Port health assessment method, device, equipment and storage medium

Abstract

The present invention discloses a port health assessment method, device, equipment and storage medium, including: obtaining multi-dimensional SMS port data within a preset time period; constructing a port relationship diagram based on the SMS port data; aggregating the SMS port data based on the port relationship diagram and the time window to obtain multiple aggregated variable data associated with any main port; and using a health assessment model to predict the health assessment result of any main port based on the multiple aggregated variable data associated with any main port. The present invention identifies potential risk points and risk propagation paths by constructing a relationship network between the main port and the sub-ports, and aggregates the SMS port data of the preset time period based on the port relationship diagram and the time window to predict the health assessment result of the main port according to the multiple aggregated variable data of the time window, thereby solving the problem that the current date data cannot accurately highlight the health status of the main port and improving the accuracy of the port health assessment.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a port health assessment method, device, equipment and storage medium. Background Art

[0002] In recent years, with the rapid development of the internet and the widespread adoption of smart devices, SMS has evolved from being just a traditional communication tool to a crucial bridge between businesses and users. However, statistics show that despite an increase in overall SMS volume, service revenue has declined. This is due to a combination of crackdowns on spam text messages and the emergence of new communication methods. Furthermore, public pressure against spam text messages has led many businesses to become more cautious before sending text messages. Therefore, it's crucial to assess the health of primary SMS endpoints to plan optimal delivery plans.

[0003] The existing SMS primary port health scoring method relies on expert scoring. This method relies on operational expertise to assess the health of the primary port number based on the current date's basic SMS primary port metrics (such as send volume, complaint volume, and complaint ratio per million). However, this expert scoring method has the following drawbacks:

[0004] (1) Lag: After a text message is sent, complaints may reach a peak on the Nth day and then decline. Therefore, the number of complaints on the current date is related to the number of messages sent on the current date, but it is not accurate enough.

[0005] (2) Biasedness: Verification of sub-ports involved in the case and user complaints are contagious. The same signature may be sent to multiple main ports, but only the number of complaints and verifications involved in the case from main port A is large, or the number of complaints from other main ports is only temporarily small. In this case, only the increased risk of main port A will be reported, and other risky main ports may be missed, which will eventually lead to the main port being restricted or shut down, affecting revenue.

[0006] (3) Subjectivity: Evaluating the health of the primary port based on business experience is subjective and cannot draw on a large amount of historical data as a reference. In addition, the real-time performance is low. Summary of the Invention

[0007] Based on this, it is necessary to provide a port health assessment method, device, equipment and storage medium for the above technical problems to solve at least one of the above technical problems.

[0008] The present invention provides a port health assessment method, comprising:

[0009] Obtain multi-dimensional SMS port data within a preset time period;

[0010] Based on the SMS port data, a port relationship diagram corresponding to each main port and each sub-port is constructed;

[0011] Aggregating the SMS port data based on the port relationship graph and a preset time window to obtain multiple aggregated variable data associated with any main port;

[0012] Based on a plurality of aggregated variable data associated with any of the master ports, a pre-built health evaluation model is used to predict and obtain a health evaluation result corresponding to any of the master ports.

[0013] Optionally, according to a port health assessment method provided by the present invention, the health assessment model is trained based on the following steps:

[0014] Obtaining sample data of port status and multiple indicator variables associated with multiple primary ports;

[0015] For any indicator variable: binning the indicator variable to obtain multiple bin intervals;

[0016] Based on the sample data corresponding to the indicator variable, determine the weight value corresponding to each bin interval;

[0017] The health assessment model is trained based on the weight value corresponding to each binning interval.

[0018] Optionally, according to a port health assessment method provided by the present invention, the health assessment result corresponding to any of the master ports is predicted using a pre-built health assessment model based on multiple aggregated variable data associated with the master port, including:

[0019] For any primary port:

[0020] Filtering and obtaining a plurality of target variable data based on the plurality of aggregate variable data associated with the primary port;

[0021] Performing a joint operation on the multiple aggregated variable data associated with the primary port to obtain multiple derived variable data;

[0022] Binning the target variable data and the derived variable data, and determining the target weight value corresponding to each bin;

[0023] Based on each of the target weight values, the health evaluation model is used to predict and obtain a health evaluation result corresponding to the primary port.

[0024] Optionally, according to a port health assessment method provided by the present invention, the port status includes a normal state, a notification state, a current limiting state, and a shutdown state;

[0025] The step of determining a weight value corresponding to each binning interval based on the sample data corresponding to the indicator variable includes:

[0026] Based on the sample data corresponding to the indicator variable, determine the number of samples in each bin interval in which the port status is in the normal state, the notification state, the current limiting state, and the shutdown state;

[0027] Determine the number of abnormal samples based on the number of samples in the notification state, current limiting state, and shutdown state and the associated weights;

[0028] Based on the number of samples in a normal state and the number of abnormal samples in each binning interval, a weight value corresponding to each binning interval is determined.

