New media operation optimization method and system based on data closed loop

Abstract

The invention discloses a new media operation optimization method and system based on a data closed loop. The system comprises a data acquisition module, a three-dimensional label system module, an intelligent analysis module and a closed loop optimization module. The data acquisition module synchronizes core operation data through an API, and captures unstructured comment data through an RPA; the three-dimensional tag system module is used for carrying out tagging processing on platform characteristics, content forms and user behaviors on the data; the intelligent analysis module generates and outputs optimization suggestions through content diagnosis, flow prediction and strategy; and the closed-loop optimization module feeds back an analysis result to an external content generation system, and new content data is transmitted back to form a closed loop. According to the system, the problems of operation data dispersion, insufficient analysis and optimization lag are solved, automatic data acquisition, structured analysis and dynamic optimization are realized, and the operation efficiency and the traffic conversion effect are improved.

Description

Technical Field

[0001] This invention belongs to the field of data processing technology, and in particular relates to a new media operation optimization method and system based on data closed loop. Background Technology

[0002] With the development of new media technologies, short video and live streaming platforms have become important channels for new farmers to promote agricultural products and expand sales channels. However, new farmers often face technical challenges in managing and optimizing operational data during new media operations, specifically as follows:

[0003] First, operational data collection is fragmented and lacks uniformity in format. New farmers' accounts are often distributed across multiple new media platforms, each with different data interfaces and statistical standards. This results in core data such as views, interactions, and user behavior being scattered across the backend systems of different platforms. Currently, new farmers largely rely on manual recording or the platform's built-in rudimentary statistical functions, making it difficult to achieve centralized collection and integration of cross-platform data. This leads to low data acquisition efficiency and a high risk of omissions.

[0004] Secondly, the data lacks structured analysis capabilities. The collected operational data includes both structured metrics such as play counts and likes, and unstructured information such as user comments and feedback. Current technologies lack systematic tagging, classification, and in-depth analysis methods for this data, making it difficult to effectively distinguish the characteristics of different platforms, the suitability of content formats, and the conversion patterns of user behavior. This makes it difficult for new farmers to identify the strengths and weaknesses of their content from the data.

[0005] Secondly, operational strategy optimization lags behind and lacks a closed-loop mechanism. Due to the inefficiency of data collection and analysis, new farmers often can only adjust operational strategies based on experience, unable to optimize content format, release timing, or conversion path in a timely manner based on data feedback. Even when strategies are adjusted, it is difficult to evaluate the optimization effect through data comparison, failing to form a complete closed loop of "data collection-analysis-optimization-verification," resulting in low operational efficiency and poor traffic conversion.

[0006] Therefore, there is an urgent need for a technical solution that can automatically collect, structure, and optimize operational data to solve the data management problems faced by new farmers in new media operations and improve the pertinence and effectiveness of operational strategies. Summary of the Invention

[0007] In view of this, the present invention aims to propose a new media operation optimization method and system based on data closed loop, so as to at least solve one of the problems in the background art.

[0008] To achieve the above objectives, the technical solution of the present invention is implemented as follows:

[0009] A new media operation optimization system based on data closed loop includes:

[0010] The data acquisition module is used to synchronize core operational data of at least one new media platform through API interface, and to capture unstructured comment data related to operational content in the platform through RPA technology. The core operational data includes play count, completion rate, likes, reposts, favorites, comments, and three-second completion rate.

[0011] A three-dimensional tagging system module, connected to the data acquisition module, is used to tag the collected core operational data and unstructured comment data. The three-dimensional tagging system includes platform characteristic tags, content format tags, and user behavior tags. The platform characteristic tags are used to distinguish the attribute characteristics of different new media platforms, the content format tags are used to mark the presentation type of operational content, and the user behavior tags are used to identify the user's behavioral path from content interaction to conversion.

