Yellow page number analysis control method and device based on deep learning

Abstract

The invention discloses a yellow page number analysis control method and device based on deep learning, and the method is characterized in that the method comprises the steps: obtaining call behavior data of a yellow page number library; inputting the call behavior data into the optimized number judgment model, and obtaining a yellow page number stability result through a gradient boosting tree algorithm; and dynamically adjusting the execution of the maintenance strategy of the yellow page number based on the stability result of the yellow page number. The yellow page number stability can be accurately analyzed, the number library data quality can be improved, unstable numbers can be timely found and classified, more reliable services can be provided for users, the cost can be reduced, and the management efficiency can be improved.

Description

Technical Field

[0001] The present application relates to the field of communication technology, and in particular to a yellow pages number analysis control method and device based on deep learning. Background Art

[0002] Yellow Pages numbers are widely used in modern business and social life. They serve as a vital bridge for effective communication and business dealings between businesses, institutions, and service providers and their customers and partners. They cover a wide range of sectors, including commercial services, public utilities, and healthcare. Yellow Pages numbers enable users to quickly and conveniently access the services and information they need.

[0003] Existing yellow pages number management technology usually relies on manual collection and organization of number information, which is then entered into a database for management; in terms of number verification, it mainly relies on a large number of outbound call verifications and manual participation to confirm the validity and accuracy of the number.

[0004] However, this technology suffers from poor timeliness. Changes to the number owner cannot be detected promptly, resulting in users frequently encountering invalid or incorrect numbers when using Yellow Pages numbers, disrupting normal communication and business operations. Furthermore, existing technology requires extensive outbound call verification and manual intervention, which not only consumes significant manpower, time, and financial costs, but also produces poor results. Summary of the Invention

[0005] In view of this, the present application provides a yellow pages number analysis and control method and device based on deep learning, which can timely change and maintain the information of yellow pages numbers, reduce costs and improve management efficiency.

[0006] The present application provides a yellow page number analysis and control method based on deep learning, the method comprising:

[0007] Obtain call behavior data from the Yellow Pages number database;

[0008] Input the call behavior data into the optimized number determination model and obtain the stability result of the yellow page number through the gradient boosting tree algorithm;

[0009] Based on the yellow page number stability result, the execution of the yellow page number maintenance strategy is dynamically adjusted.

[0010] Optionally, the call behavior characteristic data is classified and sorted to obtain the call behavior characteristic data of each yellow pages number in the user's call behavior data, wherein the call behavior characteristic data includes the call frequency within a specific time period, the difference in call behavior between weekdays and weekends, the call frequency difference, the frequency variance and the number history label characteristics.

[0011] Optionally, the number determination model adopts an XGBoost classification model, and the model training steps include:

[0012] Get a sample number dataset with labeled results;

[0013] The sample number dataset is randomly split into a training set and a validation set in a ratio of 7:3, and an XGBoost classification model is constructed for the training set based on the XGBoost algorithm;

[0014] The 5-fold cross validation method was used for model training and validation to determine the parameters of the XGBoost classification model in order to determine the optimal XGBoost classification model to achieve stability analysis of yellow pages numbers.

[0015] Optionally, dynamically adjusting the execution of the yellow page number maintenance strategy based on the yellow page number stability result includes:

[0016] The yellow pages number stability result includes a label result output based on the number determination model and a corresponding probability value.

[0017] Optionally, dynamically adjusting the execution of the yellow page number maintenance strategy based on the yellow page number stability result includes:

[0018] The label results include stable results and unstable results. When the label result of the yellow page number is a stable result, its status is updated to stable in the yellow page number database, and the label result and the corresponding probability value are recorded;

[0019] When the label result of the yellow page number is unstable, if its corresponding probability value is greater than a preset threshold, the yellow page number is marked as a high-risk state.

[0020] Optionally, when the label result of the yellow page number is unstable, if the corresponding probability value is greater than a preset threshold, marking the yellow page number as a high-risk state includes:

[0021] A secondary verification is performed on the yellow page numbers that are screened out as high-risk to confirm the stability of the yellow page numbers; the secondary verification method includes web page query, outbound call system confirmation and manual investigation.

