integrated deep neural network optimized speech recognition and noise reduction technology ensures high-precision speech-to-text conversion in complex environments, while supporting multiple languages ​​and dialects.

[0117] Through real-time sentiment analysis, the system can analyze the speaker's emotions and provide in-depth feedback on program production. Keyword monitoring and early warning functions strengthen content security control, and personalized voice feature matching ensures accurate confirmation of the identity of participants. Using semantic understanding and keyword extraction technology, content tags are automatically generated, and a sophisticated classification system and retrieval mechanism are established for radio programs. The processed audio files and data information are automatically stored in the Yundong Media Content Library, and an automatic data push mechanism is set up between the content library and the assessment database. With the help of message queues and event-driven architecture, the system can perceive data in real time.

[0118] The Fusion Communication Digital Intelligence Management Evaluation System includes a closed-door recycling mechanism, data tagging function, data rollback and recovery mechanism, and automatic cleanup function;

[0119] The closed-door recycling mechanism manages the timeliness and accuracy of assessment data. According to the preset assessment standards and real-time data monitoring, outdated or inaccurate data is automatically put into the recycling bin and not included in the assessment statistics.

[0120] The data marking function classifies and marks the recovered data in detail, recording the specific reasons, time and related evidence for the recovery;

[0121] The data rollback and recovery mechanism provides data recovery options, supports single data recovery and batch recovery configuration, and ensures that data can be restored accurately in special circumstances or when the system makes a misjudgment;

[0122] The automatic cleanup function intelligently manages data in the recycle bin, automatically deletes old data based on data tags and time thresholds, effectively controls resource usage, dynamically adjusts the cleanup strategy, and adaptively sets the cleanup cycle based on data type, size, and importance.

[0123] Closed-door recycling ensures the freshness and accuracy of assessment data, and processes data from different sources according to preset rules. For example, TV and radio data only consider the first broadcast ratings; new media data (such as Douyin and Kuaishou videos) are valid within seven days after release; hot search data is recycled in real time as the list changes; social hot news is evaluated dynamically based on progress, and relevant data is recycled immediately once the information is corrected or the event is reversed. Data tagging carefully classifies and annotates the recycled data to facilitate later auditing and tracing. This includes recording the reasons for data recycling, release time, ranking changes, etc., especially for social hot spots, tracking their development in detail, saving related links and content overviews, and enhancing transparency and responsibility traceability. Data rollback provides flexibility and error correction mechanisms, allowing the recovery of data that has been recycled due to special reasons or misjudgment. It is divided into designated data recovery and batch parameter configuration recovery to adapt to different scenario requirements and ensure the fairness and accuracy of assessment results. Automatic cleanup: Manage the storage resources of the recycle bin, dynamically adjust the cleanup strategy according to data characteristics (type, size, importance), automatically delete the recycled data that reaches a specific threshold, optimize storage space utilization, and also ensure system operation efficiency.

[0124] Through the program management interface, editors need to submit basic information about the program, including the name, type, attributes, broadcast plan, and specific requirements of superior supervision. The system has a built-in illegal content detection module, which can start the early warning mechanism at the beginning of program production. By analyzing scripts, materials, etc., it predicts potential illegal risks and assists the production team to make corrections in advance to avoid illegal broadcasts. In addition, AI technology is used to conduct in-depth analysis of the content, and the identified violations are classified into preset categories, such as "hidden vulgar content" and "misleading information". Then, according to the severity of each type of violation, the system has detailed scoring standards. For different violation categories, customized evaluation models are applied to automatically deduct points to ensure the accuracy and fairness of the assessment. After the detection and evaluation of illegal content is completed, the system automatically generates a detailed assessment report. The report not only records the basic information of the program, the type of violation and the specific content, but also lists the deduction details in detail. The report will be directly pushed to the negative assessment library for easy archiving and review.

[0125] Figure 3 A schematic diagram of the unified digital intelligence management and control service system proposed in an embodiment of the present application is shown, and the steps are: creating a service instance, activating a service instance, binding the service instance to a tenant, selecting a channel authorization service instance, selecting a data collection service instance, selecting an evaluation rule service instance, and selecting an evaluation model service instance. In selecting the channel authorization service instance, authorization is performed through scanning codes of new media such as WeChat, Weibo, and Douyin, and token authorization mechanisms for traditional media such as television and radio, as well as the inclusion of television and radio source information.

[0126] The selected data collection service instance includes video / audio scheduled collection (offline collection) and API integration (real-time collection). The selected evaluation rule service instance includes script rules, configuration files, and visualization rules.

[0127] The unified digital management and control service system covers the user's main operation permissions throughout the entire service life cycle, integrates core services, and through centralized management, flexible configuration, intelligent analysis and other functions, from creating service instances to monitoring service operations, and then to data analysis and result output.

