A telecommunication service generation system based on the Internet

Abstract

The invention discloses a telecommunication service generation system based on the Internet, which relates to the field of machine learning and data mining, and comprises a collection module for collecting flow data and preprocessing it, a verification module for a user to connect to a telecommunication service generation platform, verify the identity, and obtain user information after passing the verification, a demand module for entering a user interface, a user inputting a service type, parsing the information input by the user, extracting the service demand of the user, a configuration module for providing a service combination based on the service demand of the user, and generating a service configuration file, and a trend prediction module for detecting the network flow of a new service, and establishing a flow trend prediction model based on the preprocessed flow data to obtain a flow trend prediction result; the invention establishes a flow trend prediction model to predict the flow trend in a future period of time, thereby providing a scientific basis for the allocation of bandwidth resources, reasonably allocating bandwidth resources in different time periods, avoiding network congestion, and improving resource utilization and service quality.

Description

Technical Field

[0001] The present invention relates to the field of machine learning and data mining, and in particular to an Internet-based telecommunication service generation system. Background Art

[0002] With the development of Internet technology, telecommunication service generation systems have become an indispensable part of modern communication networks. Traditional TSGS mainly relies on static configuration and pre-set service templates to meet users' basic communication needs. However, with the diversification and complexity of user needs and the dynamic changes in network traffic patterns, traditional TSGS has gradually shown its limitations in terms of flexibility, intelligence and service quality.

[0003] Traditional methods have improved the efficiency and intelligence level of telecommunications business generation systems to a certain extent, but there are still some shortcomings in practical applications. First, most of the existing service demand analysis modules are based on fixed rule sets, lacking in-depth understanding of user input information and flexible response capabilities, which can easily lead to improper service configuration or poor user experience. Secondly, it is impossible to establish a traffic trend prediction model based on real-time traffic data, obtain traffic trend prediction results, and automatically adjust bandwidth allocation and optimize path selection. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides an Internet-based telecommunications service generation system to solve the problem of being unable to establish a traffic trend prediction model based on real-time traffic data, obtain traffic trend prediction results, and automatically adjust bandwidth allocation and optimize path selection.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] The present invention provides a telecommunication service generation system based on the Internet, which comprises:

[0008] Collection module, collects traffic data and pre-processes it;

[0009] Verification module, users connect to the telecommunications service generation platform, verify their identity, and obtain user information after passing;

[0010] Demand module: Enter the user interface, the user inputs the service type, the information entered by the user is parsed, and the user's service requirements are extracted;

[0011] Configuration module, which provides service combinations based on user service requirements and generates service configuration files;

[0012] The trend prediction module detects the network traffic of the new service and establishes a traffic trend prediction model based on the preprocessed traffic data to obtain the traffic trend prediction result;

[0013] Path planning module, which plans the network data transmission path based on the traffic trend prediction results;

[0014] Update modules to improve service quality based on user feedback and its own operating data.

[0015] As a preferred solution of the Internet-based telecommunication service generation system of the present invention, the traffic data includes data transmission rate, delay time and packet loss rate.

[0016] As a preferred solution of the Internet-based telecommunication service generation system of the present invention, the preprocessing includes filtering noise and filling missing values.

[0017] As a preferred solution of the Internet-based telecommunication service generation system of the present invention, wherein: a user connects to the telecommunication service generation platform, verifies his identity, and obtains user information after passing the identity verification, specifically including the following steps:

[0018] The user logs in to the telecommunications service generation platform, which verifies the user's identity through SMS verification code and facial recognition;

[0019] If the verification is successful, the platform accesses the backend database to extract the user's personal information and historical service records;

[0020] Enable dynamic permission management under the zero-trust network architecture to determine the functional modules and service levels that users can access;

[0021] Assign an initial set of permissions to users based on the functional modules and service levels they access.