[0012] The intelligent analysis module, connected to the three-dimensional tag system module, includes a content diagnosis unit, a traffic prediction unit, and a strategy generation unit. The content diagnosis unit is used to divide the operational content into preset performance quadrants and output targeted optimization suggestions. The traffic prediction unit is used to use an LSTM model to predict the key time nodes for the operational content to enter different levels of traffic pools. The strategy generation unit is used to automatically trigger a content format switching instruction when the interaction rate of the operational content is lower than a preset threshold for a preset number of consecutive preset days.

[0013] The closed-loop optimization module is connected to both the intelligent analysis module and the external content generation system. It is used to feed back the optimization suggestions, traffic prediction results and strategy instructions output by the intelligent analysis module to the external content generation system, so as to drive the external content generation system to iteratively generate new operational content, and to send the operational data of the newly generated operational content back to the data acquisition module, thus forming a data closed loop.

[0014] Furthermore, the data acquisition module includes a structured data interface unit and an unstructured data capture unit. The structured data interface unit is used to obtain the core operational data by calling the open API of the new media platform and convert it into a structured data format.

[0015] The unstructured data crawling unit is used to simulate user operation behavior through RPA robots, crawl text comment data in the comment section, and convert audio comment data into text data through ASR technology.

[0016] Furthermore, in the three-dimensional tag system module, the platform characteristic tags include entertainment attribute tags, social attribute tags, and information attribute tags, which are used to characterize the content distribution preferences of different platforms;

[0017] The content format tags include knowledge popularization tags, product review tags, story narrative tags, and interactive Q&A tags, which are used to distinguish the creation type of operational content;

[0018] The user behavior tags include click tags, favorite tags, forwarding tags, and purchase tags, which are used to construct a user behavior sequence model through a hidden Markov model.

[0019] Furthermore, in the content diagnosis unit of the intelligent analysis module, the preset performance quadrants include a high-playback-high-conversion quadrant, a high-playback-low-conversion quadrant, a low-playback-high-potential quadrant, and a low-playback-low-conversion quadrant. For operational content in the high-playback-low-conversion quadrant, the optimization suggestions include adjusting the conversion link nodes in the content to shorten the user's jump steps from interaction to purchase. For operational content in the low-playback-high-potential quadrant, the optimization suggestions include replacing the content cover with a comparison image containing visual guidance elements.

[0020] Furthermore, in the traffic prediction unit of the intelligent analysis module, the different levels of traffic pools include a 500-person exposure pool, a 5000-person exposure pool, and a 50000-person exposure pool.

[0021] The input features of the LSTM model include publication time, content tags, and the initial 30-minute playback volume of similar historical content. The output is the predicted time point for entering each level of traffic pool, and the model generates publication time adjustment suggestions based on the prediction results.

[0022] Furthermore, in the strategy generation unit of the intelligent analysis module, the preset number of days is 3 days, and the preset threshold is 85% of the average interaction rate of the same period in history;

[0023] The content format switching instructions include splitting long videos into short video segments of preset duration and converting text and image content into animated demonstrations. The instructions also include supplementary material information generated by calling external enterprise knowledge bases based on RAG technology.

[0024] Furthermore, this solution discloses a new media operation optimization method based on data closed loop, including the following steps:

[0025] S1. Data Collection: Synchronize core operational data of at least one new media platform through API interface, and capture unstructured comment data related to operational content in the platform through RPA technology. The core operational data includes play count, completion rate, likes, reposts, favorites, comments, and three-second completion rate.

[0026] S2. Tagging Processing: The collected core operational data and unstructured comment data are labeled with a three-dimensional tag system, which includes platform characteristic tags, content format tags, and user behavior tags.

[0027] S3 Intelligent Analysis: Based on the tagged data, it performs content diagnosis, traffic prediction and strategy generation operations. Among them, content diagnosis is used to divide the operational content into preset performance quadrants and output optimization suggestions. Traffic prediction is used to use the LSTM model to predict the time node when the content enters different traffic pools. Strategy generation is used to trigger the content format switching instruction when the interaction rate is lower than the preset threshold for a consecutive preset number of days.

[0028] S4. Closed-loop optimization: Feed the analysis results of step S3 back to the external content generation system to drive iterative generation of new operational content, and send the operational data of the new content back to step S1 to form a data closed loop.