[0029] Optionally, according to a port health assessment method provided by the present invention, the SMS port data is aggregated based on the port relationship graph and a preset time window to obtain multiple aggregate variable data associated with any primary port, including:

[0030] Dividing the SMS port data within the preset time period according to the time window to obtain target divided data within multiple time windows;

[0031] In combination with the port relationship diagram, aggregation processing is performed on the target partition data in each time window to obtain multiple aggregate variable data associated with any main port.

[0032] Optionally, according to a port health assessment method provided by the present invention, constructing a port relationship graph corresponding to each main port and each sub-port based on the SMS port data includes:

[0033] Determine the sub-port information corresponding to the SMS port data;

[0034] Extracting the main port information based on the sub-port information;

[0035] Based on the association relationship between the sub-port information and the main port information, a port relationship graph corresponding to each main port and each sub-port is constructed.

[0036] Optionally, according to a port health assessment method provided by the present invention, the step of obtaining multi-dimensional SMS port data within a preset time period includes:

[0037] Collect SMS complaint data, daily sub-port sending volume data, sub-port real-name registration data, case-related sub-port verification data, and risk signature data within a preset time period. The case-related sub-port verification data refers to the verification data of the sub-port corresponding to the risk event involved in the case;

[0038] The SMS complaint data, the sub-port daily sending volume data, the sub-port real-name registration data, the sub-port verification data involved in the case and the neutral signature data are analyzed to form the multi-dimensional SMS port data.

[0039] The present invention also provides a port health assessment device, comprising:

[0040] The acquisition module is used to obtain multi-dimensional SMS port data within a preset time period;

[0041] A construction module, configured to construct a port relationship diagram corresponding to each main port and each sub-port based on the SMS port data;

[0042] an aggregation module, configured to aggregate the SMS port data based on the port relationship graph and a preset time window to obtain a plurality of aggregated variable data associated with any primary port;

[0043] The evaluation module is used to predict the health evaluation result corresponding to any of the master ports based on multiple aggregated variable data associated with any of the master ports using a pre-built health evaluation model.

[0044] The present invention also provides a computer device, comprising a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, wherein the processor implements the above-mentioned port health evaluation method when executing the computer-readable instructions.

[0045] The present invention also provides one or more readable storage media storing computer-readable instructions, which implement the above-mentioned port health evaluation method when executed by a processor.

[0046] The above-mentioned port health assessment method, device, equipment and storage medium include: obtaining multi-dimensional SMS port data within a preset time period; constructing a port relationship diagram corresponding to each main port and each sub-port based on the SMS port data; aggregating the SMS port data based on the port relationship diagram and a preset time window to obtain multiple aggregated variable data associated with any main port; and using a pre-constructed health assessment model to predict the health assessment result corresponding to any of the main ports based on the multiple aggregated variable data associated with any of the main ports. The present invention can identify potential risk points and risk propagation paths by constructing a relationship network between the main port and each sub-port, and take timely measures to prevent the spread of risks. In addition, based on the port relationship diagram and a preset time window, the SMS port data is aggregated to obtain multiple aggregated variable data associated with any of the main ports, thereby predicting the health assessment result corresponding to the main port according to the data of each time window, solving the problem that the current date data cannot accurately highlight the health status of the main port, and effectively improving the accuracy of the port health assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0048] Figure 1 This is a flow chart of a method for evaluating port health in accordance with an embodiment of the present invention;

[0049] Figure 2 is a schematic diagram of a port relationship diagram provided by an embodiment of the present invention;

[0050] Figure 3 This is a structural diagram of a port health evaluation device according to an embodiment of the present invention;

[0051] Figure 4 FIG. 1 is a schematic diagram of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0052] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0053] The terms used in one or more embodiments of the present invention are for the purpose of describing specific embodiments only and are not intended to limit one or more embodiments of the present invention. The singular forms "a", "the" and "the" used in one or more embodiments of the present invention are also intended to include plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in one or more embodiments of the present invention refers to and includes any or all possible combinations of one or more associated listed items.

[0054] In one embodiment, specifically, Figure 1 As shown, Figure 1 FIG. 1 is a flow chart of a method for evaluating port health according to an embodiment of the present invention. The method includes the following steps:

[0055] Step S11, obtaining multi-dimensional SMS port data within a preset time period;

[0056] Specifically, SMS complaint data, daily sub-port send volume data, sub-port real-name registration data, sub-port verification data, and risk signature data within a preset time period are collected. SMS complaint data includes information such as sub-port information, complaint channel fields, and the content of SMS messages complained by end users. Daily sub-port send volume data includes information such as sub-port information and the daily SMS send volume of the sub-port. Sub-port real-name registration data includes information such as sub-port information, the signature for sub-port real-name registration, and the port service type. In one embodiment, the port service type can be set to: account registration (business management and service type), account login (business management and service type), advertising and promotion (commercial type), notification reminder (business management and service type), public service (public welfare type), etc. Sub-port verification data related to the case refers to verification data for sub-ports corresponding to the risk events involved in the case, for example, information such as sub-port information, SMS signatures, and SMS content corresponding to the risk events involved in the case. Risk signature data refers to SMS signature data containing non-compliant signatures such as fraud, gambling, or signatures that do not clearly state the sender's company name or logo.