[0022] As a preferred solution of the Internet-based telecommunication service generation system of the present invention, wherein: entering the user interface, the user inputs the service type, parsing the information input by the user, and extracting the user's service requirements, specifically includes the following steps:

[0023] After the user obtains the initial set of permissions, the platform automatically loads the user interface;

[0024] Define specific parsing functions and multimodal input fusion functions;

[0025] Users express service needs and types through voice commands, text input, and image uploads;

[0026] The service requirements and types expressed by the user are analyzed through the specific parsing function and the multimodal input fusion function to obtain the preliminary parsing results, and the preliminary parsing results are combined into a preliminary parsing result vector, which is expressed as follows:

[0027] ;

[0028] ;

[0029] in, Indicates Preliminary parsing results of class input types, Indicates that for Specific parsing functions for class input, Indicates Class input data, represents the initial parsing result vector, represents the multimodal input fusion function, Indicates the number of input types;

[0030] The maximum likelihood estimation is used to extract the user's service requirements from the preliminary parsing result vector, and its expression is:

[0031] ;

[0032] in, Represents the extracted user service requirements, Indicates service demand, Represents a given preliminary parsing result vector Next service demand Probability of occurrence.

[0033] As a preferred solution of the Internet-based telecommunication service generation system of the present invention, wherein: providing a service combination based on the user's service needs and generating a service configuration file specifically includes the following steps:

[0034] Based on the extracted user service requirements, a similarity algorithm is used to search for the service template with the highest matching degree in the service template library;

[0035] Among the service templates with the highest matching degree, select the first N closest service combinations, give the cost estimate of each service combination, and organize them into a list to display to the user on the interface;

[0036] Users can modify the recommended service combination on the interface, and the modification will be synchronously fed back to the server;

[0037] After the user confirms that the service combination is correct and submits it, the server will generate the corresponding service configuration file;

[0038] Send the generated service configuration file to the background server to execute the creation task;

[0039] After the creation task is completed, the user will receive a notification message that the new service has been successfully activated.

[0040] As a preferred solution of the Internet-based telecommunication service generation system of the present invention, the network traffic of the new service is detected, and a traffic trend prediction model is established based on the pre-processed traffic data to obtain the traffic trend prediction result, which specifically includes the following steps:

[0041] Integrate the preprocessed flow data into a flow data vector;

[0042] Divide the traffic data vector into training set and test set, and define the time window;

[0043] Select the LSTM model as the basic model for traffic trend prediction, define the mean square error as the loss function, and use the training set to train the traffic trend prediction model until the loss function converges to the minimum value;

[0044] The performance of the traffic trend prediction model is evaluated based on the test set to obtain the traffic trend prediction results.

[0045] As a preferred solution of the Internet-based telecommunication service generation system of the present invention, the obtaining of the traffic trend prediction result specifically includes the following steps:

[0046] In the time window, the traffic trend prediction model is used to analyze the traffic data vector to obtain the traffic trend prediction result, which is expressed as:

[0047] ;

[0048] in, Indicates the future time window The traffic trend vector predicted internally, represents the traffic trend prediction model, represents the flow data vector, represents the set of model parameters, represents the future time window, represents the attenuation factor, Represents a specific time point in the future time window, Represents the flow data vector In time The value of

[0049] The traffic trend vector predicted in the future time window is the traffic trend prediction result.

[0050] As a preferred solution of the Internet-based telecommunication service generation system of the present invention, wherein: planning the network data transmission path according to the traffic trend prediction result specifically includes the following steps:

[0051] Based on the traffic trend prediction results, an optimization objective function is defined to measure the error between bandwidth allocation and predicted traffic, and to penalize the change in bandwidth allocation;

[0052] The Adam optimizer is used to perform gradient descent on the optimization objective function to obtain the bandwidth allocation vector, which is expressed as:

[0053] ;

[0054] in, represents the optimal bandwidth allocation vector, represents the bandwidth allocation vector, Indicates at a point in time The predicted flow rate, At the point in time The allocated bandwidth, Indicates at a point in time Forecasted traffic The sum of the squares of represents the smoothing parameter, Indicates at a point in time allocated bandwidth;

[0055] Based on the obtained optimal bandwidth allocation vector, the optimal network data transmission path is planned.

