Virtual product distribution method and device, equipment and medium

Abstract

The invention belongs to the field of artificial intelligence, and relates to a virtual product distribution method, which comprises the steps of obtaining historical data of a first target customer, training a preset first prediction model based on a predefined service perception label by using the historical data to obtain a perception model, and training a preset second prediction model based on an activeness label to obtain a plurality of time period contact opportunity models. Second target customer data is acquired, service perception information is predicted by using the perception model, and a target service perception customer is determined according to the service perception information. And predicting the time period by using the contact opportunity model of each time period, and determining a recommended time period. And predicting service preference information by using a preset preference model. And finally, distributing a corresponding virtual product to the target service feeling customer according to the service preference information in the recommendation time period. The invention further provides a device, equipment and a medium. In addition, the invention also relates to a block chain technology, and historical data and customer data can be stored in a block chain. The method and the device can be applied to the field of financial insurance business, and personalized and accurate pushing of the service can be realized.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method, apparatus, device, and medium for distributing virtual products. Background Art

[0002] With the rapid development of internet technology and the widespread application of big data, the insurance industry is undergoing unprecedented transformation. Driven by the wave of digitalization, insurance companies face increasingly fierce market competition. To stand out in this competition, improving customer satisfaction and loyalty has become a key issue for insurance companies. Traditional insurance service models focus primarily on optimizing products, while neglecting the important role of pre-sales service in attracting potential customers and improving the customer experience.

[0003] However, in practice, insurance companies have found that the investment and output of pre-sales services are often not proportional. Specifically, insurance companies often adopt a broad-based approach in pre-sales services, lacking in-depth analysis and precise targeting of individual customer differences. This extensive service model not only wastes a significant amount of service resources but also fails to meet the diverse service needs of customers. Summary of the Invention

[0004] The purpose of the embodiments of the present application is to propose a virtual product distribution method, device, equipment and medium to solve the problem that the existing method of using a wide-net approach lacks in-depth analysis and precise positioning of individual customer differences, resulting in a large amount of waste of service resources.

[0005] In the first aspect, a virtual product distribution method is provided, which adopts the following technical solutions:

[0006] Obtain historical data of the first target customer; based on predefined service perception tags, use historical data to perform perception training on the preset first prediction model to obtain a perception model; based on predefined activity tags, use historical data to perform timing training on the preset second prediction model to obtain contact timing models for multiple time periods; obtain customer data of the second target customer, use the perception model to perform perception prediction processing on the customer data to obtain service perception information of each second target customer; based on the service perception information, determine the target service-conscious customer from the second target customer, use the contact timing model for each time period to perform time period prediction processing on the customer data of the target service-conscious customer, and determine the recommended time period for each target service-conscious customer; use the preset preference model to perform preference prediction processing on the customer data of the target service-conscious customer to obtain the service preference information of the target service-conscious customer; within the recommended contact time period, distribute corresponding virtual products to the target service-conscious customer based on the service preference information.

[0007] In a second aspect, a virtual product distribution device is provided, which adopts the following technical solution:

[0008] An acquisition module, used to acquire historical data of the first target customer;

[0009] A perception training module is used to perform perception training on a preset first prediction model using historical data based on a predefined service perception tag to obtain a perception model;

[0010] A timing training module is used to perform timing training on a preset second prediction model using historical data based on predefined activity tags to obtain contact timing models for multiple time periods;

[0011] A perception prediction module is used to obtain customer data of the second target customer, perform perception prediction processing on the customer data using a perception model, and obtain service perception information of each second target customer;

[0012] A time period prediction module is used to determine target service-conscious customers from the second target customers based on the service perception information, use a contact timing model for each time period to perform time period prediction processing on the customer data of the target service-conscious customers, and determine a recommended time period for each target service-conscious customer;

[0013] A preference prediction module is used to perform preference prediction processing on the customer data of the target service-conscious customer using a preset preference model to obtain the service preference information of the target service-conscious customer;

[0014] The distribution module is used to distribute corresponding virtual products to target service-conscious customers based on service preference information within the recommended contact period.

[0015] In a third aspect, a computer device is provided, including a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the virtual product distribution method as described above are implemented.

[0016] In a fourth aspect, a computer-readable storage medium is provided, which stores computer-readable instructions. The computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the virtual product distribution method as described above.

[0017] In the solution implemented by the aforementioned virtual product distribution method, apparatus, device, and medium, historical data of a first target customer is obtained and a pre-set first prediction model is trained based on pre-defined service perception tags. The resulting perception model can deeply explore the customer's service needs and perceived preferences, providing precise guidance for subsequent service. Next, a second prediction model is trained using pre-defined activity tags to construct contact timing models for multiple time periods. This helps identify the optimal time for customers to receive service, avoid ineffective communication, and improve the utilization of service resources. Furthermore, the perception model performs perceptual prediction processing on the customer data of the second target customer, accurately identifying target service-sensitive customers. Combined with the contact timing model, the recommended time period is determined, achieving personalized and precise service delivery. Finally, a pre-set preference model is used to predict the preferences of the target service-sensitive customers, and corresponding virtual products are distributed accordingly. This not only meets the diverse needs of customers but also enhances their identification and loyalty to the insurance company, significantly improving pre-sales service effectiveness and customer satisfaction while controlling costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the solutions in this application, a brief introduction will be given below to the drawings required for use in the description of the embodiments of this application. Obviously, the drawings described below are some embodiments of this application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0019] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0020] Figure 2 This is a flowchart of a virtual product distribution method provided by this application;

[0021] Figure 3 This is a structural diagram of a virtual product distribution device provided by this application;

[0022] Figure 4 This is a structural diagram of a computer device provided by this application. DETAILED DESCRIPTION

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meanings as commonly understood by those skilled in the art to which this application belongs. The terms used in the specification of the application are for the purpose of describing specific embodiments only and are not intended to limit this application. The terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of this application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0024] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0025] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0026] like Figure 1 As shown, system architecture 100 may include a terminal device 101, a network 102, and a server 103. Terminal device 101 may be a laptop computer 1011, a tablet computer 1012, or a mobile phone 1013. Network 102 is a medium for providing a communication link between terminal device 101 and server 103. Network 102 may include various connection types, such as wired or wireless communication links or fiber optic cables.