[0057] Furthermore, the SMS complaint data, sub-port daily sending volume data, sub-port real-name registration data, case-related sub-port verification data, and neutral signature data are analyzed. In one embodiment, the analysis process specifically includes the following operations:

[0058] SMS Signature Extraction: Extract SMS signatures from SMS complaint data. Optionally, extract SMS signatures from SMS complaint data using word segmentation and string matching. SMS signatures are categorized into three types: [Signature], [Signature], and {Signature}. The priority is [Signature] > [Signature] > {Signature}, meaning that the highest-priority SMS signature is extracted first.

[0059] In addition, main port information extraction: Based on the sub-port information in SMS complaint data, sub-port daily sending volume data, sub-port real-name registration data, case-related sub-port verification data, and risk signature data, the main port information covered by each sub-port information is extracted. The extraction steps are as follows:

[0060] If the subport subPort1 exists in the preset primary port data, it is determined that the subport subPort1 is the primary port. It should be noted that the primary port data is configured with information of multiple known primary ports.

[0061] If the subport subPort1 does not exist in the main port data, the last bit of the subport subPort1 is deleted to obtain subPort2, and it is re-determined whether the subport subPort2 with the last bit deleted exists in the main port data.

[0062] If it does not exist, continue to delete the last bit of the subport and assign it to subPort2, and then proceed to the judgment process. If it does exist, determine that subport subPort2 is the master port of subPort1.

[0063] In addition, we use open-source AI models to identify the industry categories of SMS content within SMS complaint data. These categories include internet lending, pornography and terrorism-related content, points redemption, insurance, retail, education marketing, lifestyle services, app promotion, gaming and pharmaceuticals, enterprise service marketing, banking, POS machines, stocks and securities, fund management, telecommunications, dating, travel promotion, real estate, and debt collection.

[0064] In addition, SMS type labeling: use open source AI large models for natural language recognition to identify the SMS types in SMS complaint data. For example, SMS types include marketing and notification types.

[0065] In addition, signature compliance verification: the compliance of the SMS signature is comprehensively judged based on the SMS content in the SMS complaint data, the extracted SMS signatures and the sub-port real-name registration data. For example, it is judged whether the sub-port corresponding to the SMS content in the SMS complaint data exists in the sub-port real-name registration data, and whether the SMS signature obtained in the signature extraction stage is consistent with the signature in the sub-port real-name registration data. If both are met, the SMS signature is judged to be compliant, so that a non-compliant signature can be obtained based on the signature compliance verification results.

[0066] In addition, compliance verification of mixed use of SMS entities is carried out: based on the SMS signature and SMS type extracted from the SMS complaint data, for the same sub-port, if within the preset time (for example, the same month), the SMS type is the same and there is only one SMS signature, the sub-port is judged to be compliant, otherwise it is non-compliant.

[0067] Furthermore, service type compliance verification is performed: Subports are tagged based on the port service type in the subport real-name registration basic data table. In one embodiment, port service types include 1-Account Registration, 2-Account Login, 3-Advertising and Promotion, 4-Notification and Reminder, and 5-Public Service. A compliant combination of these five types is 124, 3, and 5. Any other combination, such as 123, 125, or 25, is considered non-compliant, thus obtaining the main port service type compliance verification result.

[0068] In addition, the cancellation specifications in the SMS content are reviewed: the cancellation specifications are determined based on the SMS content in the SMS complaint data, and the cancellation method is judged to be consistent with the preset cancellation words. For example, the preset cancellation words are set to "Reply R to refuse". If they are consistent, the cancellation specifications are determined to be passed. If they are inconsistent or there is no clear cancellation method, the cancellation specifications are determined to be failed.

[0069] In addition, risk ratings are applied to the industry types of text messages: for example, industries involving Internet lending, pornography, and terrorism are classified as severe high risk; industries involving points redemption, insurance, and retail are classified as high risk; industries involving education marketing, life services, APP promotion, gaming medicine, corporate service marketing, and banking are classified as medium risk; industries involving POS machines, stocks and securities, fund management, telecommunications services, dating, travel promotion, and real estate are classified as low risk; and industries involving debt collection are classified as exempt.

[0070] In addition, severe risk signature marking: Based on the signature of the sub-port in the sub-port real-name registration data, determine whether the signature exists in the risk signature data. If not, mark it as no; if it exists, determine the million-complaint ratio and total number of complaints corresponding to the signature; then compare the million-complaint ratio with the first preset threshold, and compare the number of complaints with the second preset threshold. If the million-complaint ratio is greater than the first preset threshold, or the million-complaint ratio is less than or equal to the first preset threshold, and the number of complaints is greater than or equal to the second preset threshold, or the signature only has complaints but no data is sent, then the signature is determined to be a severe risk signature. For example, if any of the following conditions is met, it is identified as a severe risk signature:

[0071] A: The complaint ratio per million is greater than 1.18.

