User intelligent voice calling system and method based on big data

Abstract

The embodiments of the present application relate to the field of computer technology, and disclose a user intelligent voice call system and method based on big data. The user intelligent voice call method based on big data first determines a target user group with a first attribute feature based on the user's historical consumption transaction records, then determines a second attribute feature of the user's concern for credit card upgrades based on the user's historical voice interaction records, and finally adjusts the voice call strategy based on the first attribute feature and the second attribute feature to perform intelligent voice calls. In this way, multiple attribute features are comprehensively considered to perform intelligent voice precision calls, thereby improving the accuracy of user demand positioning and the efficiency of voice call transactions.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a user intelligent voice calling system and method based on big data. Background Art

[0002] With the continuous advancement of financial technology, credit card business has become an important part of the development of banking business. In order to further improve the revenue and customer satisfaction of credit card business, many banks have launched credit card upgrade promotion activities, making voice calls to cardholders through intelligent outbound calling systems to achieve business expansion and upgrade. However, in actual application, this intelligent voice calling method has obvious shortcomings. It lacks accurate user classification and demand positioning capabilities, resulting in many users who are not interested in credit card upgrades or have sufficient credit limits being included in the call range, causing adverse effects such as telephone harassment to users and reducing user experience. At the same time, due to inaccurate user screening, the large amount of human and material resources invested by banks in promotion activities are often not effectively utilized, resulting in low transaction efficiency, and may even cause users to dislike bank services, which has a negative impact on the bank's brand image and customer relationships. Summary of the invention

[0003] The main purpose of the present invention is to provide a user intelligent voice call system and method based on big data, aiming to solve the technical problem in the prior art that the lack of accurate user classification and demand positioning capabilities leads to low voice call effectiveness.

[0004] To achieve the above objectives, in a first aspect, an embodiment of the present application provides a user intelligent voice call method based on big data, which is applied to an intelligent voice call system. The intelligent voice call system is used for credit card upgrade service calls. The method includes:

[0005] Obtaining the cardholder's personal information and the historical transaction information of the credit card bound to the cardholder's personal information;

[0006] Constructing a user portrait tag based on the cardholder's personal information and credit card transaction history information;

[0007] Inputting the user portrait label into a pre-trained portrait label clustering model to obtain a target user group, wherein the target user group carries a first attribute feature;

[0008] Acquire the interaction content with the target user group users in the historical interaction process, and perform language processing analysis on the interaction content to obtain the credit card upgrade intention, wherein the credit card upgrade intention includes the second attribute feature that the user is concerned about the credit card upgrade;

[0009] Adjusting the voice call strategy according to the first attribute feature and the second attribute feature to obtain a target voice call strategy;

[0010] Intelligent voice calls are made to users in the target user group according to the target voice call strategy.

[0011] In a possible implementation, the historical transaction information of the credit card bound to the cardholder's personal information includes historical transaction information under the first external environment, and constructing a user portrait tag based on the cardholder's personal information and the credit card's historical transaction information includes:

[0012] Acquire historical transaction information of a credit card bound to the personal information of the cardholder in a first external environment;

[0013] A user portrait label is constructed based on the cardholder's personal information and historical transaction information in the first external environment.

[0014] In a possible implementation, inputting the user portrait label into a pre-trained portrait label clustering model to obtain a target user group includes:

[0015] The user portrait labels of users in the first external environment are input into a pre-trained portrait label clustering model to obtain a target user group, wherein the first attribute feature is associated with the first external environment.

[0016] In a possible implementation, the historical transaction information of the credit card bound to the cardholder's personal information includes historical transaction information under the second external environment, and constructing a user portrait tag based on the cardholder's personal information and the credit card's historical transaction information includes:

[0017] Acquire historical transaction information of the credit card bound to the personal information of the cardholder in a second external environment;

[0018] Build a user portrait label based on the cardholder’s personal information and historical transaction information in the second external environment.

[0019] In a possible implementation, inputting the user portrait label into a pre-trained portrait label clustering model to obtain a target user group includes:

[0020] The user portrait labels of the users in the second external environment are input into a pre-trained portrait label clustering model to obtain a target user group, wherein the first attribute feature is associated with the second external environment.