[0029] Furthermore, in step S1, the core operational data is obtained through the open API interface of the new media platform and then converted into a structured data table containing the video work name, release time, and values ​​of various dimensional indicators.

[0030] The unstructured comment data was crawled by an RPA robot, and after text cleaning to remove invalid characters, it was classified into positive comments, negative comments, and neutral comments using the BERT model.

[0031] Furthermore, in step S3, the specific process of content diagnosis is as follows:

[0032] Set a playback threshold and a conversion efficiency threshold, and classify content with a playback volume higher than the playback threshold and a conversion efficiency higher than the conversion efficiency threshold into a high playback and high conversion quadrant.

[0033] Content with higher play counts than the play count threshold but lower conversion efficiency than the conversion efficiency threshold is divided into the high play count low conversion quadrant, and suggestions for optimizing the conversion path are provided.

[0034] Content with a play count below the play count threshold but a conversion rate above the conversion rate threshold is categorized into the low play count, high potential quadrant, and cover optimization suggestions are provided.

[0035] Furthermore, in step S4, the closed-loop optimization also includes:

[0036] The operational data of newly generated operational content is compared and analyzed with historical data to calculate the improvement rate of indicators before and after optimization. If the improvement rate is lower than the preset value, the parameters of the intelligent analysis module are adjusted. The parameters include the training period of the LSTM model and the preset performance quadrant division threshold.

[0037] Compared with existing technologies, the new media operation optimization system based on data closed loop described in this invention has the following advantages:

[0038] (1) The new media operation optimization method and system based on data closed loop described in this invention realizes the automated and integrated collection of operation data, solves the problems of scattered data on multiple platforms and low efficiency of manual statistics in the prior art, and automatically synchronizes the core operation data and unstructured comment information of each platform by combining structured data interface with RPA technology, reduces manual operation costs, ensures the integrity and timeliness of data, and provides a reliable foundation for subsequent analysis;

[0039] (2) The new media operation optimization method and system based on data closed loop described in this invention realizes the structured analysis of data through a three-dimensional tag system, breaking through the limitations of traditional analysis methods in terms of insufficient data dimension mining. Through multi-dimensional tagging processing of platform characteristics, content forms, and user behaviors, the distribution rules of different platforms, the adaptability of content types, and the conversion path of user behaviors can be clearly identified, helping new farmers to accurately locate their operational advantages and disadvantages;

[0040] (3) The new media operation optimization method and system based on data closed loop described in this invention constructs a closed loop mechanism of "collection-analysis-optimization-verification", which solves the problems of lagging operation strategy adjustment and lack of data feedback. The targeted optimization suggestions output by the intelligent analysis module (such as content format adjustment and conversion link optimization) can directly drive content iteration, and the optimization effect can be continuously evaluated through data feedback, realizing dynamic adjustment of strategy and significantly improving content adaptability and traffic conversion efficiency;

[0041] (4) The new media operation optimization method and system based on data closed loop described in this invention are in line with the actual operation needs of new farmers, reduce the technical threshold of new media operation, and can automatically complete data processing and strategy generation through the system without professional data analysis capabilities, thereby helping new farmers to make efficient use of new media resources and improve the promotion effect and commercial conversion capability of agricultural products. Attached Figure Description

[0042] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0043] Figure 1 This is a schematic diagram of the new media operation optimization system based on data closed loop as described in an embodiment of the present invention;

[0044] Figure 2 This is a schematic diagram of the new media operation optimization method based on data closed loop as described in an embodiment of the present invention. Detailed Implementation

[0045] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0046] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0047] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0048] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0049] To make the technical solution of this invention clearer, the following describes in detail the new media operation optimization system and method based on data closed loop described in this invention, taking the actual scenario of new farmers operating new media accounts. This embodiment uses the operation of agricultural product-related accounts on mainstream short video platforms as an example to illustrate the collaboration methods and implementation details of each module of the system.