[0072] B: The complaint ratio per million is less than or equal to 1.18, and the number of complaints is greater than or equal to 27.

[0073] C: The signature only contains complaint but no sent data.

[0074] In addition, silent sub-port mark: based on the registration time of the sub-port in the sub-port real-name registration data and the date in the sending volume data, if the sub-port has not sent any text messages within the preset time (for example, 6 months), the sub-port is determined to be a silent sub-port.

[0075] Furthermore, based on the above analysis results, multi-dimensional SMS port data is formed; for example, the multi-dimensional SMS port data includes the main port to which the sub-port belongs, the compliance verification results of the mixed use of the subject, the marking results of the silent sub-port, the daily SMS sending volume of the sub-port, whether the sub-port exists in the real-name registration data, the results of the signature extraction stage, the SMS industry category marking results, the SMS type marking results, the signature compliance verification marking results, the unsubscription specification judgment marking results, the SMS industry type risk rating marking results, the complaint channel field in the SMS complaint data, the port verification volume involved in the sub-port verification data, the number of signatures, the severe risk signature marking results and other information.

[0076] Step S12: constructing a port relationship diagram corresponding to each main port and each sub-port based on the SMS port data;

[0077] Specifically, determine the sub-port information corresponding to the SMS port data and the main port information to which the sub-port information belongs; then, based on the association between the sub-port information and the main port information, construct a port relationship diagram corresponding to each main port and each sub-port. In addition, when the same SMS signature has sending records in multiple sub-ports, it proves that the multiple sub-ports have an association relationship. Based on the association relationship between the multiple sub-ports and the association relationship between the sub-port information and the main port information, a port relationship diagram corresponding to each main port and each sub-port can be constructed. The port relationship diagram can be referred to Figure 2 , Figure 2 This is a schematic diagram of a port relationship diagram provided by an embodiment of the present invention. It illustrates the relationship between a main port and its subports. The subports' send and complaint volumes are the result of subsequent aggregation of SMS port data. By building a port relationship network, potential risk points and risk transmission paths can be identified, allowing timely measures to prevent risk spread and, to a certain extent, providing early warning of risks.

[0078] Step S13: Aggregate the SMS port data based on the port relationship diagram and the preset time window to obtain multiple aggregated variable data associated with any main port;

[0079] It should be noted that the number and value of the time windows can be set according to the specific situation. For example, the time windows can be set to 3 days, 7 days, 15 days, and 30 days respectively. Specifically, the SMS port data within the preset time period is divided according to the time window to obtain the target divided data within multiple time windows; combined with the port relationship diagram, the target divided data within each time window is aggregated to obtain multiple aggregated variable data associated with any main port. The specific data aggregation process is described in the following embodiment and will not be repeated here.

[0080] Step S14 : Based on the multiple aggregated variable data associated with any master port, a pre-built health evaluation model is used to predict and obtain a health evaluation result corresponding to any master port.

[0081] Specifically, for the multiple aggregated variable data associated with any main port, the aggregated variable data is screened to select the variables with high correlation with the main port status, and obtain multiple target variable data. Because after the flow is limited or shut down, the sending volume data of the next day will decrease or be 0, but other indicators, such as complaint data, will continue to grow or have a downward trend, so these data will affect the normal judgment results and need to be filtered out. In addition, the multiple aggregated variable data associated with the main port are jointly calculated to obtain multiple derived variable data; for example, the derived variable data includes the complaint ratio per million, the real-name registration rate of sub-ports, the proportion of silent sub-ports, the growth rate of the complaint ratio per million, and other data. Further, each target variable data and each derived variable data are binned, and the target weight value WOE corresponding to each bin is determined; then, based on the target weight values ​​corresponding to each target variable data and each derived variable data, the health assessment model is used to predict the health assessment result corresponding to the main port. The training process of the health assessment model is specifically described in the following embodiments and will not be repeated here. By continuously monitoring and optimizing the health of the main port number, the sending of spam and fraudulent messages is effectively reduced, the purity and security of the information received by users is guaranteed, and the user's communication experience and satisfaction are significantly improved.

[0082] The embodiment of the present invention, through the above scheme, includes: obtaining multi-dimensional SMS port data within a preset time period; constructing a port relationship diagram corresponding to each main port and each sub-port based on the SMS port data; aggregating the SMS port data based on the port relationship diagram and a preset time window to obtain multiple aggregated variable data associated with any main port; and using a pre-built health assessment model to predict the health assessment result corresponding to any main port based on the multiple aggregated variable data associated with any main port. By constructing a relationship network between the main port and each sub-port, the embodiment of the present invention can identify potential risk points and risk propagation paths, and take timely measures to prevent the spread of risks. In addition, based on the port relationship diagram and a preset time window, the SMS port data is aggregated to obtain multiple aggregated variable data associated with any main port, thereby predicting the health assessment result corresponding to the main port according to the data of each time window, solving the problem that the current date data cannot accurately highlight the health status of the main port, and effectively improving the accuracy of the port health assessment.

