Intelligent short message management and control method, device and equipment and storage medium

Abstract

The invention relates to the technical field of artificial intelligence, can be applied to the field of intelligent medical treatment and finance, and discloses an intelligent short message management and control method, device and equipment and a storage medium, and the method comprises the steps: obtaining complaint data, predicting the complaint probability of a user, and constructing a complaint sensitive user portrait; screening the short message content, and checking the risk level of the short message content; and setting a short message sending grade according to the complaint sensitive user portrait. The method can reduce invalid touch, avoid resource waste, save operation cost, actively avoid sensitive users and reduce complaint rate.

Description

Technical Field

[0001] The present invention relates to the fields of artificial intelligence technology and natural language processing technology, and can be applied to the fields of smart medical care and finance. In particular, it relates to an intelligent text message management method, device, equipment and storage medium. Background Art

[0002] In the smart healthcare and financial sectors, SMS messages are often used to improve business efficiency, such as reminders for follow-up appointments, physical examinations, consultation appointment details, promotions for medical products, medical insurance, financial products, identity information expiration and renewal reminders, and financial product renewal reminders. However, excessive SMS messages can easily lead to user complaints. Existing systems fail to screen users who are sensitive to SMS content, resulting in a large number of invalid SMS messages, wasted resources, increased operating costs, and high complaint rates. Summary of the Invention

[0003] The present invention provides an intelligent SMS management method, device, equipment and storage medium to solve the technical problems that the inability to screen sensitive users leads to invalid SMS sending, waste of resources, increased operating costs and high complaint rate.

[0004] In a first aspect, a smart SMS management method is provided, comprising:

[0005] Obtain complaint data, predict user complaint probability, and build complaint-sensitive user profiles;

[0006] Screen SMS content and review its risk level;

[0007] Set SMS sending levels based on complaint sensitivity user profiles.

[0008] In a second aspect, an intelligent SMS management and control device is provided, comprising:

[0009] Complaint-sensitive user portrait construction module, which obtains complaint data, predicts the probability of user complaints, and constructs complaint-sensitive user portraits;

[0010] SMS review module, which screens SMS content and reviews the risk level of SMS content;

[0011] The SMS sending classification module sets the SMS sending classification based on the complaint sensitivity user profile.

[0012] In a third aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned intelligent SMS management and control method when executing the computer program.

[0013] In a fourth aspect, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the steps of the above-mentioned intelligent SMS management and control method are implemented.

[0014] In the solution implemented by the above-mentioned intelligent SMS management method, device, equipment, and storage medium, complaint data can be obtained through the client, and the server can predict the user complaint probability and construct a complaint-sensitive user profile; screen SMS content and review the risk level of SMS content; and set SMS sending levels based on the complaint-sensitive user profile. In the present invention, for intelligent assistants for promoting medical products or medical insurance in the field of smart healthcare, or for intelligent assistants for promoting financial products in the financial business, the intelligent SMS management solution can be utilized to obtain complaint data, predict the user complaint probability, and construct a complaint-sensitive user profile; screen SMS content and review the risk level of SMS content; and set SMS sending levels based on the complaint-sensitive user profile. This can effectively formulate SMS strategies based on user sensitivity, screen and stop invalid SMS messages for sensitive users, intercept invalid SMS messages, reduce invalid contacts, save communication costs, and reduce the workload of operators in handling repeated complaints. This helps free up human resources to focus on complex cases, avoid resource waste, actively avoid highly sensitive groups, reduce complaint rates at the source, and balance compliance bottom lines with development demands. Moreover, the regulatory standards can be adjusted according to actual conditions, which is highly practical. By screening SMS content, illegal language can be intercepted in real time, reducing complaints caused by inappropriate expression. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] 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.

[0016] Figure 1 This is a schematic diagram of an application environment of the intelligent SMS management method in one embodiment of the present invention;

[0017] Figure 2 This is a flow chart of an intelligent SMS management method according to an embodiment of the present invention;

[0018] Figure 3 yes Figure 1 A schematic flow chart of a specific implementation of step S10;

[0019] Figure 4 This is a structural diagram of an intelligent SMS management and control device in one embodiment of the present invention;

[0020] Figure 5is a structural diagram of a computer device in one embodiment of the present invention;

[0021] Figure 6 FIG. 2 is another structural diagram of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] 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.

