TWI777116B - Method and system for generating personalized marketing offers - Google Patents

Abstract

A system for generating personalized marketing privileges. After receiving a target customer characteristic information and a target merchant identification information from a target merchant, a service server obtains corresponding target customer characteristic information according to the target customer characteristic information and a look-up table. a target customer information, a marketing discount generating device receives the target customer information and the target merchant identification information, and obtains multiple pieces of target consumption data according to the target customer information and the target merchant identification information, and according to the target consumption data, Obtain a lifetime value data, and then generate the discount message according to the lifetime value data. After receiving the target customer information and the discount message, a push server transmits the discount message to the customer corresponding to the target customer information according to the target customer information. end.

Description

Method and system for generating personalized marketing offers

The present invention relates to a method for generating an offer, and more particularly, to a method and system for generating a personalized marketing offer.

In general e-commerce, electronic coupons are often provided to attract consumers to purchase goods. In order to dig out potential buyers, the backend of e-commerce platforms generally collects the historical data of platform users’ Internet activities, and uses data mining to predict the current purchase intentions of platform users based on their browsing records. , and provide product coupons according to the current purchase intentions of platform users.

However, in the existing physical storefronts, it is impossible to predict the current purchase intention of consumers based on the historical data of consumers' Internet activities. The existing coupons, for example, are provided to the public with the same discount serial number, which cannot be based on consumers' purchase behavior. Offer personalised offers to improve customer satisfaction in the store.

Therefore, the object of the present invention is to provide a Personalized marketing offer generation methods that act to provide personalized offers.

Therefore, the method for generating personalized marketing privileges of the present invention is executed by a marketing privilege generating system, the marketing privilege generating system is connected to a plurality of merchants via a communication network, and the marketing privilege generating system stores a lookup table of customer characteristic information to customer information and a plurality of pieces of consumption data, the lookup table includes a plurality of pieces of customer characteristic information respectively related to the customers, and a plurality of customer information respectively corresponding to the customer characteristic information and corresponding to a plurality of customer terminals, each consumption data includes a customer information, a consumption time, a consumption amount, and a merchant identification information. The method for generating personalized marketing offers includes: a step (A), a step (B), a step (C), a step (D), a step (E), and a step (F).

In step (A), after the marketing offer generation system receives a target customer characteristic information and a target merchant identification information related to the target merchant end from one of the merchant ends, the marketing offer is generated The system obtains a target customer information corresponding to the target customer characteristic information according to the target customer characteristic information and the look-up table.

In the step (B), the marketing offer generation system obtains a plurality of pieces of target consumption data including the target customer information and the target merchant identification information from the consumption data according to the target customer information and the target merchant identification information .

In the step (C), the marketing offer generation system obtains a data including a predicted consumption frequency, a predicted active time, and a predicted consumption according to the target consumption data. Lifetime value information for the fee amount.

In step (D), the marketing offer generation system determines whether to generate an offer message according to the lifetime value data.

In the step (E), when the marketing offer generating system determines that the offer message needs to be generated, the marketing offer generating system generates the offer message according to the lifetime value data.

In the step (F), the marketing offer generation system transmits the offer message to the client corresponding to the target customer information according to the target customer information.

Another object of the present invention is to provide a system for generating personalized marketing offers that can provide personalized offers according to consumers' purchasing behavior.

The system for generating personalized marketing privileges of the present invention includes a service server, a marketing privilege generating device, and a push server.

The service server stores a lookup table of customer characteristic information to customer information, the lookup table includes a plurality of pieces of customer characteristic information respectively related to the customers, and a plurality of pieces of customer characteristic information respectively corresponding to the customer characteristic information and corresponding to a plurality of clients. Customer information, the service server is connected to the merchant terminals via the first communication network, after receiving a target customer characteristic information and a target merchant identification information related to the target merchant terminal from one of the target merchant terminals of the merchant terminals , and obtain a target customer information corresponding to the target customer characteristic information according to the target customer characteristic information and the look-up table.

The marketing offer generating device stores a plurality of consumption data, each consumption data Including and a customer information, a consumption time, a consumption amount, and a merchant identification information. The marketing offer generating device is connected to the service server via a second communication network, receives the target customer information and the target merchant identification information from the service server, and, according to the target customer information and the target merchant identification information, generates a Among the consumption data, a plurality of pieces of target consumption data including the target customer information and the target merchant identification information are obtained, and according to the target consumption data, one including a predicted consumption frequency, a predicted active time, and a predicted consumption are obtained. Lifetime value data of the amount, and according to the lifetime value data, it is determined whether a discount message needs to be generated, and when it is determined that the discount message needs to be generated, the discount message is generated.