[0083] In one embodiment of the present invention, based on the port relationship graph and a preset time window, SMS port data is aggregated to obtain multiple aggregate variable data associated with any primary port, including:

[0084] According to the time window, the SMS port data within the preset time period is divided to obtain the target division data in multiple time windows; combined with the port relationship diagram, the target division data in each time window is aggregated respectively to obtain multiple aggregate variable data associated with any main port.

[0085] Specifically, to address the issue of current date data failing to accurately highlight the health status of the primary port, this embodiment divides SMS port data within a preset time period into time windows, obtaining target segmented data within multiple time windows. In the data aggregation phase, MapReduce technology is used to aggregate massive amounts of data, ultimately yielding multi-dimensional aggregated variable data associated with the primary port, including the following data:

[0086] (1) Transmission volume: According to the main port number, the transmission volume data of all sub-ports associated with the main port number are summarized and the total transmission volume is obtained after aggregation.

[0087] (2) Number of sub-ports reporting severe risk signatures: The number of sub-ports reporting severe risk signatures is obtained by aggregating the sub-port real-name registration data and the marked severe risk signatures according to the main port query.

[0088] (3) Severe risk signature sending volume: query the sub-port real-name registration data and the marked severe risk signature according to the main port number, obtain the sub-port with the severe risk signature, summarize the sending volume data of these sub-ports according to the sub-port daily sending volume data, and obtain the severe risk signature sending volume after aggregation.

[0089] (4) Number of sending subports: query the daily sending volume data of the subport according to the main port number, obtain all subports and remove duplicates, and then aggregate to obtain the number of subports with sending records.

[0090] (5) Number of sub-ports registered with real names: query the daily sending volume data of the sub-ports according to the main port number, obtain the sub-ports that have been registered with real names and remove duplicates, and after aggregation, obtain the number of sub-ports with sending records and real-name registration.

[0091] (6) Sub-port real-name registration rate: Sub-port real-name registration rate = number of real-name registered sub-ports / number of sending sub-ports.

[0092] (7) Number of silent subports: Query the data marked as silent subports according to the main port number and perform deduplication processing, and obtain the number of silent subports after aggregation.

[0093] (8) Number of sub-ports used by the subject: The data marked as subject mixed is obtained based on the main port number query and deduplication is performed. After aggregation, the number of sub-ports used by the subject mixed is obtained.

[0094] (9) Total number of complaints across all channels: Query SMS complaint data based on the main port number, and aggregate to obtain the total number of complaints across all channels.

[0095] (10) Complaint volume of 12321 channel: query SMS complaint data based on the main port number, filter data whose complaint channel is 12321 channel, and obtain the total complaint volume corresponding to 12321 channel after aggregation.

[0096] (11) Number of complaints with severe high risk: query SMS complaint data based on the main port number, and filter out data with SMS industry risk rating of severe high risk, and obtain the total number of complaints with severe high risk after aggregation.

[0097] (12) High-label complaint volume: Query SMS complaint data based on the main port number, and filter out data with SMS industry risk ratings of high risk. After aggregation, the total number of high-label complaints is obtained.

[0098] (13) Number of complaints with medium label: query SMS complaint data based on the main port number, and filter out data with SMS industry risk rating of medium risk, and obtain the total number of complaints with medium label after aggregation.

[0099] (14) Low-label complaint volume: Filter data with a low-risk SMS industry risk rating based on the main port number query, and aggregate the data to obtain the total low-risk complaint volume.

[0100] (15) Number of complaints exempted from the label: query the SMS complaint data based on the main port number, and filter the data with SMS industry risk rating as exempted. After aggregation, the total number of complaints exempted from the label is obtained.

[0101] (16) Number of complaints about non-compliance with cancellation regulations: Query SMS complaint data based on the main port number, and filter out data that are judged to be non-compliant with the cancellation regulations. After aggregation, the total number of complaints about non-compliance with cancellation regulations is obtained.

[0102] (17) Number of complaints about signature non-compliance: query the SMS complaint data based on the main port number, and filter the data with non-compliant signature compliance verification results. After aggregation, the total number of complaints about signature non-compliance is obtained.

[0103] (18) Verification volume of ports involved in the case: query the verification volume data of ports involved in the case based on the main port number, and obtain the total number of complaints of the ports involved in the case after aggregation.

[0104] In other embodiments, based on information such as the number of sub-ports and signature anomaly data on the port relationship diagram, it is determined whether there is a risk of port data anomaly spreading. For example, when the number of port-related anomaly data is large, there is a risk of spreading. Therefore, in combination with the port relationship diagram, the target division data in each time window is aggregated and processed to obtain data such as the number of spread complaints, spread transmission volume, and spread per million complaint ratio associated with any main port. If the signature corresponding to the sub-port is also transmitted at other main ports, there is actually a certain risk. The spread complaint volume is determined based on the complaint volume of the main port and the complaint volume of the signature corresponding to the sub-port at other main ports. The spread transmission volume is determined based on the transmission volume of the main port and the transmission volume of the signature corresponding to the sub-port at other main ports. The spread per million complaint ratio is determined based on the complaint volume of the main port and the complaint volume and transmission volume of the signature corresponding to the sub-port at other main ports. The aggregated variable data can be referred to as shown in Table 1 below:

[0105] Table 1

[0106]

[0107] Through the above-mentioned solution, the embodiment of the present invention realizes the comprehensive association of complaint data, sending behavior patterns and real-name registration information of the main port number, and can quickly and accurately identify the source of potential service quality degradation, abuse or security vulnerabilities, providing operators with immediate risk warnings and governance basis.