[0022] Optionally, the secondary verification of the yellow page numbers screened out as high-risk to confirm the stability of the yellow page numbers includes:

[0023] Based on the verification result of the secondary verification, the verified unstable yellow page numbers are classified into categories including repairable, unrepairable and suspected error information, so as to execute corresponding maintenance strategies.

[0024] Optionally, based on the verification result of the secondary verification, the verified unstable yellow page numbers are classified into categories including repairable, unrepairable, and suspected error information, so as to execute corresponding maintenance strategies, specifically including:

[0025] For yellow page numbers that can be repaired, notes are added and they are continuously tracked and updated after their information returns to normal; for yellow page numbers that cannot be repaired, they are deleted or marked as invalid, and their information in the yellow page number database is updated; for yellow page numbers that are suspected of having erroneous information, manual repair methods are used, and their information in the yellow page number database is updated.

[0026] Optionally, the number determination model adopts an XGBoost classification model, and the model training step further includes:

[0027] It includes parameter optimization and model evaluation of the XGBoost classification model, using the GridSearchCV function to optimize the parameters of the XGBoost classification model, the parameters of the XGBoost classification model including the learning rate, training rounds and tree structure depth; and using the confusion matrix, ROC characteristic curve and AUC value to evaluate the performance of the XGBoost classification model.

[0028] Correspondingly, the present application provides a yellow pages number analysis and control device based on deep learning, comprising:

[0029] At least one processor; and at least one memory in communication with the processor, wherein: the memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute any of the analysis and control methods described above.

[0030] The present application provides a yellow page number analysis and control method and device based on deep learning, which obtains the call behavior data of the yellow page number library, inputs it into the optimized XGBoost classification model, and obtains the stability results of the yellow page number through the gradient boosting tree algorithm. The call behavior feature data is classified and sorted, and the sample number data set is divided into a training set and a validation set during model training. The model parameters are determined by 5-fold cross-validation. The maintenance strategy is dynamically adjusted according to the stability results, including updating the status of stable numbers, marking high-risk numbers, etc. The high-risk numbers are verified twice, and based on the verification results, the unstable numbers are classified as repairable, unrepairable, and suspected error information, and the corresponding maintenance strategy is implemented. The model training step also includes parameter optimization and performance evaluation. The above scheme can accurately analyze the stability of yellow page numbers, improve the data quality of the number library, promptly discover unstable numbers and classify them, provide users with more reliable services, and optimize model parameters to improve performance and adapt to the ever-changing number data characteristics and usage environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0032] Figure 1 A flowchart of a yellow pages number analysis and control method based on deep learning provided in an embodiment of the present application;

[0033] Figure 2 A schematic diagram of the structure of a yellow pages number analysis control device based on deep learning provided in an embodiment of the present application. DETAILED DESCRIPTION

[0034] The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0035] It should be clear that the following embodiments of the present disclosure are described through specific concrete examples, and those skilled in the art can easily understand other advantages and effects of the present disclosure from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all the embodiments. The present disclosure 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 disclosure. It should be noted that the following embodiments and features in the embodiments can be combined with each other in the absence of conflict. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present disclosure.

[0036] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on this disclosure, it should be understood by those skilled in the art that an aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects described herein can be used to implement the device and / or practice the method. In addition, other structures and / or functionalities other than one or more of the aspects described herein can be used to implement this device and / or practice this method.

[0037] It should also be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present disclosure. The illustrations only show components related to the present disclosure and are not drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component can be changed at will, and the component layout type may also be more complicated.

[0038] Additionally, in the following description, specific details are provided to provide a thorough understanding of the examples. However, one skilled in the art will appreciate that the aspects described can be practiced without these specific details.

[0039] Example 1

[0040] Yellow Pages numbers serve as a crucial bridge for effective communication and business dealings between businesses, institutions, and service providers and their customers, partners, and more. Accurate Yellow Pages numbers ensure that potential customers can quickly contact relevant businesses, facilitating business transactions and collaborations. Existing Yellow Pages number management technology typically relies on manual collection and organization of number information, which is then entered into a database for management. This approach makes it difficult to promptly detect changes in number ownership, leading to users frequently encountering invalid or incorrect numbers when using Yellow Pages numbers, impacting normal communication and business operations. Furthermore, number verification relies primarily on numerous outbound calls and manual verification to confirm the validity and accuracy of numbers, which not only consumes significant manpower, time, and financial costs, but also produces poor results and is prone to omissions and errors.