[0021] In a possible implementation, obtaining the interaction content with the target user group users in the historical interaction process and performing language processing and analysis on the interaction content to obtain the credit card upgrade intention includes:

[0022] Acquire historical voice interaction content with users of the target user group during historical interactions;

[0023] Extracting target sentences from historical voice interaction content of conversations with users of the target user group using natural language processing technology to obtain target interaction content, wherein the target sentences are sentences related to credit card upgrades;

[0024] Semantic analysis is performed on the target interactive content to obtain credit card upgrade intention.

[0025] In a possible implementation, obtaining the interaction content with the target user group users in the historical interaction process and performing language processing and analysis on the interaction content to obtain the credit card upgrade intention includes:

[0026] Acquire historical voice interaction content during historical interactions with the target user group;

[0027] The credit card upgrade intention sentiment information carried in the historical voice interaction content is identified based on sentiment analysis technology, wherein the credit card upgrade intention sentiment information includes the urgency of the credit card upgrade.

[0028] In a possible implementation, after identifying the credit card upgrade intention sentiment information carried by the historical voice interaction content according to the sentiment analysis technology, the method further includes:

[0029] Prioritize the order of conversations between users in the target user group and the intelligent voice call system based on the urgency of the user's credit card upgrade;

[0030] An intelligent voice call is made to the user according to the priority sorting and using the target voice call strategy.

[0031] In a possible implementation, the identifying the credit card upgrade intention sentiment information carried by the historical voice interaction content according to the sentiment analysis technology includes:

[0032] Obtaining a positive emotion score, a negative emotion score, a neutral emotion score, and a keyword matching degree in the historical voice interaction content;

[0033] The positive sentiment score, negative sentiment score, neutral sentiment score and keyword matching degree are input into a pre-trained sentiment analysis model to obtain the credit card upgrade urgency, wherein the sentiment analysis model satisfies the following expression:

[0034] , where E is the urgency of credit card upgrade, A is the positive sentiment score, B is the negative sentiment score, C is the neutral sentiment score, and D is the keyword matching degree. 1. 2. 3. 4 are the contributions to the calculation of urgency obtained through historical big data training, E1r is the negative correction of the urgency of credit card upgrade, and the negative correction is positively correlated with the number of historical voice call complaints, E2r is the positive correction of the urgency of credit card upgrade, and the positive correction is positively correlated with the number of historical credit card upgrades.

[0035] In the second aspect, an embodiment of the present application also provides a user intelligent voice call system based on big data, including a memory and a processor, wherein the memory is used to store program code, and the processor is used to call the program code to execute the method described in the first aspect.

[0036] Different from the prior art, the user intelligent voice call method based on big data provided in the embodiment of the present application first obtains the personal information of the cardholder and the historical transaction information of the credit card bound to the personal information of the cardholder; then constructs a user portrait label based on the personal information of the cardholder and the historical transaction information of the credit card; then inputs the user portrait label into a pre-trained portrait label clustering model to obtain a target user group, and the target user group carries a first attribute feature; then obtains the interaction content with the users of the target user group in the historical interaction process, and performs language processing and analysis on the interaction content to obtain the credit card upgrade intention, and the credit card upgrade intention includes the second attribute feature of the user's concern about the credit card upgrade; then adjusts the voice call strategy based on the first attribute feature and the second attribute feature to obtain the target voice call strategy; finally, according to the target voice call strategy, an intelligent voice call is made to the users in the target user group. That is, first, the target user group with the first attribute characteristic is determined based on the user's historical consumption transaction records, and then the second attribute characteristic of the user's concern about credit card upgrades is determined based on the target user's historical voice interaction records. Finally, the voice call strategy is adjusted based on the first attribute characteristic and the second attribute characteristic to perform intelligent voice calls. In this way, multiple attribute characteristics are comprehensively considered to perform intelligent voice precision calls, which improves the accuracy of user demand positioning and thus improves the efficiency of voice call transactions. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 the structures shown in these drawings without paying creative work.

[0038] Figure 1 This is a flowchart of a user intelligent voice calling method based on big data in some embodiments of the present application;

[0039] Figure 2 A flowchart of a user intelligent voice calling method based on big data in some other embodiments of the present application;

[0040] Figure 3 This is a schematic diagram of the hardware structure of a user intelligent voice call system based on big data in some embodiments of the present application.