[0108] In one embodiment of the present invention, the health assessment model is trained based on the following steps:

[0109] Obtain sample data of the port status associated with multiple main ports and multiple indicator variables; for any indicator variable: bin the indicator variable to obtain multiple bin intervals; based on the sample data corresponding to the indicator variable, determine the weight value corresponding to each bin interval; and based on the weight value corresponding to each bin interval, train a health assessment model.

[0110] It should be noted that port status includes normal, notification, current limiting, and shutdown. Sample data for the port status of multiple primary ports and multiple indicator variables is obtained. Sample data for each indicator variable is obtained by analyzing and aggregating SMS complaint data, daily subport send volume data, subport real-name registration data, case-related subport verification data, and risk signature data within a specific time period. The data analysis and aggregation process is similar to the data analysis and aggregation process described above and will not be repeated here. The analyzed data is filtered and processed (missing value processing, outlier processing, and data cleaning) to obtain basic variable data for model training. Derivative variables are then constructed based on this basic variable data. For example, 3, 7, 15, and 30 days are used as time windows to obtain basic variable data for various indicators (e.g., complaint volume and send volume). Multiple derived variables are generated by combining the data from each time slider with the data from other indicators. For example, there are four time sliders, and other metrics include complaint volume, send volume, and complaint ratio per million. The resulting number of derived variables = four sliders * other metrics. Complaint ratio per million: Data from each slider calculated using the time slider method. Complaint ratio per million = complaint volume / send volume. Growth rate: The growth rate of the ratio of complaint volume to send volume is very important when sending volume ranks below average.

[0111] Furthermore, if the signature corresponding to a sub-port involved in the case verification is also sent from other main ports but not involved in the case verification, it actually has a certain risk and can be processed according to the custom risk factor. The calculation method is: Number of verifications involved on the main port = Number of verifications involved on this main port + Number of verifications involved on other main ports * Custom factor.

[0112] After constructing the derived variables, variables with a high correlation with the main port status are screened out according to the maximum correlation principle. Furthermore, based on the screened basic variable data and the derived variable data, sample data of multiple indicator variables are formed. Optionally, the sample data of the indicator variables include the sending volume associated with the main port, the number of complaints, the complaint ratio per million, the real-name registration rate of sub-ports, the proportion of silent sub-ports, the growth rate of the complaint ratio per million, the spread complaint ratio per million, the number of case-related inspections, the compliance rate of sub-ports and other data.

[0113] Furthermore, for any indicator variable: the indicator variable is binned to obtain multiple bin intervals; for example, the sending volume is divided into multiple bin intervals such as 10,000 to 50,000 sending volumes, 50,000 to 300,000 sending volumes, and 300,000 to 1 million sending volumes, and the samples are divided into normal and abnormal (notification, current limiting, and shutdown). It should be noted that in the process of binning, some variables are binned according to the two levels of normal and abnormal, which will cause the weights of the bins to be less accurate. When the final evaluation model outputs the results, the discrimination of the abnormal levels (notification, current limiting, shutdown) is not high. For example: the shutdown state is the most serious state, and the shutdown situation is the least, and in some variables, the main ports in the shutdown state will be concentrated in the same bin, but due to the low number, the weight is low, which will result in a low final score, affecting the inference result. Therefore, in this embodiment, a weighted operation is performed in the process of calculating the weights of these variables. Optionally, weighting is applied during the binning process, that is, when the main ports in the current limiting and shutdown states are assigned to a certain bin, multiple binning operations are performed to increase the number of abnormal variables in this bin and increase the abnormal proportion of this bin. Specifically: based on the sample data corresponding to the indicator variable, determine the number of samples in each bin interval where the port status is in the normal state, notification state, current limiting state and shutdown state, as well as the normal number in the normal state, that is, count the number of notifications in the notification state, the number of current limiting in the current limiting state, the number of shutdowns in the shutdown state, and the normal number in the normal state in each bin interval respectively; then, based on the number of notifications in the notification state, the number of current limiting in the current limiting state, and the number of shutdowns in the shutdown state in each bin interval, calculate the number of abnormalities in each bin interval according to the preset weight coefficient; then, determine the weight value WOE corresponding to each bin interval based on the number of abnormalities and normal numbers in each bin interval. The calculation formula is as follows:

[0114] Abnormal number = Notification number + Current limit number * First preset weight coefficient + Shutdown number * Second preset weight coefficient

[0115] WOE = WOE = ln((abnormal number / abnormal number of all bins) / (normal number / normal number of all bins))

[0116] Furthermore, based on the weight value corresponding to each binning interval, a health assessment model is trained, wherein the health assessment model is a logistic regression algorithm model. In other embodiments, the IV (information value) value of each binning interval is calculated, and then the IV value of each binning interval is screened among the weight values ​​corresponding to each binning interval, and the health assessment model is trained using the weight values ​​corresponding to the binning intervals obtained by screening.