[0056] As a preferred solution of the Internet-based telecommunication service generation system of the present invention, the service quality is improved based on user feedback and its own operation data, which specifically includes the following steps:

[0057] Regularly collect user feedback on new services and prediction results of traffic trend prediction models;

[0058] Evaluate the difference between the prediction results of the traffic trend prediction model and the actual traffic;

[0059] Based on user feedback and the difference between the prediction results of the traffic trend prediction model and the actual traffic, the parameters of the traffic trend prediction model are updated, and the latest security patches and upgrade packages are pushed regularly.

[0060] The beneficial effects of the present invention are as follows: by comprehensively capturing the user's multimodal input and analyzing the user's needs, more personalized service recommendations can be provided to the user; through accurate understanding of the user's needs, a highly customized and personalized service combination can be provided, thereby improving the user's satisfaction and experience; in addition, by establishing a traffic trend prediction model, the traffic trend in the future period is predicted, which provides a scientific basis for the reasonable allocation of bandwidth resources, reasonably allocates bandwidth resources in different time periods, avoids network congestion, and improves resource utilization and service quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0062] Figure 1 This is a schematic diagram of an Internet-based telecommunication service generation system in Example 1.

[0063] Figure 2 This is a schematic diagram of telecommunication service generation in Example 1. DETAILED DESCRIPTION

[0064] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0065] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0066] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0067] Example 1, reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, and which provides a telecommunication service generation system based on the Internet, comprising the following steps:

[0068] The acquisition module collects traffic data and performs preprocessing.

[0069] Deploy sensor nodes at key locations of the network (such as routers, switches, etc.) to collect data transmission rate, delay time and packet loss rate, and set the data collection frequency to once per second to ensure the real-time nature of the data.

[0070] It is further explained that by deploying sensor nodes at multiple key locations, complete network traffic data can be captured, avoiding important information that may be missed by local monitoring. Moreover, by collecting a variety of traffic data, network performance can be evaluated from different angles, providing more comprehensive service quality assurance.

[0071] For the collected data transmission rate, delay time and packet loss rate, Kalman filter is applied to remove noise data to ensure the smoothness and consistency of the data, and interpolation methods are used to fill in possible missing values. For example, linear interpolation or spline interpolation can be used to infer missing values ​​based on the data of adjacent time points to ensure data continuity.

[0072] It is further explained that through preprocessing, more accurate and reliable flow data can be obtained, which improves the accuracy of subsequent analysis and prediction.

[0073] Verification module, users connect to the telecommunications service generation platform, verify their identity, and obtain user information after passing.

[0074] The user connects to the telecommunications service generation platform through the Internet, and the telecommunications service generation platform verifies the user's identity through a text message verification code or facial recognition.

[0075] After successful verification, the backend database is accessed to extract the user's personal information and historical service records (such as previously customized service packages, usage habits, etc.).

[0076] Based on the user's identity information and role definition, the permission management rule base is instantly queried to determine the functional modules and service levels that the user can access. Based on the determination results, the user is assigned an initial permission set to restrict access to only the functions that match their role. For example, ordinary users can only view and modify personal information and customize services; administrators have more management permissions.

[0077] Further explanation: Ensure that users can only access the functional modules and services required for their work and avoid unnecessary permission exposure. Zero trust architecture and permission management significantly improve the security of the platform and reduce the risk of internal threats and data leakage.

[0078] Demand module: Enter the user interface, the user inputs the service type, the information entered by the user is parsed, and the user's service requirements are extracted;

[0079] According to the user's permission level, the platform automatically loads the corresponding user interface to ensure that the user can directly enter the functional modules to which he / she has access. The content displayed on the user interface is personalized according to the user's historical service records, providing service options that better meet the user's needs.

[0080] It can be further explained that the process from permission verification to interface loading is smooth and fluid, which reduces the user's waiting time.