[0128] The user first creates a service instance, paying special attention to channel authorization, which is achieved by integrating the code scanning authorization of new media (such as WeChat, Weibo, and Douyin) and the token verification of traditional media (TV and radio). The microservice architecture is used to process the authorization logic to ensure the safe storage of incremental data, and the real-time and continuity of the authorization status is maintained through distributed deployment, which solves the problem of authorization timeliness and prevents data collection interruption.

[0129] The system supports a variety of data collection methods and has designed special data capture protocols for different media (video, audio, image, text). For example, TV programs are processed through specific video acquisition protocols, while broadcasts use special audio acquisition technology. At the same time, it can also collect and analyze interactive data (such as reading volume, likes, comments), and integrate third-party data sources through APIs, achieving the breadth and diversity of data sources.

[0130] Users can customize assessment rules through YAML files, scripts or visual interfaces, and easily set indicators, weights and scoring standards. This approach greatly improves the flexibility and ease of use of rule configuration, ensuring that the assessment system can closely meet the needs of different business scenarios.

[0131] The engine integrates model instances for content assessment and communication assessment. Content assessment focuses on consistency analysis and uses AI technology to automatically judge content quality; communication assessment evaluates communication effects through algorithm models. The combination of the two provides data support for decision-making. Users can view and manage service instance information and parameter configurations, monitor service operation status in real time, analyze logs, and ensure stable and efficient services. This instant feedback mechanism is conducive to rapid response to problems and continuous optimization of service performance.

[0132] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0133] First, based on data assessment, the automated assessment of TV and radio program content has been improved. It can realize the automated assessment of TV and radio program content, and can also conduct in-depth mining of program evaluation indicators through big data analysis to form a digital communication evaluation system. For example, by analyzing the audience interaction after the program is broadcast (such as comments, bullet screen, etc.), it is possible to obtain audience feedback on the program in real time, and then guide the program production team to make targeted adjustments.

[0134] Second, in addition to focusing on the assessment of content production itself, we support the accurate identification, positioning, and dissemination quantification of program content after it is reproduced on the Internet, and supplement it with the content production assessment.

[0135] Third, it supports multi-tenant deployment, adopts microservice architecture, and is quickly deployed and launched to achieve data and permission isolation across tenants. The differentiated needs of different user groups were taken into consideration at the beginning of the system design. It supports independent data storage and permission management in a multi-tenant mode. Different users (such as TV stations, online video platforms, etc.) can run their own independent service instances on the same system platform without affecting each other, while ensuring the security and privacy of the data. 4. Decouple each service based on the business process, realize independent deployment of various capabilities through the microservice model, and automate through the multi-dimensional intelligent control assessment service system. The system decomposes complex business processes into multiple independent service modules, each module focuses on completing specific functions, such as scanning code authorization, appointment collection, etc. These services communicate through APIs, making the system highly flexible and scalable. When you need to add new functions or improve existing functions, you only need to adjust the multi-dimensional intelligent control assessment service system without affecting the stability of the entire system.

[0136] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. An automated processing method based on multi-dimensional fusion communication digital intelligence management, characterized in that: The method comprises: Step S1: After the system super administrator enters the system backend management page to create the tenant name and ID to be assessed, a unified digital intelligence management and control service system is built; The unified digital intelligence management and control service system covers the user's operation rights throughout the service life cycle, integrating core services including channel authorization, data collection, evaluation rules, and assessment models; Step S2: After the tenant administrator enters the system backend management page, he opens the channel management interface, enters the channel and account to be monitored, and performs the account supervision process; Step S3: After obtaining the permission, the user logs in to the Fusion Communication Digital Intelligence Management and Evaluation System, and executes the code scanning authorization service when the user is the account administrator. After the code scanning authorization service is completed, the Fusion Communication Digital Intelligence Management and Evaluation System adds the account to the monitoring queue for validity detection and performs initialization and classification storage, and generates evaluation points. In the code scanning authorization service, the new media channel data is authorized by scanning the QR code, and the traditional media channel data uses the token and site information for unified account authorization; Step S4: assess the content dissemination evaluation of each channel media, and when the channel media is television, call the reservation collection and intelligent stripping system to obtain the audio and video files to be evaluated and the dissemination data; Step S5: When the channel media is broadcast, the monitoring service is called to parse the audio file, and the speech-to-text service is initiated in real time, and the manuscript consistency is compared and then stored in the audio library; Step S6: Automatically store the data from step S3 to step S5, using technologies including data deduplication, voice recognition, image recognition, text recognition, and consistency comparison to store the data; Step S7: input the stored data into the integrated communication evaluation model, determine the comprehensive communication power assessment index after normalization processing through the entropy method, and then conduct positive assessment or negative assessment and store it in the positive assessment library or negative assessment library; The integrated communication evaluation model includes content production index model, content quality index model, user interaction index model, user scale index model, topic operation index model, and provincial and central media forwarding index model; Step S8: Users view data of different dimensions through multiple entrances including data portal, data cockpit, APP, H5 page, and mini program.