[0117] Through the above-mentioned scheme, the embodiment of the present invention realizes the comprehensive association of complaint data, sending behavior patterns and real-name registration information of the main port number, screens the variable data for model training, and trains the health assessment model by adjusting the variable weight coefficient, so as to more accurately assess the health status of the main port.

[0118] In one embodiment of the present invention, based on multiple aggregated variable data associated with any master port, a pre-built health assessment model is used to predict the health assessment result corresponding to any master port, including:

[0119] For any primary port:

[0120] Based on multiple aggregate variable data associated with the main port, multiple target variable data are screened and obtained; multiple aggregate variable data associated with the main port are jointly operated to obtain multiple derived variable data; each target variable data and each derived variable data are binned, and the target weight value corresponding to each bin is determined; based on each target weight value, the health assessment model is used to predict the health assessment result corresponding to the main port.

[0121] Specifically, the target variable data is obtained by screening multiple aggregate variable data associated with the main port.

[0122] For example, variable data such as the sending volume and complaint volume associated with the main port are selected; in addition, multiple aggregated variable data associated with the main port are jointly calculated to obtain multiple derived variable data, for example, derived variable data such as the complaint ratio per million, the real-name registration rate of sub-ports, the proportion of silent sub-ports, the growth rate of the complaint ratio per million, and the spread complaint ratio per million are calculated. Each target variable data and derived variable data are binned and the target weight value corresponding to each bin is calculated. The binning process and the calculation process of the target weight value WOE are described in the above embodiment and will not be repeated here. Furthermore, based on each target weight value, the health assessment model is used to predict the health assessment result corresponding to the main port. Therefore, based on the health assessment result of the main port, the operator can allocate resources more reasonably, provide more support to the main port numbers with excellent performance, and take targeted improvement measures for problem ports, such as adjusting the sending strategy and strengthening supervision, so as to achieve efficient resource utilization and effective cost control.

[0123] Through the above-mentioned scheme, the embodiment of the present invention realizes the complaint data, sending data and derived variable data of the main port number in each time window by joint operation of the aggregated variable data, and uses the health assessment model to predict the health assessment result corresponding to the main port, effectively solving the problem that date data cannot accurately highlight the health status of the main port, and improving the accuracy of the main port health assessment.

[0124] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0125] In one embodiment, a port health evaluation device is provided, which corresponds to the port health evaluation method in the above embodiment. Figure 3 As shown, Figure 3 FIG. 1 is a schematic diagram of a structure of a port health evaluation device according to an embodiment of the present invention, wherein the port health evaluation device comprises:

[0126] The acquisition module 21 is used to obtain multi-dimensional SMS port data within a preset time period;

[0127] A construction module 22 is used to construct a port relationship diagram corresponding to each main port and each sub-port based on the SMS port data;

[0128] Aggregation module 23, for aggregating SMS port data based on the port relationship diagram and a preset time window to obtain multiple aggregated variable data associated with any main port;

[0129] The evaluation module 24 is configured to predict a health evaluation result corresponding to any master port based on a plurality of aggregated variable data associated with any master port using a pre-built health evaluation model.

[0130] For the specific definition of the port health assessment device, please refer to the definition of the port health assessment method above and will not be repeated here. The various modules in the above-mentioned port health assessment device can be implemented in whole or in part through software, hardware, or a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of the above-mentioned modules.

[0131] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4 As shown, Figure 4This is a schematic diagram of a computer device in one embodiment of the present invention. The computer device includes a processor, a memory, a network interface, and a database connected via a device bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a readable storage medium and an internal memory. The readable storage medium stores an operating device, computer-readable instructions, and a database. The internal memory provides an environment for the operation of the operating device and computer-readable instructions in the readable storage medium. The database of the computer device is used to store data involved in the port health assessment method. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer-readable instructions are executed by the processor, a port health assessment method is implemented. The readable storage medium provided in this embodiment includes a non-volatile readable storage medium and a volatile readable storage medium.

[0132] In one embodiment, a computer device is provided. The computer device may be a terminal device, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, and a network interface connected via a device bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a readable storage medium. The readable storage medium stores computer-readable instructions. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer-readable instructions are executed by the processor, a port health assessment method is implemented. The readable storage medium provided in this embodiment includes a non-volatile readable storage medium and a volatile readable storage medium.

[0133] In one embodiment, a computer device is provided, including a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor. When the processor executes the computer-readable instructions, the steps of the port health assessment method described above are implemented.