[0023] The intelligent SMS management method provided by the embodiment of the present invention can be applied in Figure 1 In the application environment, it is used in smart assistants or smart customer service in application scenarios such as medical care, finance and insurance, and is usually implemented through the server, wherein the client communicates with the server through the network. The server can receive SMS requests through the client, and the SMS request includes an SMS draft. The client transmits the SMS request to the server, and the server obtains complaint data, predicts the probability of user complaints, and builds a complaint-sensitive user portrait; screens SMS content, and reviews the risk level of SMS content; sets SMS sending levels based on complaint-sensitive user portraits, and feeds back SMS review results and SMS sending level results to the client. In the present invention, for smart assistants for promoting medical products or medical insurance in the field of smart medical care, or for smart assistants for promoting financial products in financial services, a smart SMS management solution can be used to obtain complaint data, predict the probability of user complaints, and build a complaint-sensitive user portrait. User portraits of sensitive users; screening of SMS content, and review of the risk level of SMS content; setting SMS sending levels based on complaint sensitivity user portraits, which can effectively formulate SMS strategies based on user sensitivity, screen and stop sending invalid SMS messages to sensitive users, intercept invalid SMS sending, reduce invalid reach, save communication costs, and reduce the workload of operators in handling repeated complaints, which is conducive to freeing up human resources to focus on complex cases, avoid waste of resources, actively avoid highly sensitive groups, reduce complaint rates from the source, and take into account compliance bottom lines and development demands. Moreover, regulatory standards can be adjusted according to actual conditions and are highly practical. By screening SMS content, illegal language can be intercepted in real time, reducing complaints caused by inappropriate expressions. Among them, the client can be, but is not limited to, various personal computers, laptops, smart phones, tablets, and portable wearable devices. The server can be implemented with an independent server or a server cluster consisting of multiple servers. The present invention is described in detail below through specific embodiments.

[0024] See also Figure 2 As shown, Figure 2A flow chart of the intelligent SMS management method provided in an embodiment of the present invention includes the following steps:

[0025] S10: Obtain complaint data, predict user complaint probability, and build a complaint-sensitive user profile.

[0026] The intelligent SMS management and control method provided by the present invention can be applied to intelligent customer service or intelligent assistants in various application scenarios such as medical care, finance, and insurance. It is usually implemented through a server, which can receive SMS requests from sales staff in real time. The SMS request includes a draft SMS. According to the content of the draft SMS, the business type of the SMS can be divided into service SMS and marketing SMS. For example, in the field of medical applications, for an intelligent assistant promoting medical products or medical insurance, a service SMS may include the use details and time limit notification of the medical product, or the term and insurance details of the medical insurance; a business SMS may include an introduction to a new medical product, an introduction to a new medical insurance, and an invitation to renew the medical insurance; or, for example, in the field of finance, for an intelligent assistant promoting financial products, a service SMS may include a notification of the income of the financial product and a time limit notification; a business SMS may include an introduction to a new financial product and an invitation to renew the financial product. After receiving the SMS request from the salesperson, the intelligent assistant obtains complaint data, predicts the probability of user complaints, builds a complaint-sensitive user profile, accurately identifies high-risk users, reduces the number of marketing SMS messages sent to them, reduces complaints at the source, lowers the complaint rate, takes into account both compliance bottom lines and development demands, screens and stops sending invalid SMS messages to sensitive users, intercepts the sending of invalid SMS messages, reduces invalid contacts, saves communication costs, reduces the workload of operators in handling repeated complaints, frees up human resources to focus on complex cases, avoids waste of resources, and regulatory standards can be adjusted according to actual conditions, supports rapid adaptation of rules, facilitates regulatory upgrades, and builds a sustainable management system.