[0081] For different types of user input (such as voice commands, text input, image upload), define specific parsing functions to convert each input into preliminary parsing results.

[0082] Define a multimodal input fusion function to combine all preliminary parsing results into a preliminary parsing result vector.

[0083] Users can input service requirements through the voice assistant and use speech recognition technology to convert speech into text; secondly, users can input service requirements directly in the text box and use the BERT model to parse and understand the context and assign more precise weights to each word; finally, users can also upload pictures and use image recognition technology to analyze the picture content and extract relevant service requirements.

[0084] It is further explained that multiple input methods are supported to meet the input habits of different users and improve the convenience and efficiency of use.

[0085] Use the defined parsing function to parse each input type and get the preliminary parsing results.

[0086] For example, the user inputs the service requirements through the voice assistant, captures the user's voice input, sends the voice data to the speech recognition API (such as Google Speech-to-Text) to convert the voice into text, receives and processes the returned text data, uses natural language processing (NLP) technology to perform preliminary analysis on the text, extracts key information, and obtains preliminary analysis results. If the user enters service requirements in the text box, the text entered by the user is captured first, and then the BERT model is used to encode the text, obtain the embedded representation of each word, adjust the importance of each word according to the context, and then perform semantic analysis on the text content to extract key service requirement information and obtain preliminary analysis results. .

[0087] All preliminary parsing results are combined into a preliminary parsing result vector through a multimodal input fusion function, which is expressed as:

[0088] ;

[0089] ;

[0090] in, Indicates Preliminary parsing results of class input types, Indicates that for Specific parsing functions for class input, Indicates Class input data, represents the initial parsing result vector, represents the multimodal input fusion function, Indicates the number of input types;

[0091] It is further explained that the initial analysis result vector provides a preliminary understanding of user needs and lays the foundation for subsequent further analysis.

[0092] Using the maximum likelihood estimation method, the user's service requirements are extracted from the preliminary parsing result vector, and the expression is:

[0093] ;

[0094] in, Represents the extracted user service requirements, Indicates service demand, Represents a given preliminary parsing result vector Next service demand Probability of occurrence.

[0095] It is further explained that by accurately identifying user needs, more suitable services can be provided, thereby improving the pertinence and effectiveness of services.

[0096] The configuration module provides service combinations based on user service requirements and generates service configuration files.

[0097] Based on the extracted user service requirements, a similarity algorithm is used to find the service template with the highest matching degree in the service template library. Commonly used similarity algorithms include cosine similarity and Jaccard similarity, which can measure the similarity between user requirements and service templates.

[0098] The service template library is a set of predefined service templates covering different types of telecommunication services and their configuration options.

[0099] Further explanation: Through the similarity algorithm, the service template that best meets the user's needs is found to ensure that the provided service combination is more in line with the user's actual needs. And the best matching service template is quickly located, which can reduce the user's waiting time.

[0100] Among the service templates with the highest matching degree, select the first N (N is less than or equal to 3) closest service combinations, and calculate the cost of each service combination based on the service content in the service combination (such as package type, traffic quota, additional functions, etc.). At the same time, the recommended service combinations and their cost estimates are displayed in a list on the user interface to facilitate user comparison and selection.

[0101] It further explains that users can clearly see different service combinations and their costs, making it easier for them to make choices that suit their needs, and multiple options are provided to meet the budgets and preferences of different users.

[0102] Users can modify the recommended service combination on the interface.

[0103] For example, service package 1: 5G package, unlimited data, international roaming, monthly fee of 55 yuan;

[0104] Service package 2: broadband service, home use, Wi-Fi coverage, monthly fee 48 yuan;

[0105] Service Package 3: Comprehensive package (5G and broadband), monthly fee of 60 yuan.

[0106] The user chooses Service Package 2, but wants to add international roaming function. He checks the international roaming option on the interface and adjusts the monthly fee budget to 50 yuan.

[0107] Any changes made by the user will be immediately reflected on the server to ensure consistency between the front-end and back-end data.