[0027] The user can use the terminal device 101 to interact with the server 103 via the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0028] The terminal device 101 can be various electronic devices with a display screen and supporting web browsing. In addition to the laptop computer 1011, tablet computer 1012 or mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV) player, a laptop computer and a desktop computer, etc.

[0029] The server 103 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal device 101 .

[0030] It should be noted that the virtual product distribution method provided in the embodiment of the present application is generally executed by a server, and accordingly, the virtual product distribution device is generally set in the server.

[0031] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is merely illustrative. Any number of terminal devices, networks and servers may be provided as required.

[0032] Continue to refer Figure 2 , shows a flow chart of an embodiment of a virtual product distribution method according to the present application. The virtual product distribution method includes the following steps:

[0033] Step S201 obtains historical data of the first target customer.

[0034] The primary target customer refers to a specific customer group within the insurance business that is selected for data collection and analysis. Their historical data is used to train prediction models to optimize pre-sales service strategies.

[0035] Historical data refers to various data records generated by the first target customer during their interactions with the insurance company over a period of time. This data includes, but is not limited to, the customer's gender, vehicle age, vehicle model, vehicle price, identification of residential and commercial vehicles, whether they are WeChat users, age, customer level, whether they drive a new energy vehicle, whether they have children, income level, occupation, marital status, education level, and the issuance and use of various services within the auto insurance app.

[0036] Step S202: Based on the predefined service perception tag, historical data is used to perform perception training on the preset first prediction model to obtain a perception model.

[0037] Service perception labels are a set of predefined identifiers used to describe a customer's perception of insurance services. These labels are constructed based on customer feedback and behavioral data, and are used to quantify customer perceptions of service satisfaction and service-demand fit. For example, labels such as "high satisfaction" and "low demand fit" can be used to characterize a customer's specific feelings about insurance services and serve as key input for training perception models.

[0038] The first prediction model is a pre-set algorithm model used for perception training based on customer historical data and service perception labels. This model uses machine learning technology to learn the relationship between customer characteristics and service perception to predict the service perception of new customers.

[0039] The perception model is trained using the first prediction model and is used to predict customer service perception information for new services. This model is trained based on service perception labels in historical data and can accurately identify customer perception information such as service needs and satisfaction.

[0040] Step S203 : Based on the predefined activity tags, historical data is used to perform timing training on the preset second prediction model to obtain contact timing models for multiple time periods.

[0041] Activity tags are a set of predefined identifiers used to describe the level of customer engagement with an insurance company. These tags are constructed based on data such as customer login frequency, service usage frequency, and number of interactions, and are used to quantify customer activity levels. For example, tags such as "high activity" and "low activity" may be used to characterize the frequency and depth of customer interactions with an insurance company.

[0042] The second prediction model is a pre-set algorithm model used to train timing based on customer historical data and activity tags. This model uses machine learning to learn the relationship between customer activity and optimal contact timing, predicting the optimal service contact time for different customers.

[0043] Multiple time periods refer to dividing a day or period into several continuous or discontinuous time periods, which are used to analyze customer activity and service acceptance within these time periods. These time periods can be based on business needs, customer behavior patterns, or market trends, such as morning, noon, afternoon, and evening.

[0044] The contact timing model, trained using the second prediction model, is used to predict the optimal service contact time for different customers. This model, trained based on activity labels from historical data, accurately identifies customers' willingness to accept service and their activity levels over different time periods. It is used to predict the optimal service contact time for customers.

[0045] Step S204: Acquire customer data of the second target customer, perform perception prediction processing on the customer data using the perception model, and obtain service perception information of each second target customer.

[0046] Among them, the second target customer refers to the specific customer group selected as the service push object in the insurance business.

[0047] The service perception information is obtained by predicting and processing the customer data of the second target customer through the perception model, and is specific information about the customer's perception of insurance services. This information includes but is not limited to customer satisfaction with the service and the degree of demand matching.

[0048] Step S205: Based on the service perception information, target service-conscious customers are determined from the second target customers. The contact timing model for each time period is used to perform time period prediction processing on the customer data of the target service-conscious customers to determine the recommended time period for each target service-conscious customer.

[0049] Targeted service-sensitive customers refer to a group of customers selected from the secondary target customers, who have a high level of service awareness and willingness to accept. These customers are typically identified based on service perception information and are the primary target for insurance companies to deliver targeted services.

[0050] Recommended time periods are the optimal time periods for delivering services to target service-conscious customers, as predicted by the contact timing model. These time periods are typically determined based on factors such as customer activity and willingness to accept services, aiming to improve the accuracy and effectiveness of service delivery.

[0051] Step S206 , using a preset preference model, performs preference prediction processing on the customer data of the target service-conscious customer to obtain the service preference information of the target service-conscious customer.

[0052] The preference model is a pre-defined algorithmic model used to predict customer service preferences based on customer data. Using machine learning techniques, the model learns the relationship between customer characteristics and service preferences to predict the service preferences of new customers.

[0053] Among them, customer data refers to various information records related to insurance customers, including but not limited to basic customer information (such as gender, age, occupation, etc.), vehicle information (such as vehicle age, model, vehicle price, etc.), service usage (such as the issuance and usage of various services in the auto insurance APP, etc.) and behavioral data (such as login frequency, service usage frequency, etc.).

[0054] Service preference information, derived from predictive processing of target service-conscious customers' customer data using a preference model, provides information on specific customer preferences for insurance services. This information includes, but is not limited to, preferences for service type, service channel, and service time. Service preference information is a crucial basis for insurance companies to make personalized service recommendations and improve customer satisfaction, and it serves to characterize the specific needs and expectations of target service-conscious customers for insurance services.

[0055] Step S207: within the recommended contact period, corresponding virtual products are distributed to target service-conscious customers based on the service preference information.