[0134] In one embodiment, a readable storage medium is provided, which stores computer-readable instructions. When the computer-readable instructions are executed by a processor, the steps of the port health assessment method described above are implemented. A person skilled in the art will understand that all or part of the processes in the above-mentioned embodiment method can be implemented by instructing related hardware through computer-readable instructions. The computer-readable instructions can be stored in a non-volatile readable storage medium or a volatile readable storage medium. When the computer-readable instructions are executed, they can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0135] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0136] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A port health assessment method, characterized in that: include: Obtain multi-dimensional SMS port data within a preset time period; Based on the SMS port data, a port relationship diagram corresponding to each main port and each sub-port is constructed; Aggregating the SMS port data based on the port relationship graph and a preset time window to obtain multiple aggregated variable data associated with any main port; Based on multiple aggregated variable data associated with any of the master ports, a pre-built health assessment model is used to predict and obtain a health assessment result corresponding to any of the master ports; The step of predicting a health evaluation result corresponding to any of the master ports using a pre-built health evaluation model based on the multiple aggregated variable data associated with any of the master ports includes: For any primary port: Filtering and obtaining a plurality of target variable data based on the plurality of aggregate variable data associated with the primary port; Performing a joint operation on the multiple aggregated variable data associated with the primary port to obtain multiple derived variable data; Binning the target variable data and the derived variable data, and determining the target weight value corresponding to each bin; Based on each of the target weight values, using the health assessment model to predict a health assessment result corresponding to the primary port; The step of constructing a port relationship diagram corresponding to each main port and each sub-port based on the SMS port data includes: Determine the sub-port information corresponding to the SMS port data; Extracting the main port information based on the sub-port information; Based on the association relationship between the sub-port information and the main port information, a port relationship graph corresponding to each main port and each sub-port is constructed.

2. The port health evaluation method according to claim 1, characterized in that: The health assessment model is trained based on the following steps: Obtaining sample data of port status and multiple indicator variables associated with multiple primary ports; For any indicator variable: binning the indicator variable to obtain multiple bin intervals; Based on the sample data corresponding to the indicator variable, determine the weight value corresponding to each bin interval; The health assessment model is trained based on the weight value corresponding to each binning interval.

3. The port health evaluation method according to claim 2, characterized in that: The port status includes normal status, notification status, current limiting status and shutdown status; The step of determining a weight value corresponding to each binning interval based on the sample data corresponding to the indicator variable includes: Based on the sample data corresponding to the indicator variable, determine the number of samples in each bin interval in which the port status is in the normal state, the notification state, the current limiting state, and the shutdown state; Determine the number of abnormal samples based on the number of samples in the notification state, current limiting state, and shutdown state and the associated weights; Based on the number of samples in a normal state and the number of abnormal samples in each binning interval, a weight value corresponding to each binning interval is determined.

4. The port health evaluation method according to claim 1, characterized in that: The SMS port data is aggregated based on the port relationship diagram and the preset time window to obtain multiple aggregate variable data associated with any main port, including: Dividing the SMS port data within the preset time period according to the time window to obtain target divided data within multiple time windows; In combination with the port relationship diagram, aggregation processing is performed on the target partition data in each time window to obtain multiple aggregate variable data associated with any main port.

5. The port health evaluation method according to claim 1, characterized in that: The obtaining of multi-dimensional SMS port data within a preset time period includes: Collect SMS complaint data, daily sub-port sending volume data, sub-port real-name registration data, case-related sub-port verification data, and risk signature data within a preset time period. The case-related sub-port verification data refers to the verification data of the sub-port corresponding to the risk event involved in the case; The SMS complaint data, the sub-port daily sending volume data, the sub-port real-name registration data, the sub-port verification data involved in the case and the risk signature data are analyzed to form the multi-dimensional SMS port data.

6. A port health assessment device, characterized in that: include: The acquisition module is used to obtain multi-dimensional SMS port data within a preset time period; A construction module, configured to construct a port relationship diagram corresponding to each main port and each sub-port based on the SMS port data; an aggregation module, configured to aggregate the SMS port data based on the port relationship graph and a preset time window to obtain a plurality of aggregated variable data associated with any primary port; An evaluation module, configured to predict a health evaluation result corresponding to any of the master ports using a pre-built health evaluation model based on a plurality of aggregated variable data associated with any of the master ports; The evaluation module is further configured to: For any primary port: Filtering and obtaining a plurality of target variable data based on the plurality of aggregate variable data associated with the primary port; Performing a joint operation on the multiple aggregated variable data associated with the primary port to obtain multiple derived variable data; Binning the target variable data and the derived variable data, and determining the target weight value corresponding to each bin; Based on each of the target weight values, using the health assessment model to predict a health assessment result corresponding to the primary port; The building block is further configured to: Determine the sub-port information corresponding to the SMS port data; Extracting the main port information based on the sub-port information; Based on the association relationship between the sub-port information and the main port information, a port relationship graph corresponding to each main port and each sub-port is constructed.

7. A computer device comprising a memory, a processor, and computer-readable instructions stored in the memory and executed on the processor, wherein: When the processor executes the computer-readable instructions, the port health evaluation method according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium having computer-readable instructions stored thereon, characterized in that: When the computer-readable instructions are executed by a processor, the port health evaluation method according to any one of claims 1 to 5 is implemented.

Images