[0108] It is further explained that users can modify the service combination directly on the interface, which enhances interactivity and convenience.

[0109] Based on the service combination finally confirmed by the user, a standardized service configuration file is automatically generated, which contains all necessary configuration information and service parameters, thus reducing manual intervention and improving efficiency.

[0110] After receiving the configuration file, the backend server schedules the corresponding tasks to create and activate the new service. During the creation process, detailed logs will be recorded for subsequent tracking and auditing.

[0111] It is further explained that the background server responds and executes the creation task quickly, ensuring that the service is activated as soon as possible, and the detailed logging provides a basis for troubleshooting and performance optimization.

[0112] After the task is created, the user will be notified via SMS, email, or in-app messages. The notification also contains basic information and usage guidelines of the new service to help users understand it quickly.

[0113] The trend prediction module detects the network traffic of the new service and establishes a traffic trend prediction model based on the preprocessed traffic data to obtain the traffic trend prediction result.

[0114] The preprocessed flow data are integrated into flow data vectors, which simplifies the subsequent model input.

[0115] The traffic data vector is divided into a training set and a test set, usually in a ratio of 70%-30% or 80%-20%, and a time window is defined (for example, 7 days, 14 days, and 30 days). The ratio of the training set and the test set and the time window can be customized according to the actual situation and usage scenario.

[0116] LSTM (Long Short-Term Memory Network) is selected as the base model for traffic trend prediction because it is good at handling long-term dependencies in time series data.

[0117] Use the training set data to train the traffic trend prediction model, continuously update the LSTM model parameters, define the mean square error as the loss function, adjust the LSTM model parameters through the back propagation algorithm, and gradually reduce the loss function until the loss function converges to the minimum value. The expression of the mean square error is:

[0118] ;

[0119] in, represents the mean square error, Represents the number of samples, that is, the number of pairs of true values ​​and predicted values ​​involved in the calculation. Indicates The true value at a time point, Indicates The predicted value at a time point.

[0120] When the mean square error converges to the minimum value, the traffic trend prediction model training is completed.

[0121] The performance of the traffic trend prediction model is evaluated based on the test set.

[0122] In the time window, the traffic trend prediction model is used to analyze the traffic data vector to obtain the traffic trend prediction result, which is expressed as:

[0123] ;

[0124] in, Indicates the future time window The traffic trend vector predicted internally, represents the traffic trend prediction model, represents the flow data vector, represents the set of model parameters, represents the future time window, represents the attenuation factor, Represents a specific time point in the future time window, Represents the flow data vector In time The value of

[0125] The traffic trend vector predicted in the future time window is the traffic trend prediction result.

[0126] It is further explained that the traffic trend prediction model that selects the long short-term memory network (LSTM) as the basic model is good at processing long-term dependencies in time series data and improves the accuracy of prediction.

[0127] Path planning module, which plans the network data transmission path based on the traffic trend prediction results;

[0128] Based on the traffic trend prediction results, an optimization objective function is defined to measure the error between bandwidth allocation and predicted traffic. By minimizing the error between bandwidth allocation and predicted traffic, bandwidth allocation is ensured to be as close to actual demand as possible. The change range of bandwidth allocation is penalized and a penalty term for bandwidth change is introduced to avoid frequent fluctuations in bandwidth allocation and improve network stability.

[0129] The Adam optimizer is used to perform gradient descent on the optimization objective function to find the optimal bandwidth allocation vector, which is expressed as:

[0130] ;

[0131] in, represents the optimal bandwidth allocation vector, represents the bandwidth allocation vector, Indicates at a point in time The predicted flow rate, At the point in time The allocated bandwidth, Indicates at a point in time Forecasted traffic The sum of the squares of represents the smoothing parameter, Indicates at a point in time allocated bandwidth;

[0132] The Adam optimizer combines the advantages of the momentum method and adaptive learning rate, can converge quickly and effectively handle sparse gradient problems, and improves optimization efficiency.