[0041] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0044] In addition, the descriptions of "first", "second", etc. in the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, "and / or" in the full text includes three solutions. Taking A and / or B as an example, it includes technical solution A, technical solution B, and technical solution that satisfies both A and B. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of ordinary technicians in this field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0045] With the continuous advancement of financial technology, credit card business has become an important part of the development of banking business. In order to further improve the revenue and customer satisfaction of credit card business, many banks have launched credit card upgrade promotion activities, making voice calls to cardholders through intelligent outbound calling systems to achieve business expansion and upgrade. However, in actual application, this intelligent voice calling method has obvious shortcomings. It lacks accurate user classification and demand positioning capabilities, resulting in many users who are not interested in credit card upgrades or have sufficient credit limits being included in the call range, causing adverse effects such as telephone harassment to users and reducing user experience. At the same time, due to inaccurate screening, the large amount of human and material resources invested by banks in promotion activities are often not effectively utilized, resulting in low transaction efficiency, and may even cause users to dislike bank services, which has a negative impact on the bank's brand image and customer relationships.

[0046] In response to the above problems, the present application proposes a user intelligent voice call method based on big data, which is applied to an intelligent voice call system, and the intelligent voice call system is used for credit card upgrade service calls, such as making an intelligent voice call to a target user to promote a credit card upgrade service to the target user. The overall idea of ​​the user intelligent voice call method based on big data in the present application is: first, determine the target user group with a first attribute feature according to the user's historical consumption transaction record (for example, the first attribute feature is "frequent large-amount credit consumption at night"), then determine the second attribute feature of the user's attention to credit card upgrade according to the target user's historical voice interaction record (for example, the second attribute feature is "increasing the credit limit"), and finally adjust the voice call strategy according to the first attribute feature and the second attribute feature to make an intelligent voice call (make a voice call at night, and the theme of the voice call is to increase the credit limit for the user). In this way, a plurality of attribute features are comprehensively considered to make an intelligent voice precise call, which improves the accuracy of user demand positioning and the efficiency of voice call transactions (making a voice call to promote the upgrade service to the user at night when the user needs a large amount of consumption, which can improve the transaction efficiency).

[0047] like Figure 1-2 As shown, the following will mainly describe the specific steps of the user intelligent voice call method based on big data. It should be noted that although the logical order is shown in the flowchart, in some cases, the steps shown or described can be performed in a different order from here. Figure 1 , the method comprises the following steps:

[0048] S100, obtaining the cardholder's personal information and the historical transaction information of the credit card bound to the cardholder's personal information;

[0049] The cardholder's personal information and the historical transaction information of the credit card bound to the cardholder's personal information can be obtained from the bank's backend management system. For example, the cardholder's personal information can be obtained from the user information provided when applying for a credit card, and the credit card's historical transaction information can be obtained through the bank's settlement system.

[0050] S200, constructing a user portrait tag based on the cardholder's personal information and credit card transaction history information;

[0051] After obtaining the cardholder's personal information and credit card consumption transaction record information, user portrait tags can be constructed based on the above information in order to classify users into groups.

[0052] It is understandable that the results of user portrait labels under different portrait label generation rules will be quite different. For example, user portrait labels can be constructed based on gender and consumption amount, or based on credit card consumption frequency. The user portrait labels obtained under different standards may also have different usage purposes.

[0053] In one embodiment, the historical transaction information of the credit card bound to the cardholder's personal information includes the historical transaction information under the first external environment, and the step S200: constructing a user portrait tag according to the cardholder's personal information and the credit card historical transaction information includes:

[0054] S210, obtaining historical transaction information of the credit card bound to the cardholder's personal information in the first external environment;

[0055] S220. Construct a user portrait tag based on the cardholder's personal information and historical transaction information in the first external environment.

[0056] Specifically, the first external environment may be ambient light brightness, such as daytime or nighttime. The following description takes nighttime as the first external environment as an example.

[0057] For example, first obtain the cardholder's historical transaction information at night, and then build a user portrait label based on the cardholder's personal information and the historical transaction information at night. The user portrait label may include a user portrait who does not make any credit card purchases at night, a user portrait who occasionally makes credit card purchases at night, and a user portrait who frequently makes credit card purchases at night.

[0058] In other examples, the user portrait tags may also include a user portrait who does not make any credit card purchases at night, a user portrait who occasionally makes low-value credit card purchases at night, and a user portrait who frequently makes high-value credit card purchases at night.