[0050] I. System Architecture and Module Configuration

[0051] The system described in this embodiment includes a data acquisition module, a 3D labeling system module, an intelligent analysis module, and a closed-loop optimization module. Each module interacts with data through a conventional server cluster and is built using programming languages ​​and database technologies commonly used in the field to ensure the stability and timeliness of data transmission between modules.

[0052] Data acquisition module

[0053] This module includes a structured data interface unit and an unstructured data capture unit, which are used to acquire different types of operational data respectively.

[0054] The structured data interface unit synchronizes core data related to operational content by calling the open APIs provided by various new media platforms. This includes, but is not limited to, content identifiers, publication time, play counts, completion rates, interaction volume (such as likes, shares, and favorites), and user dwell time metrics. The interface call frequency is set according to the platform's data update mechanism and actual operational needs to ensure data timeliness.

[0055] The unstructured data crawling unit employs Robotic Process Automation (RPA) technology to simulate user operations and retrieve text information from the content comment section. For audio comments, speech-to-text technology is used to convert them into text data. After preprocessing (such as removing invalid information and filtering out distracting content), the crawled comment data is stored in the corresponding database to provide a foundation for subsequent analysis.

[0056] 3D label system module

[0057] This module categorizes and labels the collected data based on preset labeling rules, constructing a multi-dimensional labeling system, specifically including:

[0058] Platform characteristic tags: Based on the content distribution characteristics of different new media platforms, preset tags that reflect the platform attributes (such as focusing on entertainment dissemination, focusing on social sharing, etc.). By identifying the platform identifier of the content publication, the corresponding tags are automatically matched to distinguish the differences in the operating environment of different platforms.

[0059] Content format tags: By combining the title information, presentation format and core theme of the content, tags reflecting the content type (such as knowledge explanation, product display, scene record, etc.) are automatically marked, which facilitates subsequent targeted analysis of the operational effectiveness of different content formats.

[0060] User behavior tags: By analyzing user interaction records with content (such as clicks, dwell times, conversions, etc.), tags reflecting user behavior paths are generated. Combined with behavior sequence analysis technology, potential patterns from user contact with content to conversion are identified.

[0061] Intelligent Analysis Module

[0062] This module performs multi-dimensional analysis based on tagged data, including content diagnostics, traffic prediction, and strategy generation. The specific implementation is as follows:

[0063] Content Diagnosis Unit: Based on content playback performance and conversion results, content is categorized into different performance categories. For example, for content with good playback performance but poor conversion results, suggestions for optimizing the conversion path are provided (such as adjusting the position of the onboarding nodes in the content); for content with average playback performance but high conversion potential, suggestions for optimizing the content presentation are provided (such as adjusting the cover design to increase click-through rates).

[0064] Traffic Prediction Unit: This unit uses a time-series prediction model to analyze the changing trends of content exposure and predict the possible stages at which content enters different levels of traffic pools. This provides a reference for adjusting content release times to adapt to the platform's traffic distribution patterns.

[0065] Strategy Generation Unit: Continuously monitors the interactive performance of content. When the interaction level of a certain type of content is lower than the normal level for a period of time, it automatically triggers content format adjustment instructions (such as splitting long content into short segments or converting single-format content into combined formats), and supplements content materials with related product knowledge information to improve content relevance.

[0066] Closed-loop optimization module

[0067] This module feeds back the optimization suggestions, prediction results, and adjustment instructions output by the intelligent analysis module to the external content generation system, driving it to generate new operational content. After the new content is released, its operational data is transmitted back to the system through the data acquisition module for comparison and analysis with historical data. If the optimization effect does not meet expectations, the analysis parameters of the intelligent analysis module are adjusted in reverse (such as adjusting the performance category classification criteria and optimizing the training method of the prediction model), forming a complete data closed loop.

[0068] II. Method and Flow Steps

[0069] Based on the above system modules, the operation optimization method of this embodiment includes the following steps:

[0070] Step S1: Data Acquisition

[0071] The structured data interface unit calls the corresponding platform's open API to obtain basic operational data of the account's published content (such as play count, interaction count, user dwell time, etc.); at the same time, the unstructured data crawling unit obtains user feedback information from the content comment section, which is then preprocessed and stored in the database.