[0133] The shortest path or the path with the least delay is selected according to the obtained optimal bandwidth allocation vector to ensure the high efficiency of network data transmission.

[0134] Update modules to improve service quality based on user feedback and its own operating data.

[0135] Regularly collect user feedback on new services and prediction results of traffic trend prediction models through questionnaires, online reviews, customer service records, etc.

[0136] Compare the prediction results of the traffic trend prediction model with the actual traffic data to evaluate the accuracy of the model. Collect actual traffic data every day or every week to ensure the real-time and accuracy of the data. Compare the predicted values ​​with the actual values ​​to calculate various evaluation indicators. Commonly used evaluation indicators include mean square error, mean absolute error, root mean square error, etc.

[0137] Based on user feedback and evaluation results, update the parameters of the traffic trend prediction model to improve the model's prediction accuracy. If the model performs poorly, the model can be retrained using the latest traffic data.

[0138] It is further explained that by continuously updating the model parameters, the model is ensured to be always in the best state, which improves the accuracy and stability of the prediction.

[0139] Conduct security assessments regularly to identify potential security risks and vulnerabilities, develop corresponding security patches and upgrade packages for the problems found, and push them to users.

[0140] It is further explained that timely repair of known vulnerabilities can prevent potential security threats and improve operational efficiency and stability.

[0141] In summary, the present invention provides users with more personalized service recommendations by comprehensively capturing users' multimodal inputs and analyzing user needs. By accurately understanding user needs, it can provide highly customized and personalized service combinations, thereby improving user satisfaction and experience. In addition, by establishing a traffic trend prediction model, the traffic trend in the future is predicted, which provides a scientific basis for the reasonable allocation of bandwidth resources, reasonably allocates bandwidth resources in different time periods, avoids network congestion, and improves resource utilization and service quality.

[0142] Example 2, referring to Table 1, is the second example of the present invention. To further verify the technical solution of the present invention, experimental simulation data of an Internet-based telecommunication service generation system is provided.

[0143] In order to verify the Internet-based telecommunication service generation system, a set of experiments was designed to evaluate its performance in actual applications. The experimental object is a service platform of a telecommunications operator, aiming to improve service response speed, reduce delay time and reduce packet loss rate through the present invention. The following are the detailed experimental steps:

[0144] Collection module: First, a traffic data collector is deployed on the service platform to monitor and record key parameters of network traffic in real time, including data transmission rate (Mbps), latency (ms), and packet loss rate (%). These raw data are collected into a database and then preprocessed. The preprocessing step includes using a high-pass filter to remove high-frequency noise and using linear interpolation to fill in possible missing values.

[0145] Verification module: When a user logs in, the user's identity is verified through a text message verification code combined with facial recognition technology. Once the verification is successful, the backend database will be accessed to extract the user's personal information and historical service records. On this basis, the dynamic permission management function under the zero-trust network architecture is activated to determine the user's access rights and service level based on the user's historical behavior pattern, and then assign the initial permission set.

[0146] Demand module: After obtaining the initial permission set, the user enters a customized interface and can express specific service requirements through voice commands, text input, or image upload. A multimodal input fusion function is used to parse different forms of input, and the parsing results are combined into a preliminary parsing result vector. The maximum likelihood estimation method is used to extract the most likely service requirements from the vector.

[0147] Configuration module: Based on the extracted service requirements, the service template library is automatically searched for the service combination with the highest matching degree. The top three best matching service options are selected and displayed to the user for selection and modification. The final confirmed service combination will be converted into a service configuration file and sent to the backend server to execute the creation task. When the creation task is completed, the user will receive a notification message informing him that the new service has been successfully activated.

[0148] Trend prediction module: During the activation of new services, network traffic changes are continuously monitored. The preprocessed traffic data is integrated into a traffic data vector and divided into a training set and a test set at a ratio of 70%-30%. A 7-day time window is defined. The long short-term memory network (LSTM) is selected as the basic model for traffic trend prediction, the mean square error (MSE) is defined as the loss function, and the model parameters are continuously updated through the back propagation algorithm until convergence. After that, the model performance is evaluated based on the test set to obtain the traffic trend prediction results.