[0059] In another embodiment, the historical transaction information of the credit card bound to the cardholder's personal information includes the historical transaction information under the second external environment, and the step S200: constructing a user portrait tag according to the cardholder's personal information and the credit card historical transaction information includes:

[0060] S230, obtaining historical transaction information of the credit card bound to the cardholder's personal information in the second external environment;

[0061] S240. Construct a user portrait label based on the cardholder's personal information and historical transaction information in the second external environment.

[0062] Specifically, the second external environment may be a working time environment, such as a working day environment and a non-working day environment. The following description takes a non-working day environment as an example of the second external environment.

[0063] For example, first obtain the historical transaction information of the cardholder on non-working days, and then build a user portrait label based on the cardholder's personal information and the historical transaction information on non-working days. The user portrait label may include user portraits who do not have any credit card consumption on non-working days, user portraits who occasionally make credit card consumption on non-working days, and user portraits who frequently make credit card consumption on non-working days.

[0064] In other examples, the user portrait tags may also include user portraits of users who do not make any credit card purchases on non-working days, user portraits of users who occasionally make low-value credit card purchases on non-working days, and user portraits of users who frequently make high-value credit card purchases on non-working days.

[0065] S300, inputting the user portrait label into a pre-trained portrait label clustering model to obtain a target user group, wherein the target user group carries a first attribute feature;

[0066] Since cardholders have different credit card spending habits and spending records, there are many types of user portrait tags for cardholders. It is necessary to group the user portrait tags of cardholders to obtain the target user group.

[0067] In one embodiment, the step S300: inputting the user portrait label into a pre-trained portrait label clustering model to obtain a target user group includes:

[0068] S310. Inputting the user portrait label of the user in the first external environment into a pre-trained portrait label clustering model to obtain a target user group, wherein the first attribute feature is associated with the first external environment.

[0069] Exemplarily, when the first external environment is a night environment, the user portrait labels of all cardholders in the first external environment can be input into a pre-trained portrait label clustering model to obtain a target user group. For example, a user group that makes high-frequency and high-value consumption at night can be screened by the portrait label clustering model. This user group can be used as the target user group for subsequent voice call management.

[0070] It can be understood that when the first external environment is a night environment, the first attribute feature can be an attribute feature related to the night environment - "high frequency and high credit consumption at night".

[0071] In other embodiments, the step S300: inputting the user portrait label into a pre-trained portrait label clustering model to obtain a target user group includes:

[0072] S320. Input the user portrait label of the user in the second external environment into a pre-trained portrait label clustering model to obtain a target user group, wherein the first attribute feature is associated with the second external environment.

[0073] Exemplarily, when the second external environment is a non-working day environment, the user portrait labels of all cardholders in the second external environment can be input into a pre-trained portrait label clustering model to obtain a target user group. For example, a user group that makes high-frequency and high-value consumption on non-working days can be screened by the portrait label clustering model. This user group can be used as the target user group for subsequent voice call management.

[0074] It can be understood that when the second external environment is a non-working day environment, the first attribute feature can be an attribute feature related to the non-working day environment - "high frequency and high amount of credit consumption on non-working days".

[0075] S400, obtaining the interaction content with the target user group users in the historical interaction process, and performing language processing analysis on the interaction content to obtain the credit card upgrade intention, wherein the credit card upgrade intention includes a second attribute feature that the user is concerned about the credit card upgrade;

[0076] In order to increase the target users' interest in and receptivity to intelligent voice calls, when making voice calls, the topics that users are more concerned about can be obtained based on historical interactions. For example, if the target users are interested in increasing their credit card limit, annual fee reduction, or additional credit card services (VIP services), the target users' interest direction is the user's intention to upgrade their credit card. In other words, the second attribute feature can be one or more of "credit limit increase", "annual fee reduction", or "additional services".

[0077] In one embodiment, the step S400: obtaining the interaction content with the target user group users in the historical interaction process, and performing language processing and analysis on the interaction content to obtain the credit card upgrade intention, includes:

[0078] S410, obtaining historical voice interaction content in the historical interaction process with users of the target user group;

[0079] S420, extracting target sentences from historical voice interaction contents of conversations with users of the target user group according to natural language processing technology to obtain target interaction contents, wherein the target sentences are sentences related to credit card upgrades;

[0080] S430: Perform semantic analysis on the target interactive content to obtain credit card upgrade intention.