[0027] Among them, Figure 3 As shown, step S10, i.e. obtaining complaint data, predicting user complaint probability, and building a complaint-sensitive user profile, includes the following steps:

[0028] S11: Obtain basic user information from the CRM (Customer Relationship Management) system and construct user sensitivity and activity features. User basic information includes personal identity information, complaint data, number of SMS messages received, number of SMS replies, account status, and overdue records. Personal identity information includes name, age, gender, and region. Complaint data includes complaint history, complaint frequency, and complaint type. Complaint history indicates whether a complaint has been made within a preset timeframe, which can be three months. Complaint types include operator complaints, administrative department complaints, and escalated complaints, which can be categorized based on complaint content and / or complaint intent. Sensitivity is used to determine a user's propensity to complain about SMS content, i.e., to determine the probability of a complaint. User sensitivity features include the number of historical complaints and complaint type. Activity is used to determine a user's propensity to reply to SMS messages after receiving them. User activity features include active days and interaction frequency. Active days refers to the number of days SMS messages were sent to the user within the set timeframe, and interaction frequency refers to the frequency of SMS replies received within the set timeframe, which can be one month.

[0029] Specifically, step S11, i.e., constructing user sensitivity features and user activity features, includes:

[0030] Based on the complaint data of the obtained user basic information, the NLP (Natural Language Processing) model is used to analyze the complaint content of the complaint data to obtain the historical number of complaints and complaint types, and the user sensitivity features are constructed based on the obtained historical number of complaints and complaint types; among them, the historical number of complaints can be used to calculate the average number of complaints per month, and the complaint type can be used to calculate the complexity of the complaint type. The complaint weight of upgraded complaints is the highest, and the weight ratio of upgraded complaints, operator complaints and management department complaints can be 5:2.5:2.5; when the NLP model is used to analyze the complaint content of the complaint data, the complaint keywords in the complaint content can be captured, and by calculating the coverage rate of the complaint keywords in the total number of SMS complaints on that day, combined with the preset high-frequency complaint word coverage rate threshold, it can be determined whether the corresponding complaint keyword is a high-frequency complaint item. The preset high-frequency complaint word coverage rate threshold can be 20%. If the coverage rate of the complaint keywords in the total number of SMS complaints on that day is greater than 20%, the corresponding complaint keywords on that day are high-frequency complaint items;

[0031] Based on the number of SMS messages received and the number of SMS replies obtained from the user's basic information, the number of active days and the frequency of interaction are obtained to construct user activity characteristics.

[0032] S12: Predicting the probability of user complaints and calculating the user activity score. The user complaint probability can be predicted by the Logistic Regression algorithm or the XGBoost algorithm based on the constructed user sensitivity features and user activity features. Then step S12, i.e., predicting the user complaint probability, is specifically as follows:

[0033] Define the binary complaint label as the target variable, use the logistic regression algorithm or gradient boosting tree algorithm to predict the user complaint probability and calculate the user activity score based on the constructed user sensitivity features and user activity features.

[0034] The binary complaint labels include "complaint" and "no complaint." A complaint label is recorded as 1, and a no complaint label is recorded as 0. The predicted probability of a user complaint ranges from 0 to 1. The logistic regression algorithm is suitable for scenarios where the data is linearly separable and the relationship between features and the target variable is relatively simple, facilitating analysis of the impact of each feature on complaints. The gradient boosting tree algorithm is suitable for scenarios with complex data, significant nonlinear relationships, and high prediction accuracy requirements. It facilitates processing high-dimensional data and has strong nonlinear relationship capabilities. Depending on the needs, either the logistic regression algorithm or the gradient boosting tree algorithm can be used to predict the probability of user complaints.

[0035] Preferably, in step S12, i.e. calculating the user activity score, specifically:

[0036] Normalization is used to convert the user activity index into a user activity score. The user activity index may be the number of received text messages, and the user activity score ranges from 0 to 1.

[0037] S13: Based on the average monthly platform complaints and the distribution of basic user information, the user sensitivity threshold and user activity threshold are set as the quadrant boundary between sensitivity and activity. Combined with the predicted user complaint probability and user activity score, a complaint-sensitive user profile is constructed.