[0056] Virtual products refer to digitally accessible services or products within the insurance business. These products can include electronic policies, online claims processing, various value-added services within auto insurance apps (such as roadside assistance and vehicle inspections), and service packages. Virtual products are targeted at increasing customer acceptance and satisfaction with insurance services through targeted push notifications based on service preferences during recommended contact times.

[0057] The embodiment of the present application can obtain the historical data of the first target customer and perform perception training on the preset first prediction model based on the predefined service perception label. The obtained perception model can deeply explore the customer's service needs and perception preferences, and provide accurate guidance for subsequent services. Then, the second prediction model is trained on the timing using the predefined activity label to construct a contact timing model for multiple time periods, which helps to grasp the best time for customers to receive services, avoid ineffective communication, and improve the utilization rate of service resources. Furthermore, by performing perception prediction processing on the customer data of the second target customer through the perception model, the target service-conscious customer can be accurately identified, and the recommended time period can be determined in combination with the contact timing model to achieve personalized and accurate service push. Finally, the preset preference model is used to predict the preferences of the target service-conscious customer, and the corresponding virtual products are distributed accordingly, which not only meets the diverse needs of customers, but also enhances the customer's sense of identity and loyalty to the insurance company, thereby significantly improving the effect of pre-sales service and customer satisfaction while controlling costs.

[0058] In some optional implementations of this embodiment, step S202, based on the predefined service perception tag, uses historical data to perform perception training on the preset first prediction model to obtain a perception model, specifically including the following steps:

[0059] Obtain insurance sales cycle information of the first target customer, and based on the insurance sales cycle, divide historical data into training data and prediction data; based on predefined perception feature information, perform feature extraction on the training data to obtain perception feature information; based on predefined service perception tags, perform perception annotation on the perception feature information to obtain perception feature information carrying perception annotations; based on the perception feature information carrying perception annotations and an optimization algorithm, perform binary classification training on the preset first prediction model to obtain a perception model; use the prediction data to perform performance verification on the perception model to obtain a verified perception model.

[0060] Insurance sales cycle information refers to the time span and related data records from when a customer is introduced to an insurance product to when they purchase or renew it. This information typically comes from the insurance company's sales and customer management systems, and is obtained by collecting and analyzing data such as the customer's purchase time, renewal time, and policy validity period.

[0061] Perception feature information refers to characteristic data that reflects a customer's perception of insurance products, services, or marketing activities. It represents a customer's attention, interest, satisfaction, and potential purchase intention towards insurance products.

[0062] Perceptual feature information is the result of extracting features from training data using predefined perceptual feature information. It is the key features that reflect customer perception extracted after processing and analyzing the original data.

[0063] Optimization algorithms are a class of algorithms used to find optimal or near-optimal solutions, such as tuning a grid. In machine learning and model training, optimization algorithms are used to adjust model parameters to minimize the loss function or maximize the model's performance.

[0064] In one example, a large insurance company will illustrate how to build and validate a perception model based on the insurance sales cycle information of the first target customer, combined with perceptual feature information, to optimize customer service strategies. First, historical data on the first target customer can be obtained, including basic customer information, insurance purchase history, and service usage. Information on the insurance sales cycle of these customers, such as the time interval between insurance purchases and renewal periods, is also collected. Based on the insurance sales cycle, the historical data is divided into training data and prediction data. For example, if the current month is July, the historical data for three months (i.e., the expiration dates in April, May, and June) is used as the training set, and the historical data for the expiration date in July is used as the prediction set. Feature extraction is performed on the training data based on predefined perceptual feature information, such as customer interest in insurance products, service usage frequency, claims satisfaction, customer age, occupation, and income level. Data mining techniques are used to filter features closely related to customer service perception from the massive data to form perceptual feature information. These features comprehensively reflect customers' insurance needs and service preferences. Perceptual feature information is perceptually labeled using predefined service perception labels, such as "high satisfaction" and "low demand match." This step aims to convert customers' perceived preferences into quantifiable labels to provide supervisory signals for subsequent model training. For example, for customers whose historical write-off rates are greater than the average write-off rate, the label is "1", indicating high satisfaction; the labels of other customers are "0", indicating low demand matching. Based on the perceptual feature information carrying perceptual annotations and optimization algorithms (such as tuned grid, random forest, gradient boosting tree or deep neural network, etc.), the preset first prediction model is trained for binary classification. By continuously adjusting the model parameters, the model can accurately predict the customer's service perception. After training is completed, the perceptual model is obtained. The performance of the perceptual model can be verified using predicted data. By comparing the model prediction results with actual customer feedback, the model's accuracy, recall rate, F1 score and other indicators are evaluated. If the model performance does not meet expectations, return to adjust the feature extraction, perceptual annotation or model training steps until the model performance meets the requirements. The verified perceptual model will be used for subsequent customer service optimization.

[0065] The embodiment of the present application can achieve effective organization and utilization of data by obtaining the insurance sales cycle information of the first target customer and dividing the historical data into training data and prediction data based on the cycle. This step ensures that the model training can closely fit the actual cycle of the insurance business and improves the prediction accuracy of the model. Furthermore, based on the predefined perceptual feature information, feature extraction is performed on the training data to obtain perceptual feature information. This process deeply explores the customer's potential demand for insurance services and provides strong support for subsequent precise services. By perceptually labeling the perceptual feature information and performing binary classification training on the preset first prediction model based on the feature information carrying the perceptual labeling and the optimization algorithm, a perceptual model is finally obtained. This model can accurately predict the customer's perception of insurance services and provide a scientific decision-making basis for insurance companies. The performance of the perceptual model is verified using prediction data to ensure the stability and reliability of the model. The comprehensive application of this series of technical means not only improves the accuracy of insurance services and customer satisfaction, but also effectively reduces service costs.