[0081] Specifically, the historical voice interaction content with the target user group users in the historical interaction process is first obtained, and then the target sentence is extracted from the historical voice interaction content with the target user group users according to the natural language processing technology to obtain the target interaction content, and finally the target interaction content is semantically analyzed to obtain the credit card upgrade intention. In this way, after extracting the sentences related to the credit card upgrade, the semantic analysis of the relevant sentences is performed to obtain the credit card upgrade intention, which can improve the accuracy and efficiency of the credit card upgrade intention analysis.

[0082] In other embodiments, the step S400: obtaining the interaction content with the target user group users in the historical interaction process, and performing language processing and analysis on the interaction content to obtain the credit card upgrade intention includes:

[0083] S440, obtaining historical voice interaction content in the historical interaction process with the target user group;

[0084] S450. Identify the credit card upgrade intention sentiment information carried in the historical voice interaction content according to sentiment analysis technology, wherein the credit card upgrade intention sentiment information includes the urgency of credit card upgrade.

[0085] Specifically, the urgency of the user's credit card upgrade is analyzed according to the historical voice interaction content. After the user's urgency for the credit card upgrade is obtained, the order of conversations between users in the target user group and the intelligent voice call system can be prioritized according to the urgency of the user's credit card upgrade; finally, an intelligent voice call is made to the user according to the priority sorting and the target voice call strategy; in this way, the urgency of the user's credit card upgrade is judged through the historical voice interaction records. When the user's credit card upgrade urgency is large, it means that the user is in urgent need of the credit card upgrade service, and the user's voice call can be put in front to give priority to the voice call.

[0086] It should be noted that the urgency of credit card upgrade can be obtained by using sentiment analysis technology during voice interaction with the user. In one embodiment, step S450: identifying the sentiment information of credit card upgrade intention carried by the historical voice interaction content according to the sentiment analysis technology includes:

[0087] S451, obtaining a positive emotion score, a negative emotion score, a neutral emotion score, and a keyword matching degree in the historical voice interaction content;

[0088] S452: Input the positive sentiment score, negative sentiment score, neutral sentiment score and keyword matching degree into a pre-trained sentiment analysis model to obtain the urgency of credit card upgrade, wherein the sentiment analysis model satisfies the following expression:

[0089] , where E is the urgency of credit card upgrade, A is the positive sentiment score, B is the negative sentiment score, C is the neutral sentiment score, and D is the keyword matching degree. 1. 2. 3. 4 are the contributions to the calculation of urgency obtained through historical big data training, E1r is the negative correction of the urgency of credit card upgrade, and the negative correction is positively correlated with the number of historical voice call complaints, E2r is the positive correction of the urgency of credit card upgrade, and the positive correction is positively correlated with the number of historical credit card upgrades.

[0090] Specifically, the emotional information and keyword information in the historical voice interaction content are converted into specific numerical values, so as to quantitatively evaluate the user's urgency for credit card upgrade. By comprehensively considering multiple aspects such as the user's emotions, historical behaviors, usage habits, and complaints, the user's urgency for credit card upgrade can be evaluated more comprehensively.

[0091] S500, adjusting the voice call strategy according to the first attribute feature and the second attribute feature to obtain a target voice call strategy;

[0092] In one embodiment, after the first attribute feature and the second attribute feature are obtained, the voice call strategy may be adjusted according to the first attribute feature and the second attribute feature to obtain a target voice call strategy.

[0093] In other embodiments, after obtaining the target voice call strategy according to the first attribute feature and the second attribute feature, intelligent voice calls can be made to the target user group in sequence in combination with the voice call sequence priority to give priority to meeting the needs of urgent users.

[0094] S600: Performing intelligent voice calls on users in a target user group according to the target voice call strategy.

[0095] After the target voice call strategy is obtained, intelligent voice calls can be made to users in the target user group according to the target voice call strategy.

[0096] For example, when the first attribute feature is "high frequency and high amount of credit consumption at night" and the second attribute feature is "credit limit increase", the original call strategy can be adjusted to make intelligent voice calls at night to users who have high frequency and high amount of credit consumption at night and are interested in credit limit increase, and provide credit upgrade explanation services for credit limit increase. In this way, voice calls are made to promote the upgrade services corresponding to the user's interests when the user is in need of large-amount consumption, which can meet the user's immediate needs and greatly improve transaction efficiency.

[0097] For example, when the first attribute feature is "high frequency credit consumption on weekdays" and the second attribute feature is "annual fee reduction", the original call strategy can be adjusted to make intelligent voice calls on weekdays to users who have high frequency credit consumption on weekdays and are interested in annual fee reduction, and provide credit upgrade explanation services for annual fee reduction (such as upgrading to a platinum credit card). In this way, voice calls are made to promote corresponding upgrade services to users when users are in need of high frequency consumption, which can meet users' immediate needs and greatly improve transaction efficiency.