[0149] Path planning module: Based on the above prediction results, an optimization objective function is formulated, taking into account the error between bandwidth allocation and predicted traffic and the bandwidth change range. The Adam optimizer is used to perform gradient descent on the objective function, calculate the optimal bandwidth allocation vector, and plan the optimal data transmission path accordingly to ensure the effective use of network resources.

[0150] Update module: Regularly collect user feedback on new services and compare the prediction results of the traffic trend prediction model with the actual traffic difference. Adjust model parameters based on these feedback and analysis results, push the latest security patches and upgrade packages, and continuously improve service quality.

[0151] The details are shown in Table 1 below:

[0152] Table 1 Network speed comparison table

[0153]

[0154] By comparing and analyzing the indicators in the above table, it can be clearly seen that the solution of the present invention has significant advantages over the comparative solution:

[0155] Data transmission rate: Traditional telecommunication service solutions can usually provide a data transmission rate of about 82.3 Mbps, but through the present invention, this value is increased to 119.5 Mbps. This means that under the same conditions, the present invention can complete the data transmission task faster, greatly improving the user experience.

[0156] Delay time: In terms of delay time, the comparison scheme reached an average of 48.7 ms, which may have an adverse effect in some application scenarios with high timeliness requirements. However, after the optimization of the present invention, the delay time was significantly shortened to 29.8 ms, a reduction of nearly 38.8%, making the interaction smoother and meeting the needs of more types of applications.

[0157] Packet loss rate: Packet loss rate is one of the key factors to measure network stability. The packet loss rate under the comparison scheme is 1.45%, that is, about 1.45 packets are lost for every 100 packets transmitted. In contrast, the scheme of the present invention controls the packet loss rate to 0.47%, which is almost two-thirds lower, greatly improving the reliability of data transmission.

[0158] Service quality score: Finally, from the comprehensive service quality score, the comparison solution scored 76 points, while the solution of the present invention scored a high score of 94 points. This shows that the present invention not only performs well in technical indicators, but also has a qualitative leap in the overall service level.

[0159] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A telecommunication service generation system based on the Internet, characterized in that: include, Collection module, collects traffic data and pre-processes it; Verification module, users connect to the telecommunications service generation platform, verify their identity, and obtain user information after passing; Demand module: Enter the user interface, the user inputs the service type, the information entered by the user is parsed, and the user's service requirements are extracted; The configuration module provides service combinations based on user service requirements and generates service configuration files. After receiving the configuration files, the backend server schedules corresponding tasks to create and activate new services. The trend prediction module detects the network traffic of the new service and establishes a traffic trend prediction model based on the preprocessed traffic data to obtain the traffic trend prediction result; Path planning module, which plans the network data transmission path based on the traffic trend prediction results; Update modules to improve service quality based on user feedback and its own operating data; Enter the user interface, the user enters the service type, the user's input information is parsed, and the user's service requirements are extracted. The specific steps include the following: After the user obtains the initial set of permissions, the platform automatically loads the user interface; Define specific parsing functions and multimodal input fusion functions; Users express service needs and types through voice commands, text input, and image uploads; The service requirements and types expressed by users are analyzed through specific parsing functions and multimodal input fusion functions to obtain preliminary parsing results, and the preliminary parsing results are combined into a preliminary parsing result vector, which is expressed as follows: ； ； in, Indicates Preliminary parsing results of class input types, Indicates that for Specific parsing functions for class input, Indicates Class input data, represents the initial parsing result vector, represents the multimodal input fusion function, Indicates the number of input types; The maximum likelihood estimation is used to extract the user's service requirements from the preliminary parsing result vector, and its expression is: ； in, Represents the extracted user service requirements, Indicates service demand, Represents a given preliminary parsing result vector Next service demand Probability of occurrence; Planning the network data transmission path based on the traffic trend prediction results includes the following steps: Based on the traffic trend prediction results, an optimization objective function is defined to measure the error between bandwidth allocation and predicted traffic, and to penalize the change in bandwidth allocation; The Adam optimizer is used to perform gradient descent on the optimization objective function to obtain the bandwidth allocation vector, which is expressed as: ； in, represents the optimal bandwidth allocation vector, represents the bandwidth allocation vector, Indicates at a point in time The predicted flow rate, At the point in time The allocated bandwidth, Indicates at a point in time Forecasted traffic The sum of the squares of represents the smoothing parameter, Indicates at a point in time allocated bandwidth; Based on the obtained optimal bandwidth allocation vector, the optimal network data transmission path is planned.