[0066] In some optional implementations of this embodiment, the step of “perception-labeling the perception feature information based on the predefined service perception tag to obtain the perception feature information carrying the perception label” specifically includes the following steps:

[0067] The historical average write-off rate of the first target customer corresponding to the training data is calculated; the write-off rate of the first target customer is compared with the historical average write-off rate to obtain a comparison result; based on the comparison result and predefined service perception labels, the perception feature information is perception-labeled to obtain perception feature information carrying the perception label.

[0068] Among them, the historical average write-off rate refers to the average proportion of insurance write-offs (i.e., customers actually use or redeem insurance services) of the first target customers within a specific historical period.

[0069] Among them, the write-off rate refers to the proportion of insurance write-offs for each first target customer within a specific time period.

[0070] The comparison result refers to the conclusion obtained by comparing the write-off rate of the first target customer with the historical average write-off rate. This result is based on the difference between the two values and is derived through logical reasoning.

[0071] In one example, historical data can be collected for a first target customer, including insurance purchase and write-off records, and basic customer information. Information about the first target customer's insurance sales cycle, such as purchase and renewal dates, is obtained to divide the historical data into training data and prediction data. The write-off records in the training data are filtered and organized, and the number of write-offs and total purchases for each first target customer within a specific time period are counted. The historical average write-off rate is calculated, which is the average of the number of write-offs and total purchases for all first target customers within the specific time period. For each first target customer, the current write-off rate (i.e., the ratio of write-offs to total purchases within the current time period) is calculated. The current write-off rate is compared with the historical average write-off rate to obtain a comparison result. If the current write-off rate is higher than the historical average, it is labeled "high perception"; if it is lower, it is labeled "low perception." Based on the comparison result and predefined service perception labels (such as "high satisfaction" and "low satisfaction"), the perception feature information is perceptually labeled. For example, for “high perception” customers, their perception feature information is marked as positive features under the label of “high satisfaction”; for “low perception” customers, it is marked as negative features under the label of “low satisfaction”.

[0072] The embodiment of the present application can achieve a quantitative analysis of customer service perception by statistically analyzing the historical average write-off rate of the first target customer corresponding to the training data, and comparing the write-off rate of the first target customer with the historical average write-off rate. The perception feature information is perceptually labeled based on the comparison results and predefined service perception tags, which further improves the accuracy and practicality of the perception feature information. The perception feature information carrying the perception annotation not only reflects the customer's actual write-off behavior, but also incorporates the subjective evaluation of service perception, providing a rich and accurate data foundation for subsequent model training. This labeling method helps the model to more deeply understand the customer's service perception pattern, thereby more accurately predicting the customer's future service needs and behavior. Through this technical processing, high-value customers can be identified more accurately, personalized service strategies can be formulated, and the utilization efficiency of service resources can be improved.

[0073] In some optional implementations of this embodiment, step S203, based on the predefined activity tags, uses historical data to perform timing training on a preset second prediction model to obtain contact timing models for multiple time periods, specifically including the following steps:

[0074] Based on predefined contact timing feature information, feature extraction is performed on the training data to obtain timing feature information; the daily time of the first target customer is divided into multiple time periods to obtain the activity of the first target customer in each time period; based on the predefined activity label and activity, activity annotation is performed on the timing feature information to obtain timing feature information with activity annotation; based on the timing feature information with activity annotation, timing training is performed on the preset second prediction model to obtain contact timing models for multiple time periods.

[0075] In one example, feature extraction can be performed on training data based on predefined contact timing characteristics (such as time, date, and customer behavior) to obtain timing characteristics. For example, behavioral characteristics such as login frequency and browsing duration can be extracted for different time periods throughout the day. The first target customer's daily schedule can be divided into multiple time periods, such as morning, midday, afternoon, evening, and night. The first target customer's activity level in each time period can be measured by counting metrics such as the number of interactions and duration of stay during that time period. Based on predefined activity labels (such as high activity, medium activity, and low activity), the timing characteristics are labeled with activity levels based on the customer's activity level in each time period. For example, if a customer's login frequency and browsing duration are both high in the morning, the timing characteristics for that time period are labeled as high activity. Based on the activity-labeled timing characteristics, a preset second prediction model is trained to obtain contact timing models for multiple time periods. During model training, an optimization algorithm can be used to continuously adjust model parameters to improve the model's accuracy in predicting contact timing.

[0076] The embodiment of the present application can extract features from the training data based on predefined contact timing feature information to obtain timing feature information, and perform activity labeling based on the activity of the first target customer in each time period. This processing flow significantly improves the accuracy of contact timing prediction. In the field of finance and insurance, customer activity is often closely related to purchase intention, service demand, etc. Therefore, activity labeling of timing feature information can more accurately reflect customers' service contact preferences at different time periods. Based on the timing feature information carrying activity labels, the preset second prediction model is trained on timing, and the contact timing models for multiple time periods obtained can accurately predict the possibility of service contact for customers at different time periods, providing insurance companies with scientific customer contact strategies. This technical process not only avoids the waste of resources in the traditional wide-net service model, but also enables insurance companies to accurately locate customers based on individual differences and provide personalized service experiences.

[0077] In some optional implementations of this embodiment, the step of “obtaining the activity level of the first target customer in each time period” specifically includes the following steps:

[0078] Count the average number of active times of the first target customer in each time period; compare the number of active times of each first target customer in each time period with the average number of active times; and determine the activity level of the first target customer in each time period based on the comparison result.

[0079] In one example, the first target customer's daily schedule can be divided into multiple time periods, such as morning, midday, afternoon, evening, and night. The average number of active visits for all first target customers during each time period is calculated and used as a baseline activity indicator for that time period. For each first target customer, the number of active visits during each time period is calculated. The number of active visits during each time period is compared with the average number of active visits during the corresponding time period. Based on the comparison results, the customer's activity level during that time period is determined. For example, if a customer's activity level during the morning period is higher than the average number of active visits, the customer is determined to be highly active during the morning period.