[0041] like Figure 1 As shown, the present application provides a yellow page number analysis and control method based on deep learning, which uses a deep learning algorithm to analyze yellow page numbers, can timely change and maintain the information of yellow page numbers, reduce costs and improve management efficiency, and the method includes:

[0042] S1. Obtain call behavior data from the yellow pages number database.

[0043] Optionally, there are many ways to obtain call behavior data from the yellow pages number database, among which querying the database is one of the common methods. If there are fields or tables storing call behavior data in the database, SQL statements can be used to accurately extract it. It also includes calling the data interface. After finding the interface documentation, write code in a programming language to call the interface to obtain call behavior data. You can also analyze system log files and parse call behavior data from the logs with the help of log analysis tools or custom scripts. In addition, cooperation with telecom operators can obtain more accurate and comprehensive call behavior data. On this basis, the call behavior data is preprocessed, and the specific steps of preprocessing include cleaning the acquired call behavior data, removing duplicate data, abnormal data and erroneous data, and standardizing the data, converting the call time into a unified time format, and converting the call duration into a unified unit.

[0044] Optionally, the call behavior feature data is classified and organized to obtain the call behavior feature data of each yellow page number in the user's call behavior data, wherein the call behavior feature data includes the call frequency within a specific time period, the difference in call behavior between weekdays and weekends, the call frequency difference, the frequency variance, and the number history label feature. By classifying and organizing the call behavior data, the usage and stability of the yellow page numbers can be analyzed more comprehensively and accurately. Specifically, abnormal call behavior, such as changes in call frequency, large differences between weekdays and weekends, etc., can be discovered in a timely manner, so that yellow page numbers with possible problems can be dealt with in a focused manner; at the same time, incorporating the number history label features of the yellow page numbers into the call behavior feature data can improve the accuracy of the number reliability judgment, reduce the troubles caused to users by wrong or harassing numbers, help improve the overall quality and service level of the yellow page number database, and provide users with a more reliable and efficient information query experience.

[0045] For example, call behavior feature data includes multiple types of data, among which the call frequency within a specific time interval is used to count the number of calls made by the Yellow Pages number within the corresponding time period; it also includes the call frequency in the past 1-4 weeks and the call frequency in the past 1-3 months that distinguish between weekdays / holidays, as well as the frequency that distinguishes working hours (9-18 o'clock). In addition, there are number history label features, such as label type and label number in the past 7-90 days. By counting the call frequency and changes in different time periods and scenarios, as well as recording the label status, it is possible to fully reflect the usage status and stability of the Yellow Pages number, providing an important basis for subsequent number judgment and analysis.

[0046] S2. Input the call behavior data into the optimized number determination model and obtain the stability results of the yellow page numbers through the gradient boosting tree algorithm.

[0047] Optionally, the number determination model adopts an XGBoost classification model, and the model training steps include: obtaining a sample number data set with labeled results; randomly splitting the sample number data set into a training set and a validation set in a ratio of 7:3, and constructing an XGBoost classification model for the training set based on the XGBoost algorithm; using a 5-fold cross-validation method to train and validate the model, determine the XGBoost classification model parameters, and determine the optimal XGBoost classification model to implement the yellow page number stability analysis. In the above process, the parameters of the XGBoost classification model are continuously adjusted and determined to find the optimal model configuration, thereby achieving accurate analysis of the stability of the yellow page numbers.

[0048] Optionally, the number determination model adopts an XGBoost classification model, and the model training step further includes:

[0049] It includes parameter optimization and model evaluation of the XGBoost classification model. The GridSearchCV function is used to optimize the parameters of the XGBoost classification model. The parameters of the XGBoost classification model include learning rate, training rounds and tree structure depth; and the confusion matrix, ROC characteristic curve and AUC value are used to evaluate the performance of the XGBoost classification model.

[0050] Specifically, in terms of parameter optimization for the XGBoost classification model, the advanced GridSearchCV function was introduced. The GridSearchCV function can systematically search the parameter space of the XGBoost classification model to find the optimal parameter combination. Key parameters of the XGBoost classification model include the learning rate, the number of training rounds, and the tree structure depth. The learning rate determines the step size of the model's parameter update in each iteration. If the learning rate is too high, the model may oscillate during the optimization process and have difficulty converging to the optimal solution. The number of training rounds represents the number of times the model learns from the training data. The tree structure depth controls the complexity of the decision tree in the XGBoost model.