[0072] Step S2: Tagging Processing

[0073] The collected operational data is labeled with three-dimensional tags: platform characteristic tags are matched according to the publishing platform, content type tags are matched according to the content format, and user behavior tags are matched according to user operation records, forming a structured analysis dataset.

[0074] Step S3: Intelligent Analysis

[0075] Multi-dimensional analysis based on tagged data: The content diagnosis unit categorizes content performance and provides optimization suggestions; the traffic prediction unit analyzes content exposure trends and provides a reference for release timing; the strategy generation unit monitors interaction performance and triggers content format adjustment instructions when needed.

[0076] Step S4: Closed-loop optimization

[0077] The analysis results are fed back to the content generation system to drive the generation of optimized new content. After the new content is released, its operational data is sent back to the data collection module to compare with historical data to evaluate the optimization effect. If necessary, the analysis parameters are adjusted to achieve closed-loop data optimization.

[0078] III. Technical Effects Description

[0079] Through the system and method of this embodiment, new farmers can achieve the following effects in new media operations:

[0080] Automation of data processing: Replacing traditional manual statistical methods, realizing the automatic collection and organization of operational data, and reducing manual operation costs;

[0081] Targeted operational strategies: By analyzing multiple dimensions, we can identify the strengths and weaknesses of the content, making optimization suggestions more aligned with actual operational needs.

[0082] Continuous optimization process: Content format and operational strategies are continuously iterated through a data closed-loop mechanism to improve content adaptability and conversion rates. The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A new media operation optimization system based on data closed loop, characterized in that, include: The data acquisition module is used to synchronize core operational data of at least one new media platform through API interface, and to capture unstructured comment data related to operational content in the platform through RPA technology. The core operational data includes play count, completion rate, likes, reposts, favorites, comments, and three-second completion rate. A three-dimensional tagging system module, connected to the data acquisition module, is used to tag the collected core operational data and unstructured comment data. The three-dimensional tagging system includes platform characteristic tags, content format tags, and user behavior tags. The platform characteristic tags are used to distinguish the attribute characteristics of different new media platforms, the content format tags are used to mark the presentation type of operational content, and the user behavior tags are used to identify the user's behavioral path from content interaction to conversion. The intelligent analysis module, connected to the three-dimensional tag system module, includes a content diagnosis unit, a traffic prediction unit, and a strategy generation unit. The content diagnosis unit is used to divide the operational content into preset performance quadrants and output targeted optimization suggestions. The traffic prediction unit is used to use an LSTM model to predict the key time nodes for the operational content to enter different levels of traffic pools. The strategy generation unit is used to automatically trigger a content format switching instruction when the interaction rate of the operational content is lower than a preset threshold for a preset number of consecutive preset days. The closed-loop optimization module is connected to both the intelligent analysis module and the external content generation system. It is used to feed back the optimization suggestions, traffic prediction results and strategy instructions output by the intelligent analysis module to the external content generation system, so as to drive the external content generation system to iteratively generate new operational content, and to send the operational data of the newly generated operational content back to the data acquisition module, thus forming a data closed loop.

2. The new media operation optimization system based on data closed loop according to claim 1, characterized in that, The data acquisition module includes a structured data interface unit and an unstructured data capture unit. The structured data interface unit is used to obtain the core operational data by calling the open API of the new media platform and convert it into a structured data format. The unstructured data crawling unit is used to simulate user operation behavior through RPA robots, crawl text comment data in the comment section, and convert audio comment data into text data through ASR technology.

3. The new media operation optimization system based on data closed loop according to claim 1, characterized in that, In the three-dimensional tag system module, the platform characteristic tags include entertainment attribute tags, social attribute tags, and information attribute tags, which are used to characterize the content distribution preferences of different platforms; The content format tags include knowledge popularization tags, product review tags, story narrative tags, and interactive Q&A tags, which are used to distinguish the creation type of operational content; The user behavior tags include click tags, favorite tags, forwarding tags, and purchase tags, which are used to construct a user behavior sequence model through a hidden Markov model.