[0080] The embodiment of the present application can determine the activity level of the first target customer in each time period by counting the average number of active times of the first target customer in each time period, and comparing the number of active times of each first target customer in each time period with the average number of active times. In the field of finance and insurance, customer activity is an important indicator that reflects the customer's interest in services or products, and is closely related to the customer's purchasing intention and service needs. Through refined time period division and active number statistics, it is possible to more accurately capture the customer's behavior patterns in different time periods, and provide data support for subsequent personalized services. The customer activity determined based on the comparison results not only reflects the customer's activity level in a specific time period, but also reveals the time preference of the customer's behavior.

[0081] In some optional implementations of this embodiment, step S205, determining a target service-conscious customer from the second target customers based on the service perception information, specifically includes the following steps:

[0082] Based on the service perception information, the second target customers are sorted in descending order to obtain a second target customer sequence; the number of pre-sales service personnel in the forecast period is obtained; and according to the number of pre-sales service personnel, a corresponding number of target service-conscious customers is obtained from the second target customer sequence.

[0083] In one example, the second target customers can be sorted in descending order based on the service perception information to obtain a second target customer sequence. For example, if the service perception information is based on satisfaction, customers with high satisfaction are ranked at the front of the sequence. Next, the number of pre-sales service personnel for the predicted period is obtained. For example, an insurance company predicts that 10 pre-sales service personnel will be available during the next business period. Based on the number of pre-sales service personnel, the corresponding number of target service-conscious customers is obtained from the second target customer sequence. In this example, the first 10 customers in the sequence are selected as target service-conscious customers.

[0084] The embodiment of the present application can sort the second target customers in descending order by the service perception information, and determine the target service-conscious customers in combination with the number of pre-sales service personnel in the forecast period. This processing flow has a significant technical effect. In the field of finance and insurance, the customer's service perception information directly reflects their satisfaction and potential value with the insurance company's services. By sorting the service perception information in descending order, customers with high service perception and high potential value can be preferentially screened out, providing strong support for subsequent precise services. At the same time, combined with the number of pre-sales service personnel in the forecast period, a corresponding number of target service-conscious customers are selected from the sorted customer sequence, avoiding the waste of service resources and ensuring the efficient use of service resources.

[0085] In some optional implementations of this embodiment, step S205 uses the contact timing model for each time period to perform time period prediction processing on the customer data of the target service-conscious customers and determine the recommended time period for each target service-conscious customer, specifically including the following steps:

[0086] Input the customer data of the target service-oriented customers into the contact timing model for each time period to obtain the activity probability of each target service-oriented customer in each time period; compare the activity probabilities for each time period to determine the time period when each target service-oriented customer has the highest activity; and determine the time period with the highest activity as the recommended time period for each target service-oriented customer.

[0087] In one example, the customer data of target service-oriented customers is input into a contact timing model for each time period to obtain the activity probability of each target service-oriented customer in each time period. For example, for customer A, the model predicts an activity probability of 0.7 in the morning, 0.3 in the afternoon, and 0.8 in the evening. The activity probabilities for each time period are compared to determine the time period in which each target service-oriented customer is most active. In this example, customer A's most active time period is the evening. The time period with the highest activity is determined as the recommended time period for each target service-oriented customer. That is, the recommended contact time period for customer A is the evening. Using a preset preference model, preference prediction processing is performed on the customer data of the target service-oriented customer to obtain the service preference information of the target service-oriented customer. For example, customer A prefers online consultation and insurance product B. During the recommended contact time period, the corresponding virtual product is distributed to the target service-oriented customer based on the service preference information. For example, an online consultation link for insurance product B is distributed to customer A in the evening.

[0088] The embodiment of the present application can input the customer data of the target service-oriented customers into the contact timing model of each time period to obtain the activity probability of each target service-oriented customer in each time period, and determine the time period with the highest activity as the recommended time period accordingly. In the field of finance and insurance, the customer's active time period is the key to formulating a precise contact strategy. Predicting the customer's active probability through the contact timing model can more accurately grasp the customer's behavior pattern in different time periods, thereby avoiding the waste of resources of traditional wide-ranging services. Specifically, through data analysis and model prediction, accurate positioning of customer active time periods is achieved. This not only helps insurance companies optimize service resource allocation and improve service efficiency, but also allows them to contact customers during the time periods when they are most active and most receptive to services, significantly improving customer response rates and satisfaction.

[0089] It should be emphasized that in order to further ensure the above historical data, customer data, service perception information, and service preference information, the above historical data, customer data, service perception information, and service preference information can also be stored in a node of a blockchain.

[0090] The blockchain referred to in this application is a new application model for computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Blockchain is essentially a decentralized database, a series of data blocks generated using cryptographic methods. Each data block contains information about a batch of network transactions, which is used to verify the validity of this information (to prevent counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, the platform product service layer, and the application service layer.

[0091] The embodiments of the present application can build and optimize related models and networks based on artificial intelligence technology, such as perception models, contact timing models, and preference models. Artificial Intelligence (AI) models are the theoretical and practical crystallization of solving complex problems, predicting future trends, or automating tasks by simulating the decision-making process of human intelligence through algorithms and data analysis. These models utilize large amounts of historical data and real-time information, and are trained and optimized through specific algorithmic frameworks to achieve efficient, accurate, and reliable performance.

[0092] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware via computer-readable instructions. The computer-readable instructions can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes in the above-described method embodiments. The aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0093] It should be understood that although the steps in the flowcharts of the accompanying drawings are shown in sequence as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some of the steps in the flowcharts of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0094] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of a virtual product distribution device. Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0095] like Figure 3 As shown, the virtual product distribution device 400 of this embodiment includes: an acquisition module 401, a perception training module 402, a timing training module 403, a perception prediction module 404, a time period prediction module 405, a preference prediction module 406 and a distribution module 407. Among them:

[0096] An acquisition module 401 is used to acquire historical data of a first target customer;

[0097] A perception training module 402 is configured to perform perception training on a preset first prediction model using historical data based on a predefined service perception tag to obtain a perception model;

[0098] A timing training module 403 is configured to perform timing training on a preset second prediction model using historical data based on predefined activity tags to obtain contact timing models for multiple time periods;

[0099] A perception prediction module 404 is configured to obtain customer data of the second target customers, perform perception prediction processing on the customer data using a perception model, and obtain service perception information of each second target customer;

[0100] The time period prediction module 405 is used to determine target service-conscious customers from the second target customers based on the service perception information, use the contact timing model for each time period to perform time period prediction processing on the customer data of the target service-conscious customers, and determine the recommended time period for each target service-conscious customer;

[0101] The preference prediction module 406 is used to perform preference prediction processing on the customer data of the target service-conscious customer using a preset preference model to obtain service preference information of the target service-conscious customer;

[0102] The distribution module 407 is used to distribute corresponding virtual products to target service-conscious customers based on the service preference information within the recommended contact period.