[0051] Specifically, to accurately evaluate the performance of the XGBoost classification model, we employed a variety of evaluation metrics, including the confusion matrix, ROC curve, and AUC. The confusion matrix provides a visual representation of the model's classification accuracy. It compares the model's predicted results with the actual results, categorizing them into four categories: true positives, false positives, true negatives, and false negatives. Using the confusion matrix, we can calculate metrics such as the model's accuracy, precision, and recall, providing a comprehensive understanding of the model's performance across different categories. The ROC curve plots the relationship between the true positive rate and the false positive rate by varying the classification threshold. The closer the ROC curve is to the upper left corner, the better the model's performance. The AUC (Area Under the Curve) is the area under the ROC curve and is a quantitative metric used to measure the model's overall performance. AUC values ​​closer to 1 indicate better model performance; AUC values ​​closer to 0.5 indicate that the model's performance is comparable to random guessing. By combining these evaluation metrics, we can comprehensively and objectively evaluate the performance of the XGBoost classification model, providing a strong basis for further model optimization and application.

[0052] S3. Based on the stability results of the yellow page numbers, dynamically adjust the execution of the yellow page number maintenance strategy.

[0053] Optionally, the yellow pages number stability result includes a label result and a corresponding probability value output based on the number determination model. The label result includes a stable result and an unstable result. When the label result for a yellow pages number is a stable result, its status is updated to stable in the yellow pages number database, and the label result and corresponding probability value are recorded. When the label result for a yellow pages number is unstable, if the corresponding probability value is greater than a preset threshold, the yellow pages number is marked as high-risk.

[0054] Specifically, the stability results for Yellow Pages numbers are output by the number determination model, which includes clear label results and corresponding probability values. Label results are categorized into two types: stable and unstable. When a Yellow Pages number's label result is displayed as stable, its status is promptly updated in the Yellow Pages number database, and the label result and corresponding probability value are recorded in detail.

[0055] For example, when the label result of a yellow page number is unstable, if its corresponding probability value is greater than a preset threshold, the yellow page number will be marked as a high-risk state. For example, if the label result of another number is unstable and the probability value is 0.92, if the preset threshold is 0.9, then the number will be identified as a high-risk state. In actual applications, the above-mentioned model can be used to conduct a comprehensive stability assessment of the numbers to obtain accurate label results and corresponding probability values ​​for each number. Based on the prediction results of the model, the numbers to be verified for the second time can be effectively screened. For example, numbers with an "unstable" label and a probability value greater than a specific value (such as 0.9) can be selected to form a high-risk set of numbers to be verified. And the preset threshold is not fixed and can be flexibly adjusted according to specific circumstances such as actual application scenarios and manpower arrangements. For example, when manpower is relatively sufficient, the probability value standard can be appropriately lowered to more widely screen numbers that may be at risk for verification.

[0056] Optionally, when the label result of the yellow page number is unstable, if the corresponding probability value is greater than a preset threshold, the yellow page number is marked as a high-risk state, including:

[0057] A secondary verification is conducted on the yellow page numbers that have been screened out as high-risk to confirm the stability of the yellow page numbers; the secondary verification methods include web page query, outbound call system confirmation and manual investigation.

[0058] Based on the verification results of the secondary verification, the unstable yellow page numbers after verification are classified into repairable, unrepairable and suspected error information, so as to implement the corresponding maintenance strategy.

[0059] For yellow page numbers that can be repaired, notes are added and they are continuously tracked and updated after their information returns to normal; for yellow page numbers that cannot be repaired, they are deleted or marked as invalid, and their information in the yellow page number database is updated; for yellow page numbers that are suspected of having erroneous information, manual repair methods are used, and their information in the yellow page number database is updated.

[0060] Optionally, after obtaining the verification results through secondary verification, the yellow pages numbers that are still unstable after verification are carefully classified into three main types: repairable, unrepairable, and suspected error information, so as to implement the corresponding maintenance strategy. For yellow pages numbers of the repairable type, notes will be added to mark them, and they will be continuously tracked. Once their information returns to normal, they will be updated in a timely manner. For yellow pages numbers of the unrepairable type, they will be deleted or marked as invalid, and their information status in the yellow pages number database will be updated simultaneously. For yellow pages numbers of the suspected error information type, manual repair will be adopted, and professionals will carefully verify and correct them, and update their information in the yellow pages number database.