2. The Internet-based telecommunication service generation system as claimed in claim 1, characterized in that: The traffic data includes data transmission rate, delay time and packet loss rate.

3. The Internet-based telecommunication service generation system as claimed in claim 2, characterized in that: The preprocessing includes filtering noise and filling missing values.

4. The Internet-based telecommunication service generation system as claimed in claim 3, characterized in that: The user connects to the telecommunications service generation platform, verifies his identity, and obtains user information after passing the verification. The specific steps include the following: The user logs in to the telecommunications service generation platform, which verifies the user's identity through SMS verification code and facial recognition; If the verification is successful, the platform accesses the backend database to extract the user's personal information and historical service records; Enable dynamic permission management under the zero-trust network architecture to determine the functional modules and service levels that users can access; Assign an initial set of permissions to users based on the functional modules and service levels they access.

5. The Internet-based telecommunication service generation system as claimed in claim 4, characterized in that: Providing service combinations based on user service requirements and generating service configuration files specifically includes the following steps: Based on the extracted user service requirements, a similarity algorithm is used to search for the service template with the highest matching degree in the service template library; Among the service templates with the highest matching degree, select the first N closest service combinations, give the cost estimate of each service combination, and organize them into a list to display to the user on the interface; Users can modify the recommended service combination on the interface, and the modification will be synchronously fed back to the server; After the user confirms that the service combination is correct and submits it, the server will generate the corresponding service configuration file; Send the generated service configuration file to the background server to execute the creation task; After the creation task is completed, the user will receive a notification message that the new service has been successfully activated.

6. The Internet-based telecommunication service generation system according to claim 5, characterized in that: Detect the network traffic of the new service, and establish a traffic trend prediction model based on the preprocessed traffic data to obtain the traffic trend prediction results. The specific steps include the following: Integrate the preprocessed flow data into a flow data vector; Divide the traffic data vector into training set and test set, and define the time window; Select the LSTM model as the basic model for traffic trend prediction, define the mean square error as the loss function, and use the training set to train the traffic trend prediction model until the loss function converges to the minimum value; The performance of the traffic trend prediction model is evaluated based on the test set to obtain the traffic trend prediction results.

7. The Internet-based telecommunication service generation system as claimed in claim 6, characterized in that: The method of obtaining the traffic trend prediction result specifically includes the following steps: In the time window, the traffic trend prediction model is used to analyze the traffic data vector to obtain the traffic trend prediction result, which is expressed as: ； in, Indicates the future time window The traffic trend vector predicted internally, represents the traffic trend prediction model, represents the flow data vector, represents the set of model parameters, represents the future time window, represents the attenuation factor, Represents a specific time point in the future time window, Represents the flow data vector In time The value of The traffic trend vector predicted in the future time window is the traffic trend prediction result.

8. The Internet-based telecommunication service generation system as claimed in claim 7, characterized in that: Improve service quality based on user feedback and own operation data, including the following steps: Regularly collect user feedback on new services and prediction results of traffic trend prediction models; Evaluate the difference between the prediction results of the traffic trend prediction model and the actual traffic; Based on user feedback and the difference between the prediction results of the traffic trend prediction model and the actual traffic, the parameters of the traffic trend prediction model are updated, and the latest security patches and upgrade packages are pushed regularly.

Images