[0103] The embodiment of the present application can obtain the historical data of the first target customer and perform perception training on the preset first prediction model based on the predefined service perception label. The obtained perception model can deeply explore the customer's service needs and perception preferences, and provide accurate guidance for subsequent services. Then, the second prediction model is trained on the timing using the predefined activity label to construct a contact timing model for multiple time periods, which helps to grasp the best time for customers to receive services, avoid ineffective communication, and improve the utilization rate of service resources. Furthermore, by performing perception prediction processing on the customer data of the second target customer through the perception model, the target service-conscious customer can be accurately identified, and the recommended time period can be determined in combination with the contact timing model to achieve personalized and accurate service push. Finally, the preset preference model is used to predict the preferences of the target service-conscious customer, and the corresponding virtual products are distributed accordingly, which not only meets the diverse needs of customers, but also enhances the customer's sense of identity and loyalty to the insurance company, thereby significantly improving the effect of pre-sales service and customer satisfaction while controlling costs.

[0104] In one embodiment, the perception training module 402 includes:

[0105] A first acquisition submodule is used to obtain insurance sales cycle information of a first target customer and divide historical data into training data and prediction data based on the insurance sales cycle;

[0106] A first extraction submodule is used to extract features from the training data based on predefined perceptual feature information to obtain perceptual feature information;

[0107] The perception annotation submodule is used to perform perception annotation on the perception feature information based on the predefined service perception tag to obtain the perception feature information carrying the perception annotation;

[0108] A first training submodule is configured to perform binary classification training on a preset first prediction model based on the perceptual feature information carrying the perceptual annotation and an optimization algorithm to obtain a perceptual model;

[0109] The verification submodule is used to use the prediction data to verify the performance of the perception model and obtain a verified perception model.

[0110] The embodiment of the present application can achieve effective organization and utilization of data by obtaining the insurance sales cycle information of the first target customer and dividing the historical data into training data and prediction data based on the cycle. This step ensures that the model training can closely fit the actual cycle of the insurance business and improves the prediction accuracy of the model. Furthermore, based on the predefined perceptual feature information, feature extraction is performed on the training data to obtain perceptual feature information. This process deeply explores the customer's potential demand for insurance services and provides strong support for subsequent precise services. By perceptually labeling the perceptual feature information and performing binary classification training on the preset first prediction model based on the feature information carrying the perceptual labeling and the optimization algorithm, a perceptual model is finally obtained. This model can accurately predict the customer's perception of insurance services and provide a scientific decision-making basis for insurance companies. The performance of the perceptual model is verified using prediction data to ensure the stability and reliability of the model. The comprehensive application of this series of technical means not only improves the accuracy of insurance services and customer satisfaction, but also effectively reduces service costs.

[0111] In one embodiment, the perception labeling submodule is also used to count the historical average write-off rate of the first target customer corresponding to the training data; compare the write-off rate of the first target customer with the historical average write-off rate to obtain a comparison result; based on the comparison result and the predefined service perception label, perform perception labeling on the perception feature information to obtain perception feature information carrying the perception label.

[0112] The embodiment of the present application can achieve a quantitative analysis of customer service perception by statistically analyzing the historical average write-off rate of the first target customer corresponding to the training data, and comparing the write-off rate of the first target customer with the historical average write-off rate. The perception feature information is perceptually labeled based on the comparison results and predefined service perception tags, which further improves the accuracy and practicality of the perception feature information. The perception feature information carrying the perception annotation not only reflects the customer's actual write-off behavior, but also incorporates the subjective evaluation of service perception, providing a rich and accurate data foundation for subsequent model training. This labeling method helps the model to more deeply understand the customer's service perception pattern, thereby more accurately predicting the customer's future service needs and behavior. Through this technical processing, high-value customers can be identified more accurately, personalized service strategies can be formulated, and the utilization efficiency of service resources can be improved.

[0113] In one embodiment, the timing training module 403 includes:

[0114] A second extraction submodule is configured to extract features from the training data based on predefined contact timing feature information to obtain timing feature information;

[0115] A division submodule is used to divide the first target customer's daily time into multiple time periods and obtain the first target customer's activity level in each time period;

[0116] An activity labeling submodule is used to label the timing feature information with activity based on a predefined activity tag and activity level, thereby obtaining the timing feature information carrying the activity label;

[0117] The second training submodule is used to perform timing training on a preset second prediction model based on the timing feature information carrying the activity annotation to obtain contact timing models for multiple time periods.

[0118] The embodiment of the present application can extract features from the training data based on predefined contact timing feature information to obtain timing feature information, and perform activity labeling based on the activity of the first target customer in each time period. This processing flow significantly improves the accuracy of contact timing prediction. In the field of finance and insurance, customer activity is often closely related to purchase intention, service demand, etc. Therefore, activity labeling of timing feature information can more accurately reflect customers' service contact preferences at different time periods. Based on the timing feature information carrying activity labels, the preset second prediction model is trained on timing, and the contact timing models for multiple time periods obtained can accurately predict the possibility of service contact for customers at different time periods, providing insurance companies with scientific customer contact strategies. This technical process not only avoids the waste of resources in the traditional wide-net service model, but also enables insurance companies to accurately locate customers based on individual differences and provide personalized service experiences.