4. The new media operation optimization system based on data closed loop according to claim 1, characterized in that, In the content diagnosis unit of the intelligent analysis module, the preset performance quadrants include a high-playback-high-conversion quadrant, a high-playback-low-conversion quadrant, a low-playback-high-potential quadrant, and a low-playback-low-conversion quadrant. For operational content in the high-playback-low-conversion quadrant, the optimization suggestions include adjusting the conversion link nodes in the content to shorten the jump steps from user interaction to purchase. For operational content in the low-playback-high-potential quadrant, the optimization suggestions include replacing the content cover with a comparison image containing visual guidance elements.

5. The new media operation optimization system based on data closed loop according to claim 1, characterized in that, In the traffic prediction unit of the intelligent analysis module, the different levels of traffic pools include a 500-person exposure pool, a 5000-person exposure pool, and a 50000-person exposure pool. The input features of the LSTM model include publication time, content tags, and the initial 30-minute playback volume of similar historical content. The output is the predicted time point for entering each level of traffic pool, and the model generates publication time adjustment suggestions based on the prediction results.

6. The new media operation optimization system based on data closed loop according to claim 1, characterized in that, In the strategy generation unit of the intelligent analysis module, the preset number of days is 3 days, and the preset threshold is 85% of the average interaction rate of the same period in history. The content format switching instructions include splitting long videos into short video segments of preset duration and converting text and image content into animated demonstrations. The instructions also include supplementary material information generated by calling external enterprise knowledge bases based on RAG technology.

7. A new media operation optimization method based on data closed loop, characterized in that, Includes the following steps: S1. Data Collection: Synchronize core operational data of at least one new media platform through API interface, and capture unstructured comment data related to operational content in the platform through RPA technology. The core operational data includes play count, completion rate, likes, reposts, favorites, comments, and three-second completion rate. S2. Tagging Processing: The collected core operational data and unstructured comment data are labeled with a three-dimensional tag system, which includes platform characteristic tags, content format tags, and user behavior tags. S3 Intelligent Analysis: Based on the tagged data, it performs content diagnosis, traffic prediction and strategy generation operations. Among them, content diagnosis is used to divide the operational content into preset performance quadrants and output optimization suggestions. Traffic prediction is used to use the LSTM model to predict the time node when the content enters different traffic pools. Strategy generation is used to trigger the content format switching instruction when the interaction rate is lower than the preset threshold for a consecutive preset number of days. S4. Closed-loop optimization: Feed the analysis results of step S3 back to the external content generation system to drive iterative generation of new operational content, and send the operational data of the new content back to step S1 to form a data closed loop.

8. The new media operation optimization method based on data closed loop according to claim 1, characterized in that, In step S1, the core operational data is obtained through the open API interface of the new media platform and then converted into a structured data table containing the video work name, release time, and values ​​of various dimensional indicators. The unstructured comment data was crawled by an RPA robot, and after text cleaning to remove invalid characters, it was classified into positive comments, negative comments, and neutral comments using the BERT model.

9. The new media operation optimization method based on data closed loop according to claim 1, characterized in that, In step S3, the specific process of content diagnosis is as follows: Set a playback threshold and a conversion efficiency threshold, and classify content with a playback volume higher than the playback threshold and a conversion efficiency higher than the conversion efficiency threshold into a high playback and high conversion quadrant. Content with higher play counts than the play count threshold but lower conversion efficiency than the conversion efficiency threshold is divided into the high play count low conversion quadrant, and suggestions for optimizing the conversion path are provided. Content with a play count below the play count threshold but a conversion rate above the conversion rate threshold is categorized into the low play count, high potential quadrant, and cover optimization suggestions are provided.

10. The new media operation optimization method based on data closed loop according to claim 1, characterized in that, In step S4, the closed-loop optimization further includes: The operational data of newly generated operational content is compared and analyzed with historical data to calculate the improvement rate of indicators before and after optimization. If the improvement rate is lower than the preset value, the parameters of the intelligent analysis module are adjusted. The parameters include the training period of the LSTM model and the preset performance quadrant division threshold.

Images