2. The automated processing method according to claim 1, characterized in that: The account supervision process of step S2 includes: Step S21, access and analyze the account registration information, crawl relevant web page content or call API interface, parse and extract core data, the core data includes registration number, registration subject name and website name; Step S22: determining whether the registered account is officially registered based on the confirmation that the account has been registered; Step S23: For the accounts that have been confirmed as officially registered, perform information capture and use intelligent recognition technology to analyze the activity to determine whether the account is a zombie account. If it is a zombie account, start the deregistration procedure. If the judgment result shows that the account is active and normal, activate the mechanism waiting for the user to scan the code for authorization.

3. The automated processing method according to claim 1, characterized in that: The manuscript consistency comparison in step S5 includes: Intelligent proofreading can identify and correct sensitive words, typos, misused idioms, improper punctuation, incorrect quantifiers, incorrect names of leaders and their job information, disordered order, grammatical flaws, English spelling, phonetic errors, and mixed number formats in the manuscript, and obtain the corrected manuscript; An automated comparative analysis is performed on the corrected manuscripts for the same news content published on different media or platforms, followed by a precise verification of the knowledge content, which includes but is not limited to: precise verification of speeches, accuracy verification of personal names and positions, consistency check of geographical terms and institution names, application of professional terminology and proper nouns, compliance review of nouns related to laws and regulations, expression of basic common sense, and use of current political keywords.

4. The automated processing method according to claim 1, characterized in that: The method further comprises: Build a supervision service framework for the entire life cycle of the departments and accounts to be assessed, which integrates account location and status monitoring; The framework includes an account location module, which is used to perform in-depth analysis and pattern mining on multi-source heterogeneous cross-platform data, identify official account attributes through unsupervised learning methods for influence assessment, and mine the associations and influence paths between accounts; The cross-platform account deduplication mechanism of the framework relies on a similarity measurement algorithm, comprehensively considers multi-dimensional features, accurately identifies and merges duplicate accounts, crawls through the Internet, sets a threshold for the account content release cycle, and warns of zombie accounts when the threshold is exceeded, cancels supervision and evaluation of the account, and enables account supervision if it is not a zombie account, and adds it to the assessment and evaluation queue.

5. The automated processing method according to claim 1, characterized in that: The method further comprises: Automatically collect TV serial data and automatically extract key metadata by establishing a real-time API communication link with the TV serial system; The integrity and consistency of metadata can be verified through the built-in verification mechanism, and the list of assessors can be matched to conduct accurate assessments of the staff involved in program production.

6. The automated processing method according to claim 1, characterized in that: The method further comprises: Perform XML parsing on the received broadcast data stream, where the XML document complies with metadata standards, including PRSS and RDS; The parsing process extracts key information, including but not limited to program ID, broadcast time, and channel ID; The collection mechanism is implemented by integrating with cloud storage services and sending requests using the RESTful API interface; The system schedules and configures recording tasks based on the parsed program information to ensure the complete recording of the target program. The recording process adopts an event-driven microservice architecture, in which each microservice is responsible for handling different types of recording requirements. Based on the monitoring system, we introduced intelligent speech recognition and noise reduction technology optimized by deep neural network, which can maintain high-precision speech-to-text effect even in complex environments; Synchronously analyze the speaker's emotional state through real-time sentiment analysis.

7. The automated processing method according to claim 1, characterized in that: The Fusion Communication Digital Intelligence Management Evaluation System includes a closed-door recycling mechanism, a data marking function, a data rollback and recovery mechanism, and an automatic cleanup function; The closed-door recycling mechanism manages the timeliness and accuracy of assessment data. According to the preset assessment standards and real-time data monitoring, outdated or inaccurate data is automatically put into the recycling bin and not included in the assessment statistics. The data marking function classifies and marks the recovered data in detail, recording the specific reasons, time and related evidence for the recovery; The data rollback and recovery mechanism provides data recovery options, supports single data recovery and batch recovery configuration, and ensures that data can be restored accurately in special circumstances or when the system makes a misjudgment; The automatic cleanup function intelligently manages data in the recycle bin, automatically deletes old data based on data tags and time thresholds, effectively controls resource usage, dynamically adjusts the cleanup strategy, and adaptively sets the cleanup cycle based on data type, size, and importance.

8. The automated processing method according to claim 1, wherein: The unified digital management and control service system covers the user's main operation permissions throughout the entire service life cycle, integrating core services, from creating service instances to monitoring service operations, to data analysis and result output.

Images