[0119] In one embodiment, the partitioning submodule is further used to count the average number of active times of the first target customer in each time period; compare the number of active times of each first target customer in each time period with the average number of active times; and determine the activity level of the first target customer in each time period based on the comparison result.

[0120] The embodiment of the present application can determine the activity level of the first target customer in each time period by counting the average number of active times of the first target customer in each time period, and comparing the number of active times of each first target customer in each time period with the average number of active times. In the field of finance and insurance, customer activity is an important indicator that reflects the customer's interest in services or products, and is closely related to the customer's purchasing intention and service needs. Through refined time period division and active number statistics, it is possible to more accurately capture the customer's behavior patterns in different time periods, and provide data support for subsequent personalized services. The customer activity determined based on the comparison results not only reflects the customer's activity level in a specific time period, but also reveals the time preference of the customer's behavior.

[0121] In one embodiment, the time period prediction module 405 includes:

[0122] an arrangement submodule, configured to arrange the second target customers in descending order based on the service perception information to obtain a second target customer sequence;

[0123] The second acquisition submodule is used to obtain the number of pre-sales service personnel in the forecast period;

[0124] The third acquisition submodule is configured to acquire a corresponding number of target service-sense customers from the second target customer sequence according to the number of pre-sales service personnel.

[0125] The embodiment of the present application can sort the second target customers in descending order by the service perception information, and determine the target service-conscious customers in combination with the number of pre-sales service personnel in the forecast period. This processing flow has a significant technical effect. In the field of finance and insurance, the customer's service perception information directly reflects their satisfaction and potential value with the insurance company's services. By sorting the service perception information in descending order, customers with high service perception and high potential value can be preferentially screened out, providing strong support for subsequent precise services. At the same time, combined with the number of pre-sales service personnel in the forecast period, a corresponding number of target service-conscious customers are selected from the sorted customer sequence, avoiding the waste of service resources and ensuring the efficient use of service resources.

[0126] In one embodiment, the time period prediction module 405 includes:

[0127] The input submodule is used to input the customer data of the target service-sense customer into the contact timing model of each time period to obtain the activity probability of each target service-sense customer in each time period;

[0128] The comparison submodule is used to compare the activity probability of each time period and determine the time period with the highest activity of each target service-sense customer;

[0129] The determination submodule is used to determine the time period with the highest activity as the recommended time period for each target service-conscious customer.

[0130] The embodiment of the present application can input the customer data of the target service-oriented customers into the contact timing model of each time period to obtain the activity probability of each target service-oriented customer in each time period, and determine the time period with the highest activity as the recommended time period accordingly. In the field of finance and insurance, the customer's active time period is the key to formulating a precise contact strategy. Predicting the customer's active probability through the contact timing model can more accurately grasp the customer's behavior pattern in different time periods, thereby avoiding the waste of resources of traditional wide-ranging services. Specifically, through data analysis and model prediction, accurate positioning of customer active time periods is achieved. This not only helps insurance companies optimize service resource allocation and improve service efficiency, but also allows them to contact customers during the time periods when they are most active and most receptive to services, significantly improving customer response rates and satisfaction.

[0131] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0132] The computer device 6 includes a memory 61, a processor 62, and a network interface 63 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 6 with a memory 61, a processor 62, and a network interface 63, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to a microprocessor, an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a digital signal processor (DSP), an embedded device, etc.

[0133] Computer devices can be desktop computers, laptops, PDAs, cloud servers, etc. Computer devices can interact with users through keyboards, mice, remote controls, touchpads, or voice-activated devices.

[0134] Memory 61 includes at least one type of readable storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic storage, magnetic disk, optical disk, etc. In some embodiments, memory 61 may be an internal storage unit of computer device 6, such as the hard disk or memory of computer device 6. In other embodiments, memory 61 may also be an external storage device of computer device 6, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash memory card, etc. equipped on computer device 6. Of course, memory 61 may also include both internal storage units and external storage devices of computer device 6. In this embodiment, memory 61 is generally used to store the operating system and various application software installed on computer device 6, such as computer-readable instructions for the virtual product distribution method. In addition, memory 61 may also be used to temporarily store various types of data that have been output or are about to be output.

[0135] In some embodiments, processor 62 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip. Processor 62 is generally used to control the overall operation of computer device 6. In this embodiment, processor 62 is used to execute computer-readable instructions stored in memory 61 or process data, such as computer-readable instructions for executing a virtual product distribution method.

[0136] The network interface 63 may include a wireless network interface or a wired network interface. The network interface 63 is generally used to establish a communication connection between the computer device 6 and other electronic devices.

[0137] The embodiment of the present application can obtain the historical data of the first target customer and perform perception training on the preset first prediction model based on the predefined service perception label. The obtained perception model can deeply explore the customer's service needs and perception preferences, and provide accurate guidance for subsequent services. Then, the second prediction model is trained on the timing using the predefined activity label to construct a contact timing model for multiple time periods, which helps to grasp the best time for customers to receive services, avoid ineffective communication, and improve the utilization rate of service resources. Furthermore, by performing perception prediction processing on the customer data of the second target customer through the perception model, the target service-conscious customer can be accurately identified, and the recommended time period can be determined in combination with the contact timing model to achieve personalized and accurate service push. Finally, the preset preference model is used to predict the preferences of the target service-conscious customer, and the corresponding virtual products are distributed accordingly, which not only meets the diverse needs of customers, but also enhances the customer's sense of identity and loyalty to the insurance company, thereby significantly improving the effect of pre-sales service and customer satisfaction while controlling costs.

[0138] The present application also provides another embodiment, namely, providing a computer-readable storage medium, which stores computer-readable instructions. The computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the virtual product distribution method as described above.