[0061] Specifically, web search, outbound call system confirmation, and manual investigation are all methods for confirming the stability of Yellow Pages numbers. Web search uses search engines to search for information related to the number, such as the company's official website, business information, and user reviews. It is convenient and widely used, but the authenticity of the information must be verified. Outbound call system confirmation uses specialized equipment or software to call the number. This can automate batch dialing and combine voice recognition to improve efficiency. The number's status is determined based on the answering situation, but is affected by the answer rate and cooperation level. Manual investigation is personally verified by professionals, who use various channels such as direct communication and searching for other contact information to gain an in-depth understanding of the actual situation of the number. It has high accuracy but low efficiency and is affected by personal experience and ability. The three technical methods can be used in combination to fully leverage their respective advantages, improve the accuracy and comprehensiveness of Yellow Pages number stability confirmation, and provide a reliable basis for subsequent classification and maintenance strategies.

[0062] Example 2

[0063] Based on the above embodiments, Figure 2 FIG. 1 is a structural diagram of a yellow pages number analysis control device based on deep learning in an embodiment of the present invention. Figure 2As shown, the device includes: at least one processor 301; and at least one memory 302 communicatively connected to the processor 301, wherein: the memory 302 stores program instructions that can be executed by the processor 301, and the processor 301 calls the program instructions to execute the yellow pages number analysis and control method based on deep learning provided by the above-mentioned embodiments, for example, including: obtaining call behavior data of the yellow pages number library; inputting the call behavior data into the optimized number determination model, and obtaining the yellow pages number stability result through the gradient boosting tree algorithm; based on the yellow pages number stability result, dynamically adjusting the execution of the yellow pages number maintenance strategy.

[0064] The present invention provides a deep learning-based yellow page number analysis and control method and device. This method acquires call behavior data from a yellow page number database, feeds it into an optimized XGBoost classification model, and uses a gradient boosting tree algorithm to determine the stability of the yellow page numbers. The method then dynamically adjusts maintenance strategies based on the stability results, including marking high-risk numbers, performing secondary verification on these numbers, and classifying unstable numbers as repairable, unrepairable, or suspected errors based on the verification results, and implementing corresponding maintenance strategies. This approach accurately analyzes the stability of yellow page numbers, improves database data quality, and promptly identifies and categorizes unstable numbers, providing users with more reliable services. It also optimizes model parameters to improve performance and adapt to the ever-changing characteristics of number data and usage environments.

[0065] The basic principles of the present disclosure have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, and effects mentioned in this disclosure are merely illustrative and not restrictive, and should not be construed as necessarily possessed by each embodiment of the present disclosure. Furthermore, the specific details disclosed above are provided for illustrative purposes and to facilitate understanding, rather than as limitations. These details do not limit the present disclosure to necessarily being implemented using these specific details.

[0066] In the present disclosure, relational terms such as first and second, etc. are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. The block diagrams of the devices, devices, equipment, and systems involved in the present disclosure are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As will be appreciated by those skilled in the art, these devices, devices, equipment, and systems can be connected, arranged, or configured in any manner. Words such as "including," "comprising," "having," and the like are open-ended words, meaning "including but not limited to," and can be used interchangeably therewith. The words "or" and "and" used herein refer to the words "and / or" and can be used interchangeably therewith, unless the context clearly indicates otherwise. The word "such as" used herein refers to the phrase "such as but not limited to," and can be used interchangeably therewith.

[0067] Additionally, as used herein, "or" used in a list of items beginning with "at least one" indicates a separate list, so that, for example, a list of "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the word "exemplary" does not mean that the example described is preferred or better than other examples.

[0068] It should also be noted that in the system and method of the present disclosure, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present disclosure.

[0069] Various changes, substitutions, and modifications may be made to the technology described herein without departing from the teachings defined by the appended claims. Moreover, the scope of the claims of this disclosure is not limited to the specific aspects of the processes, machines, manufactures, compositions of things, means, methods, and actions described above. Currently existing or later developed processes, machines, manufactures, compositions of things, means, methods, or actions that perform substantially the same function or achieve substantially the same results as the corresponding aspects described herein may be utilized. Accordingly, the appended claims include within their scope such processes, machines, manufactures, compositions of things, means, methods, or actions.