[0038] When the user complaint probability is below the user sensitivity threshold, the user is considered low-sensitive, meaning the complaint probability is low; otherwise, the user is considered highly sensitive, meaning the complaint probability is high. When the user activity score is below the user activity threshold, the user is considered low-active; otherwise, the user is considered highly active. The complaint sensitivity user profile includes user sensitivity and activity levels, as well as the SMS policies corresponding to each level. For example, the user sensitivity threshold could be 0.8, meaning a user with a complaint probability below 0.8 is considered low-sensitive, while a user with a complaint probability at or above 0.8 is considered highly sensitive. The user activity threshold could be 0.7, meaning a user with an activity score below 0.7 is considered low-active, while a user with an activity score at or above 0.7 is considered highly active. The user sensitivity threshold can be determined based on the distribution of complaint data within user basic information. By taking the quantile of the complaint probability, the top 20% are considered highly sensitive. For months with low complaint volume, the threshold can be set based on business experience. For example, users who file at least two complaints in a quarter are considered highly sensitive. The user activity threshold can be determined based on the distribution of user activity metrics, such as the median number of SMS messages received.

[0039] When the user's sensitivity and activity are classified as highly sensitive and highly active, that is, the probability of complaints is high and the activity is high, the SMS strategy is to completely avoid and avoid contact, that is, suspend SMS marketing and SMS push, reduce the frequency of contact, and optimize the SMS process to reduce the triggering of complaints; when the user's sensitivity and activity are classified as highly sensitive and low active, that is, the probability of complaints is high but the activity is low, the SMS strategy is to reach out with caution and optimize the SMS content, that is, reduce the frequency of SMS marketing and SMS push to avoid frequent interruptions, and collect problems through user surveys or feedback channels for targeted optimization; when the user's sensitivity and activity are classified as low sensitive and highly active, that is, the activity is high but the probability of complaints is low, the SMS strategy is normal push, and users in this category are key maintenance users. Data analysis can be used to optimize the user experience and further improve satisfaction; when the user's sensitivity and activity are classified as low sensitive and low active, that is, both the probability of complaints and the activity are low, the SMS strategy is normal push and continuous observation, recording behavioral changes.

[0040] Specifically, after step S13, that is, according to the average monthly platform complaint volume and the distribution of basic user information, the user sensitivity threshold and the user activity threshold are set as the quadrant boundary between sensitivity and activity, and after the predicted user complaint probability and user activity score are combined to construct the complaint sensitivity user profile, the following is also included:

[0041] When the proportion of highly sensitive users exceeds the preset high-sensitivity ratio threshold, the user sensitivity threshold is adjusted; wherein the preset high-sensitivity ratio threshold may be 10%;

[0042] When the proportion of low-activity users exceeds the preset user ratio threshold, adjust the user activity threshold based on business needs.

[0043] The complaint-sensitive user profile can be rebuilt every quarter based on actual needs to update the quadrant division between sensitivity and activity based on the latest data, and adjust SMS strategies based on changes in user behavior to support rapid adaptation of rules, facilitate regulatory upgrades, and build a sustainable management system.

[0044] Specifically, after step S10, that is, after building the complaint-sensitive user profile, the following steps may also be included:

[0045] Analyze the actual complaint rate of users and verify the accuracy of the complaint sensitivity user portrait.

[0046] S20: Screen the content of SMS messages and review the risk level of the content.

[0047] For SMS requests from sales staff, screening the SMS content before sending can help proactively reduce complaints. In step S20, the SMS content is screened and the risk level of the SMS content is reviewed, specifically including:

[0048] Collect historical complaint text messages to generate a sensitive word list, and use the BERT model to determine the contextual tone intensity of the historical complaint text messages. Sensitive words can include deduction, overdue, breach of contract, and freezing.

[0049] The system obtains text message drafts and extracts sensitive words, tone intensity, and formatting standards from their content. It then uses pre-set rules and the BERT model to determine the draft's risk level. The draft's content is then colored according to the risk level, generating a risk level heat map. This visually highlights problematic areas in the draft and facilitates revisions by business personnel. Pre-set rules include legal provisions and historical complaint standards. Business personnel can upload and submit draft text messages based on SMS templates. Draft text messages can be ranked high, medium, or low risk. The risk level heat map indicates high, medium, and low risk content in the draft text messages in red, yellow, and green, respectively. The NLP model analyzes the text message semantics, extracts sensitive words from a sensitive word list, and uses the BERT model for contextual analysis and tone intensity analysis. This allows for content analysis and review of the draft text, intercepting illegal language in real time and reducing complaints caused by inappropriate language.