[0098] Based on this, the user intelligent voice call method based on big data of the present application first determines the target user group with the first attribute feature according to the user's historical consumption transaction records, and then determines the second attribute feature of the user's concern about credit card upgrades according to the target user's historical voice interaction records. Finally, the voice call strategy is adjusted according to the first attribute feature and the second attribute feature to make an intelligent voice call. In this way, multiple attribute features are comprehensively considered to make intelligent voice precise calls, thereby improving the accuracy of user demand positioning and the efficiency of voice call transactions.

[0099] The present application also provides a user intelligent voice call system 100 based on big data. Figure 3 , Figure 3 A schematic diagram of the hardware structure of a big data-based user intelligent voice call system provided for some embodiments of the present application; comprising a memory 110 and a processor 120, wherein the memory 110 is used to store program code, and the processor 120 is used to call the program code to execute the method described above.

[0100] Among them, the processor 120 is used to provide computing and control capabilities to control the user intelligent voice call system based on big data to perform corresponding tasks, for example, to control the user intelligent voice call system based on big data to perform the user intelligent voice call method based on big data in any of the above method embodiments, the method comprising: obtaining the personal information of the cardholder and the historical transaction information of the credit card bound to the cardholder's personal information; constructing a user portrait label based on the cardholder's personal information and the credit card historical transaction information; inputting the user portrait label into a pre-trained portrait label clustering model to obtain a target user group, and the target user group carries a first attribute feature; obtaining the interaction content with the users of the target user group in the historical interaction process, and performing language processing and analysis on the interaction content to obtain a credit card upgrade intention, and the credit card upgrade intention includes a second attribute feature of the user's concern about the credit card upgrade; adjusting the voice call strategy based on the first attribute feature and the second attribute feature to obtain a target voice call strategy; and performing intelligent voice calls on users in the target user group according to the target voice call strategy.

[0101] The processor 120 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), a hardware chip, or any combination thereof; it may also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The above-mentioned PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

[0102] The memory 110, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer executable programs and modules, such as program instructions / modules corresponding to the user intelligent voice call method based on big data in the embodiment of the present application. The processor 120 can implement the user intelligent voice call method based on big data in any of the above method embodiments by running the non-transitory software programs, instructions and modules stored in the memory 110.

[0103] Specifically, the memory 110 may include a volatile memory (VM), such as a random access memory (RAM); the memory 110 may also include a non-volatile memory (NVM), such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD) or a solid-state drive (SSD) or other non-transitory solid-state storage device; the memory 110 may also include a combination of the above types of memory.

[0104] To sum up, the user intelligent voice call system based on big data of the present application adopts the technical solution of any one of the above-mentioned user intelligent voice call method embodiments based on big data. Therefore, it has at least the beneficial effects brought by the technical solutions of the above-mentioned embodiments, which will not be repeated here one by one.

[0105] The embodiment of the present application also provides a computer-readable storage medium, such as a memory including a program code, and the program code can be executed by a processor to complete the user intelligent voice call method based on big data in the above embodiment. For example, the computer-readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CDROM), a magnetic tape, a floppy disk, and an optical data storage device.

[0106] The present application also provides a computer program product, which includes one or more program codes, which are stored in a computer-readable storage medium. The processor of the user intelligent voice call system based on big data reads the program code from the computer-readable storage medium, and the processor executes the program code to complete the steps of the user intelligent voice call method based on big data provided in the above embodiment.