[0139] The embodiment of the present application can obtain the historical data of the first target customer and perform perception training on the preset first prediction model based on the predefined service perception label. The obtained perception model can deeply explore the customer's service needs and perception preferences, and provide accurate guidance for subsequent services. Then, the second prediction model is trained on the timing using the predefined activity label to construct a contact timing model for multiple time periods, which helps to grasp the best time for customers to receive services, avoid ineffective communication, and improve the utilization rate of service resources. Furthermore, by performing perception prediction processing on the customer data of the second target customer through the perception model, the target service-conscious customer can be accurately identified, and the recommended time period can be determined in combination with the contact timing model to achieve personalized and accurate service push. Finally, the preset preference model is used to predict the preferences of the target service-conscious customer, and the corresponding virtual products are distributed accordingly, which not only meets the diverse needs of customers, but also enhances the customer's sense of identity and loyalty to the insurance company, thereby significantly improving the effect of pre-sales service and customer satisfaction while controlling costs.

[0140] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods of each embodiment of the present application.

[0141] Obviously, the embodiments described above are only some of the embodiments of the present application, rather than all of the embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application has been described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific embodiments, or to make equivalent replacements for some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is also within the scope of patent protection of this application. The non-company enterprise software tools or components that appear in the embodiments of this application are merely examples and do not represent actual use.

Claims

1. A virtual product distribution method, characterized in that: The steps include: Obtain historical data of the first target customer; Based on the predefined service perception tag, using the historical data to perform perception training on a preset first prediction model to obtain a perception model; Based on the predefined activity tags, the historical data is used to perform timing training on a preset second prediction model to obtain contact timing models for multiple time periods; Acquiring customer data of second target customers, and performing perception prediction processing on the customer data using the perception model to obtain service perception information of each second target customer; Based on the service perception information, target service-conscious customers are determined from the second target customers, and a contact timing model for each time period is used to perform time period prediction processing on the customer data of the target service-conscious customers to determine a recommended time period for each target service-conscious customer; Using a preset preference model, performing preference prediction processing on the customer data of the target service-conscious customer to obtain service preference information of the target service-conscious customer; During the recommended contact period, corresponding virtual products are distributed to the target service-conscious customers based on the service preference information.

2. The method according to claim 1, characterized in that The step of performing perception training on a preset first prediction model using the historical data based on the predefined service perception tag to obtain a perception model specifically includes: Obtaining insurance sales cycle information of the first target customer, and dividing the historical data into training data and prediction data based on the insurance sales cycle; Based on predefined perceptual feature information, feature extraction is performed on the training data to obtain perceptual feature information; Based on a predefined service perception tag, the perception feature information is perception-labeled to obtain perception feature information carrying the perception label; Based on the perceptual feature information carrying the perceptual annotation and the optimization algorithm, a preset first prediction model is trained in binary classification to obtain a perceptual model; The prediction data is used to perform performance verification on the perception model to obtain a verified perception model.

3. The method according to claim 2, characterized in that The step of performing perception annotation on the perception feature information based on the predefined service perception tag to obtain the perception feature information carrying the perception annotation specifically includes: Calculate the historical average write-off rate of the first target customer corresponding to the training data; Comparing the write-off rate of the first target customer with the historical average write-off rate to obtain a comparison result; Based on the comparison result and a predefined service perception tag, the perception feature information is perception-labeled to obtain perception feature information carrying the perception label.

4. The method according to claim 2, characterized in that The step of performing timing training on a preset second prediction model using the historical data based on the predefined activity tag to obtain contact timing models for multiple time periods specifically includes: Based on predefined contact timing feature information, feature extraction is performed on the training data to obtain timing feature information; Divide the first target customer's daily time into multiple time periods, and obtain the first target customer's activity level in each time period; Based on a predefined activity tag and the activity, the timing feature information is annotated with activity to obtain the timing feature information carrying the activity annotated information; Based on the timing feature information carrying the activity annotation, timing training is performed on a preset second prediction model to obtain contact timing models for the multiple time periods.

5. The method according to claim 4, characterized in that The step of obtaining the activity level of the first target customer in each time period specifically includes: Counting the average number of active times of the first target customer in each time period; comparing the number of active times of each first target customer in each time period with the average number of active times; Based on the comparison result, the activity level of the first target customer in each time period is determined.

6. The method according to claim 1, characterized in that The step of determining target service-conscious customers from the second target customers based on the service perception information specifically includes: Arranging the second target customers in descending order based on the service perception information to obtain a second target customer sequence; Get the number of pre-sales service personnel in the forecast period; According to the number of the pre-sales service personnel, a corresponding number of target service-sense customers is obtained from the second target customer sequence.

7. The method according to claim 1, characterized in that The step of using the contact timing model for each time period to perform time period prediction processing on the customer data of the target service-conscious customers and determining the recommended time period for each target service-conscious customer specifically includes: Inputting the customer data of the target service-conscious customers into the contact timing model for each time period to obtain the activity probability of each target service-conscious customer in each time period; Comparing the activity probabilities of each time period to determine the time period when each target service-conscious customer has the highest activity; The time period with the highest activity is determined as the recommended time period for each target service-conscious customer.

8. A virtual product distribution device, characterized in that: include: An acquisition module, used to acquire historical data of the first target customer; A perception training module, configured to perform perception training on a preset first prediction model using the historical data based on a predefined service perception tag to obtain a perception model; a timing training module, configured to perform timing training on a preset second prediction model using the historical data based on a predefined activity tag to obtain contact timing models for multiple time periods; a perception prediction module, configured to obtain customer data of second target customers, perform perception prediction processing on the customer data using the perception model, and obtain service perception information of each second target customer; a time period prediction module, configured to determine target service-conscious customers from the second target customers based on the service perception information, perform time period prediction processing on the customer data of the target service-conscious customers using a contact timing model for each time period, and determine a recommended time period for each target service-conscious customer; a preference prediction module, configured to perform preference prediction processing on the customer data of the target service-conscious customer using a preset preference model to obtain service preference information of the target service-conscious customer; A distribution module is used to distribute corresponding virtual products to the target service-conscious customers based on the service preference information within the recommended contact period.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the virtual product distribution method according to any one of claims 1 to 7 when executing the computer-readable instructions.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the virtual product distribution method according to any one of claims 1 to 7.

Images