[0050] Specifically, the BERT model is used to determine the contextual tone strength of historical complaint text messages, including:

[0051] Collect historical complaint text messages, annotate them by tone intensity, mark the locations of sensitive words, and construct a training set. Segment the text messages, add special tags for the BERT model, convert the specially marked historical complaint text messages into the BERT model input format, and then input them into the pre-trained BERT model to fine-tune the pre-trained BERT model and train the classification layer to identify tone intensity features.

[0052] The BERT model is used to extract contextual embeddings of historical complaint text messages, capturing the relevance of sensitive words and surrounding semantics.

[0053] The output vector of the CLS marker or the hidden state of a specific layer is used in combination with the classification results to determine the tone intensity level. The tone intensity level can be divided into three levels: high, medium, and low. The CLS marker is one of the special markers of the BERT model.

[0054] Specifically, after using the BERT model to determine the contextual tone strength of the historical complaint text message, the method further includes:

[0055] Based on the output of the BERT model and the sensitive word list, the weight of sensitive words can be adjusted in real time or an early warning mechanism can be triggered to dynamically update the sensitive word list, which can improve the accuracy of sensitive word recognition.

[0056] S30: Set SMS sending levels based on complaint sensitivity user profiles.

[0057] SMS sending is categorized into three levels: high-risk, medium-risk, and low-risk. High-risk corresponds to highly sensitive and highly active users; medium-risk corresponds to highly sensitive and inactive users; and low-risk corresponds to both low-sensitivity and highly active users, and low-sensitivity and inactive users. Users in the high-risk category are only allowed to send service-related SMS messages; users in the medium-risk category can send marketing SMS messages, but frequency limits apply; and users in the low-risk category can send marketing SMS messages, but are still subject to other policy restrictions.

[0058] Specifically, step S30, i.e., after setting the SMS sending level according to the complaint sensitivity user profile, further includes:

[0059] Monitor the SMS messages of different BUs (Business Units) to avoid multiple SMS bombardment and reduce the increase in complaints caused by excessive outreach.

[0060] The monitoring of SMS messages of different BUs is as follows:

[0061] Get the marketing SMS sending status of different BUs to the same user within the preset time period. If a BU has sent a marketing SMS to the user within the preset time period, cancel the sending of marketing SMS to the user by the remaining BUs within the preset time period.

[0062] Specifically, monitoring the SMS status of different BUs also includes:

[0063] Obtain the monthly number of marketing SMS messages received by a user and determine whether the number exceeds the threshold. If the number exceeds the threshold, alert the business personnel when a new SMS request is received.

[0064] It can be seen that in the above scheme, the intelligent assistant for promoting medical products or medical insurance in the field of smart medical care, or the intelligent assistant for promoting financial products in the financial business, can use the intelligent SMS management solution to obtain complaint data, predict the probability of user complaints, and build a complaint-sensitive user portrait; screen SMS content and review the risk level of SMS content; set SMS sending levels according to the complaint-sensitive user portrait, which can effectively formulate SMS strategies based on user sensitivity, screen and stop sending invalid SMS to sensitive users, intercept invalid SMS sending, reduce invalid reach, save communication costs, and reduce the workload of operators in handling repeated complaints, which is conducive to freeing up human resources to focus on complex cases, avoiding waste of resources, actively avoiding highly sensitive people, reducing the complaint rate from the source, and taking into account both compliance bottom line and development demands. Moreover, the regulatory standards can be adjusted according to actual conditions and are highly practical. By screening SMS content, illegal language can be intercepted in real time, reducing complaints caused by inappropriate expression.

[0065] 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.

[0066] In one embodiment, a smart SMS management device is provided, which corresponds to the smart SMS management method in the above embodiment. Figure 4 As shown, the intelligent SMS control device includes a complaint-sensitive user construction module 101, an SMS review module 102, and an SMS sending classification module 103. The functional modules are described in detail as follows:

[0067] Complaint-sensitive user profile building module 101, which obtains complaint data, predicts user complaint probability, and builds complaint-sensitive user profiles;

[0068] SMS review module 102, screening SMS content and reviewing the risk level of SMS content;

[0069] The SMS sending classification module 103 sets the SMS sending classification according to the complaint sensitivity user profile.