Claims

1. A user intelligent voice calling method based on big data is applied to an intelligent voice calling system, wherein the intelligent voice calling system is used for credit card upgrade service calls, characterized in that: The method comprises: Obtaining the cardholder's personal information and the historical transaction information of the credit card bound to the cardholder's personal information; Constructing a user portrait tag based on the cardholder's personal information and credit card transaction history information; Inputting the user portrait label into a pre-trained portrait label clustering model to obtain a target user group, wherein the target user group carries a first attribute feature; Acquire the interaction content with the target user group users in the historical interaction process, and perform language processing analysis on the interaction content to obtain the credit card upgrade intention, wherein the credit card upgrade intention includes the second attribute feature that the user is concerned about the credit card upgrade; Adjusting the voice call strategy according to the first attribute feature and the second attribute feature to obtain a target voice call strategy; Performing intelligent voice calls on users in the target user group according to the target voice call strategy; The acquiring of the interaction content with the target user group users in the historical interaction process, and performing language processing and analysis on the interaction content to obtain the credit card upgrade intention includes: Acquire historical voice interaction content during historical interactions with the target user group; Identifying the credit card upgrade intention sentiment information carried by the historical voice interaction content according to sentiment analysis technology, wherein the credit card upgrade intention sentiment information includes the urgency of credit card upgrade; After identifying the credit card upgrade intention sentiment information carried by the historical voice interaction content according to the sentiment analysis technology, the method further includes: Prioritize the order of conversations between users in the target user group and the intelligent voice call system based on the urgency of the user's credit card upgrade; An intelligent voice call is made to the user according to the priority sorting and using the target voice call strategy.

2. The user intelligent voice calling method based on big data as claimed in claim 1, characterized in that: The historical transaction information of the credit card bound to the cardholder's personal information includes the historical transaction information in the first external environment. The user portrait tag is constructed based on the cardholder's personal information and the credit card historical transaction information, including: Acquire historical transaction information of a credit card bound to the personal information of the cardholder in a first external environment; A user portrait label is constructed based on the cardholder's personal information and historical transaction information in the first external environment.

3. The user intelligent voice calling method based on big data as claimed in claim 2, characterized in that: The step of inputting the user portrait labels into a pre-trained portrait label clustering model to obtain a target user group includes: The user portrait labels of users in the first external environment are input into a pre-trained portrait label clustering model to obtain a target user group, wherein the first attribute feature is associated with the first external environment.

4. The user intelligent voice calling method based on big data as claimed in claim 1, characterized in that: The historical transaction information of the credit card bound to the cardholder's personal information includes the historical transaction information in the second external environment. The user portrait tag is constructed based on the cardholder's personal information and the credit card historical transaction information, including: Acquire historical transaction information of the credit card bound to the personal information of the cardholder in a second external environment; Build a user portrait label based on the cardholder’s personal information and historical transaction information in the second external environment.

5. The user intelligent voice calling method based on big data as claimed in claim 4, characterized in that: The step of inputting the user portrait labels into a pre-trained portrait label clustering model to obtain a target user group includes: The user portrait labels of the users in the second external environment are input into a pre-trained portrait label clustering model to obtain a target user group, wherein the first attribute feature is associated with the second external environment.

6. The user intelligent voice calling method based on big data as claimed in claim 1, characterized in that: The acquiring of the interaction content with the target user group users in the historical interaction process, and performing language processing and analysis on the interaction content to obtain the credit card upgrade intention includes: Acquire historical voice interaction content with users of the target user group during historical interactions; Extracting target sentences from historical voice interaction content of conversations with users of the target user group using natural language processing technology to obtain target interaction content, wherein the target sentences are sentences related to credit card upgrades; Semantic analysis is performed on the target interactive content to obtain credit card upgrade intention.

7. The user intelligent voice calling method based on big data as claimed in claim 1, characterized in that: The step of identifying the credit card upgrade intention sentiment information carried by the historical voice interaction content according to the sentiment analysis technology includes: Obtaining a positive emotion score, a negative emotion score, a neutral emotion score, and a keyword matching degree in the historical voice interaction content; The positive sentiment score, negative sentiment score, neutral sentiment score and keyword matching degree are input into a pre-trained sentiment analysis model to obtain the credit card upgrade urgency, wherein the sentiment analysis model satisfies the following expression: E=A*Ф1+B*Ф2+C*Ф3+D*Ф4-E1r+E2r, where E is the urgency of credit card upgrade, A is the positive sentiment score, B is the negative sentiment score, C is the neutral sentiment score, D is the keyword matching degree, Ф1, Ф2, Ф3, and Ф4 are the contributions to the calculation of urgency obtained based on historical big data training, E1r is the negative correction of the urgency of credit card upgrade, and the negative correction is positively correlated with the number of historical voice call complaints, and E2r is the positive correction of the urgency of credit card upgrade, and the positive correction is positively correlated with the number of historical credit card upgrades.

8. The user intelligent voice call system based on big data is characterized by: The method comprises a memory and a processor, wherein the memory is used to store program codes, and the processor is used to call the program codes to execute the method according to any one of claims 1 to 7.

Images