The push server is connected to the marketing offer generating device via the second communication network, and after the push server receives the target customer information and the offer message from the marketing offer generating device, the push server transmits the information via the target customer information according to the target customer information. The first communication network transmits the preferential message to the customer end corresponding to the target customer information.

The effect of the present invention lies in: obtaining the lifetime value data according to the target consumption data by the marketing preferential generating device, and generating the discount message according to the lifetime value data, so as to improve the consumer satisfaction in the store.

11: Service Server

12: Marketing Offer Generation Device

13: Push server

100: First Communication Network

101: Merchant side

102: Second Communication Network

21~27: Steps

241, 242: Substeps

251, 252: Substeps

Other features and effects of the present invention will be clearly presented in the embodiments with reference to the drawings, wherein: 1 is a block diagram illustrating an embodiment of a system for generating personalized marketing offers of the present invention; FIG. 2 is a flowchart illustrating an embodiment of a method for generating personalized marketing offers according to the present invention; FIG. 2 illustrates the sub-steps of step 23; and FIG. 4 is a flow chart that assists FIG. 2 in illustrating the sub-steps of step 24. FIG.

Before the present invention is described in detail, it should be noted that in the following description, similar elements are designated by the same reference numerals.

Referring to FIG. 1 , an embodiment of the system for generating personalized marketing offers of the present invention includes a service server 11 , a marketing offer generating device 12 , and a push server 13 .

The service server 11 stores a lookup table of customer characteristic information to customer information, the lookup table includes a plurality of pieces of customer characteristic information respectively related to the customers, and a plurality of pieces of customer characteristic information respectively corresponding to the corresponding customer characteristic information and corresponding to a plurality of customers respectively The customer information of the terminal (not shown) is connected to a plurality of merchant terminals 101 via the first communication network 100, such as the Internet. It should be noted that the merchant terminals 101 are, for example, Internet of Things (IoT) devices. For each merchant terminal 101 , the merchant terminal 101 can capture the facial features of a target customer and generate Generate a target customer feature information including the intercepted facial features; and in another implementation aspect, for each merchant end 101, the merchant end 101 sends a beacon (Beacon) related to its corresponding The beacon data of the merchant 101, after a target customer used by the target customer receives the beacon data, the target customer sends a target customer characteristic information including an identification data related to the target customer to the target customer. The merchant terminal 101, wherein the target customer terminal is, for example, a mobile device installed with an application program related to the beacon.

The marketing privilege generating device 12 stores a plurality of consumption data and a plurality of threshold values corresponding to the corresponding merchant terminals 101 respectively. Each consumption data includes a customer information, a consumption time, a consumption amount, and a merchant identification information. The second communication network 102 is connected to the service server 11 . The second communication network 102 is, for example, an intranet of an enterprise. The second communication network 102 can also be the Internet, but not limited thereto.

The push server 13 is connected to the first communication network 100 and is connected to the marketing offer generating device 12 via the second communication network 102 .

Referring to FIG. 1 and FIG. 2 , an embodiment of how the system for generating a personalized marketing offer of the present invention executes an embodiment of the method for generating a personalized marketing offer of the present invention is described.

In step 21 , the service server 11 receives, via the first communication network 100 , a target customer characteristic information and a target merchant related to the target merchant terminal 101 from one of the merchant terminals 101 . After identifying the information, obtain the target customer characteristic information corresponding to the target customer characteristic information according to the target customer characteristic information and the look-up table. a target customer information, and transmit the target customer information to the marketing offer generating device 12 via the second communication network 102 .

In step 22, the marketing offer generating device 12 obtains a plurality of pieces of target consumption data including the target customer information and the target merchant identification information from the consumption data according to the target customer information and the target merchant identification information.

In step 23, the marketing offer generating device 12 obtains a lifetime value including a predicted consumption frequency, a predicted active time, a predicted consumption amount, and a target customer lifetime value according to the target consumption data material.

其中，Y _i 為顧客i的消費次數，λ _i 為顧客i在一預設單位時間內的平均消費次數，τ _i 為顧客i流失的時間點，T _i 為一特定時間點，τ _i >T為顧客i在T _i 尚未流失，該預設單位時間例如為一年，在其他實施方式，該預設單位時間亦可為一個月，不以此為限。值得注意的是，式(1)的該預測消費頻率模型中的λ _i 是視為隨機變數(random variable)，然而顧客i於不同期間的λ _i 存在異質性 (heterogeneity)，例如於特定期間消費頻率高，但其他時間消費頻率低，故λ _i 異質性模式的該預測消費頻率模型為伽瑪分配(Gamma distribution)，如下式：