[0070] In one embodiment, the complaint sensitivity user profile building module 101 is specifically used to:

[0071] Obtain basic user information from the CRM system and construct user sensitivity and activity features. Basic user information includes personal information, complaint data, number of SMS messages received, and number of SMS replies.

[0072] Predict user complaint probability and calculate user activity score;

[0073] Based on the average monthly platform complaints and the distribution of user basic information, the user sensitivity threshold and user activity threshold are set as the quadrant dividing line between sensitivity and activity. Combined with the predicted user complaint probability and user activity score, a complaint-sensitive user profile is constructed.

[0074] In one embodiment, the complaint sensitivity user profile building module 101 is specifically used to:

[0075] Based on the complaint data obtained from user basic information, the NLP model is used to analyze the complaint content of the complaint data to obtain the historical number of complaints and complaint types, and user sensitivity features are constructed based on the obtained historical number of complaints and complaint types;

[0076] Based on the number of SMS messages received and the number of SMS replies obtained from the user's basic information, the number of active days and the frequency of interaction are obtained to construct user activity characteristics.

[0077] In one embodiment, the complaint sensitivity user profile building module 101 is specifically used to:

[0078] Define the binary complaint label as the target variable, and use the logistic regression algorithm or gradient boosting tree algorithm to predict the user complaint probability based on the constructed user sensitivity features and user activity features.

[0079] In one embodiment, the SMS review module 102 is specifically configured to:

[0080] Collect historical complaint text messages to generate a sensitive word list, and use the BERT model to determine the contextual tone intensity of historical complaint text messages;

[0081] Obtain SMS drafts, extract sensitive words, tone intensity, and format specifications based on the content of the obtained SMS drafts, combine preset rules and the BERT model to determine the risk level of the SMS drafts, color the content of the SMS drafts according to the risk level, and generate a risk level heat map of the SMS.

[0082] In one embodiment, the SMS review module 102 is specifically configured to:

[0083] Collect historical complaint text messages, annotate the tone intensity, mark the location of sensitive words, and build a training set;

[0084] Segment the historical complaint text messages, add special tags for the BERT model, convert the specially marked historical complaint text messages into the BERT model input format, and then input them into the pre-trained BERT model to fine-tune the pre-trained BERT model and train the classification layer to identify tone intensity features.

[0085] The BERT model is used to extract contextual embeddings of historical complaint text messages, capturing the relevance of sensitive words and surrounding semantics.

[0086] The output vector of the CLS tag or the hidden state of a specific layer is used in combination with the classification results to determine the tone intensity level.

[0087] In one embodiment, the SMS sending classification module 103 is further configured to:

[0088] Monitor the SMS status of different BUs.

[0089] The present invention provides an intelligent SMS management and control device, which obtains complaint data, predicts the probability of user complaints, and constructs a complaint-sensitive user portrait; screens SMS content, and reviews the risk level of SMS content; sets SMS sending levels according to the complaint-sensitive user portrait, and can effectively formulate SMS strategies based on user sensitivity, screen and stop sending invalid SMS to sensitive users, intercept the sending of invalid SMS, reduce invalid contacts, save communication costs, and reduce the workload of operators in handling repeated complaints, which is conducive to freeing up human resources to focus on complex cases, avoiding waste of resources, actively avoiding highly sensitive people, reducing complaint rates from the source, and taking into account compliance bottom lines and development demands. Moreover, regulatory standards can be adjusted according to actual conditions, and are highly practical. By screening SMS content, illegal language can be intercepted in real time, reducing complaints caused by inappropriate expressions.

[0090] The specific definition of the intelligent SMS control device can be found in the definition of the intelligent SMS control method above and will not be repeated here. Each module in the above-mentioned intelligent SMS control device can be implemented in whole or in part through software, hardware, or a combination thereof. Each of 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 each of the above modules.

[0091] 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 5As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, it realizes the functions or steps on the service side of an intelligent SMS management method.

[0092] In one embodiment, a computer device is provided. The computer device may be a client, and its internal structure diagram may be as follows: Figure 6 As shown. The computer device includes a processor, memory, network interface, display screen and input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it realizes the functions or steps of the client side of an intelligent SMS management method.