[0070] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0071] The above description has been provided for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present disclosure to the forms disclosed herein. Although a number of example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A yellow pages number analysis and control method based on deep learning, characterized in that: The method comprises: Obtain call behavior data from the Yellow Pages number database; Input the call behavior data into the optimized number determination model, and obtain the stability result of the yellow page number through the gradient boosting tree algorithm; Based on the yellow pages number stability result, the execution of the yellow pages number maintenance strategy is dynamically adjusted.

2. The yellow page number analysis control method based on deep learning according to claim 1, characterized in that: The call behavior characteristic data are classified and sorted to obtain the call behavior characteristic data of each yellow pages number in the user's call behavior data, wherein the call behavior characteristic data includes the call frequency within a specific time period, the difference in call behavior between weekdays and weekends, the call frequency difference, the frequency variance and the number history label characteristics.

3. The yellow page number analysis and control method based on deep learning according to claim 1, characterized in that: The number determination model adopts the XGBoost classification model, and the model training steps include: Get a sample number dataset with labeled results; The sample number data set is randomly split into a training set and a validation set in a ratio of 7:3, and an XGBoost classification model is constructed for the training set based on the XGBoost algorithm; The 5-fold cross validation method was used for model training and verification to determine the parameters of the XGBoost classification model, so as to determine the optimal XGBoost classification model to realize the stability analysis of yellow pages numbers.

4. The yellow page number analysis control method based on deep learning according to claim 1, characterized in that: The dynamically adjusting the execution of the yellow page number maintenance strategy based on the yellow page number stability result includes: The yellow pages number stability result includes a label result output based on the number determination model and a corresponding probability value.

5. The yellow page number analysis control method based on deep learning according to claim 4 is characterized in that: The dynamically adjusting the execution of the yellow page number maintenance strategy based on the yellow page number stability result includes: The label result includes a stable result and an unstable result. When the label result of the yellow page number is a stable result, its status is updated to stable in the yellow page number database, and the label result and the corresponding probability value are recorded; When the label result of the yellow page number is unstable, if its corresponding probability value is greater than a preset threshold, the yellow page number is marked as a high-risk state.

6. The yellow page number analysis control method based on deep learning according to claim 5, characterized in that: When the label result of the yellow page number is unstable, if the corresponding probability value is greater than a preset threshold, the yellow page number is marked as a high-risk state, including: A secondary verification is performed on the yellow page numbers that are screened out as high-risk to confirm the stability of the yellow page numbers; the secondary verification method includes web page query, outbound call system confirmation and manual investigation.

7. The yellow page number analysis control method based on deep learning according to claim 6, characterized in that: The secondary verification of the yellow page numbers screened out as high-risk numbers to confirm the stability of the yellow page numbers includes: Based on the verification result of the secondary verification, the verified unstable yellow page numbers are classified into repairable, unrepairable and suspected error information, so as to execute corresponding maintenance strategies.

8. The yellow page number analysis control method based on deep learning according to claim 7, characterized in that: The verification result based on the secondary verification is used to classify the verified unstable yellow page numbers into repairable, unrepairable and suspected error information, so as to execute corresponding maintenance strategies, including: For yellow page numbers that can be repaired, notes are added, and they are continuously tracked and updated after their information returns to normal; yellow page numbers that cannot be repaired are deleted or marked as invalid, and their information in the yellow page number database is updated; yellow page numbers that are suspected of having erroneous information are manually repaired, and their information in the yellow page number database is updated.

9. The yellow page number analysis control method based on deep learning according to claim 3, characterized in that: The number determination model adopts an XGBoost classification model, and the model training step also includes: It includes parameter optimization and model evaluation of the XGBoost classification model, using the GridSearchCV function to optimize the parameters of the XGBoost classification model, the parameters of the XGBoost classification model include learning rate, training rounds and tree structure depth; and using confusion matrix, ROC characteristic curve and AUC value to evaluate the performance of the XGBoost classification model.

10. A yellow pages number analysis control device based on deep learning, characterized in that: include: at least one processor; And at least one memory communicatively connected to the processor, wherein: the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute the analysis and control method as described in any one of claims 1 to 9.

Images