[0093] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are performed:

[0094] Obtain complaint data, predict user complaint probability, and build complaint-sensitive user profiles;

[0095] Screen SMS content and review its risk level;

[0096] Set SMS sending levels based on complaint sensitivity user profiles.

[0097] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0098] Obtain complaint data, predict user complaint probability, and build complaint-sensitive user profiles;

[0099] Screen SMS content and review its risk level;

[0100] Set SMS sending levels based on complaint sensitivity user profiles.

[0101] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can be found in the relevant descriptions of the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0102] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it 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 can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0103] 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.

[0104] The embodiments described above 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. An intelligent SMS management and control method, characterized in that: include: Obtain complaint data, predict user complaint probability, and build complaint-sensitive user profiles; Screen SMS content and review its risk level; Set SMS sending levels based on complaint sensitivity user profiles.

2. The intelligent SMS management and control method according to claim 1, wherein: The acquisition of complaint data, prediction of user complaint probability, and construction of complaint-sensitive user profiles include: Obtain basic user information from the CRM system and construct user sensitivity and activity features. Basic user information includes personal information, complaint data, number of SMS messages received, and number of SMS replies. Predict user complaint probability and calculate user activity score; Based on the average monthly platform complaints and the distribution of user basic information, the user sensitivity threshold and user activity threshold are set as the quadrant dividing line between sensitivity and activity. Combined with the predicted user complaint probability and user activity score, a complaint-sensitive user profile is constructed.

3. The intelligent SMS management and control method according to claim 2, characterized in that: The constructing of user sensitivity features and user activity features includes: Based on the complaint data obtained from user basic information, the NLP model is used to analyze the complaint content of the complaint data to obtain the historical number of complaints and complaint types, and user sensitivity features are constructed based on the obtained historical number of complaints and complaint types; Based on the number of SMS messages received and the number of SMS replies obtained from the user's basic information, the number of active days and the frequency of interaction are obtained to construct user activity characteristics.

4. The intelligent SMS management and control method according to claim 2, wherein: The predicted user complaint probability is specifically: Define the binary complaint label as the target variable, and use the logistic regression algorithm or gradient boosting tree algorithm to predict the user complaint probability based on the constructed user sensitivity features and user activity features.

5. The intelligent SMS management and control method according to claim 1, wherein: Screening SMS content and reviewing the risk level of SMS content specifically include: Collect historical complaint text messages to generate a sensitive word list, and use the BERT model to determine the contextual tone intensity of historical complaint text messages; Obtain SMS drafts, extract sensitive words, tone intensity, and format specifications based on the content of the obtained SMS drafts, combine preset rules and the BERT model to determine the risk level of the SMS drafts, color the content of the SMS drafts according to the risk level, and generate a risk level heat map of the SMS.

6. The intelligent SMS management and control method according to claim 5, characterized in that: The BERT model is used to determine the contextual tone strength of historical complaint text messages, including: Collect historical complaint text messages, annotate the tone intensity, mark the location of sensitive words, and build a training set; Segment the historical complaint text messages, add special tags for the BERT model, convert the specially marked historical complaint text messages into the BERT model input format, and then input them into the pre-trained BERT model to fine-tune the pre-trained BERT model and train the classification layer to identify tone intensity features. The BERT model is used to extract contextual embeddings of historical complaint text messages, capturing the relevance of sensitive words and surrounding semantics. The output vector of the CLS tag or the hidden state of a specific layer is used in combination with the classification results to determine the tone intensity level.

7. The intelligent SMS management and control method according to claim 1, wherein: After setting the SMS sending level according to the complaint sensitivity user profile, the method further includes: Monitor the SMS status of different BUs.

8. An intelligent SMS management and control device, characterized in that: include: Complaint-sensitive user portrait construction module, which obtains complaint data, predicts the probability of user complaints, and constructs complaint-sensitive user portraits; SMS review module, which screens SMS content and reviews the risk level of SMS content; The SMS sending classification module sets the SMS sending classification based on the complaint sensitivity user profile.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the intelligent SMS management and control method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the intelligent SMS management and control method according to any one of claims 1 to 7 are implemented.

Images