其中，γ及α為預設參數，又稱超參數(hyperparameter)。 The marketing discount generating device 12 obtains the predicted consumption frequency of the lifetime value data according to a predicted consumption frequency model, and the predicted consumption frequency model is a Poisson distribution, as shown in the following formula

Among them, Y _i is the consumption times of customer i , λ _i is the average consumption times of customer i in a preset unit time, τ _i is the time point when customer i is lost, T _i is a specific time point, τ _i > T Since the customer i has not been lost in T _i , the preset unit time is, for example, one year. In other embodiments, the preset unit time can also be one month, which is not limited thereto. It is worth noting that λ _i in the predicted consumption frequency model of Eq. (1) is regarded as a random variable, but there is heterogeneity in λ _i of customer i in different periods, such as consumption in a specific period The frequency is high, but the consumption frequency at other times is low, so the predicted consumption frequency model of the λ _i heterogeneity model is the Gamma distribution, as follows:

Among them, γ and α are preset parameters, also known as hyperparameters.

其中，τ _i 為顧客i流失的時間點，μ _i 為顧客i在該預設單位時間內的平均流失率。值得注意的是，式(3)的該預測活躍時間模型中的μ _i 是視為隨機變數，然而顧客i於不同期間的μ _i 存在異質性，例如於特定期間流失率高，但其他時間流失率低，故μ _i 異質性模式的該預測活躍時間模型為伽瑪分配，如下式：

其中，s及β為預設參數，又稱超參數。 The marketing privilege generating device 12 obtains the predicted active time of the lifetime value data according to a predicted active time model, and the predicted active time model is an exponential distribution, as follows:

Among them, τ _i is the time point when customer i churns, and μ _i is the average churn rate of customer i in the preset unit time. It is worth noting that μ _i in the predictive active time model of Eq. (3) is regarded as a random variable, but there is heterogeneity in μ _i of customer i in different periods, for example, the churn rate is high in a certain period, but churn in other times. Therefore, the predicted active time model of the μ _i heterogeneity pattern is a gamma allocation, as follows:

Among them, s and β are preset parameters, also known as hyperparameters.

It should be noted that the above equations (1)~(4) can be used to estimate the parameters γ , α , s , β and predict the parameters γ , α , s , β after using the Monte Carlo Markov Chain Monte Carlo (MCMC) iterative algorithm. .

其中，m _i 為顧客i在該預設單位時間內的平均消費金額，p為伽瑪分配的形狀數(Shape Parameter)，v為伽瑪分配的尺度參數(Scale Parameter)。值得注意的是，式(5)的該預測消費金額模型中的m _i 是視為隨機變數，其期望值(Expected Value)為θ _i =p/v，然而顧客i於不同期間的θ _i 存在異質性，例如於特定期間消費金額高，但其他時間消費金額低，故θ _i 異質性模式的該預測消費金額模型為伽瑪分配，如下式：

其中，θ _i 為顧客i在該預設單位時間內的期望值，q及r為預設參數，又稱超參數。 The marketing discount generating device 12 obtains the predicted consumption amount of the lifetime value data according to a predicted consumption amount model, and the predicted consumption amount model is gamma distribution, as follows:

Among them, m _i is the average consumption amount of customer i in the preset unit time, p is the shape number (Shape Parameter) allocated by gamma, and v is the scale parameter (Scale Parameter) allocated by gamma. It is worth noting that m _i in the predicted consumption amount model of Eq. (5) is regarded as a random variable, and its expected value (Expected Value) is θ _i = p / v , but there is heterogeneity in θ _i of customer i in different periods For example, the consumption amount is high in a certain period, but the consumption amount is low at other times, so the predicted consumption amount model of the θ _i heterogeneity model is a gamma distribution, as follows:

Among them, θ _i is the expected value of customer i in the preset unit time, and q and r are preset parameters, also known as hyperparameters.

其中，M _i 為顧客i在該預設單位時間內的平均消費金額(Mean Transaction Monetary Value)。 It should be noted again that equations (5) and (6) can be combined into an inverse beta distribution, as follows:

Wherein, M _i is the average consumption amount (Mean Transaction Monetary Value) of customer i in the preset unit time.

It is worth noting that, in this embodiment, the marketing discount generating device 12 first obtains the predicted consumption frequency, the predicted active time, and the predicted consumption amount, and then obtains the predicted consumption frequency, the predicted active time, and the predicted consumption amount. The predicted consumption amount obtains the target customer lifetime value, wherein the target customer lifetime value is the product of the predicted consumption frequency, the predicted active time, and the predicted consumption amount. For example, if the predicted consumption frequency is 50 times per year, the predicted active time is 5 years, and the predicted consumption amount is 100 yuan per time, the lifetime value of the target customer is 50*5*100=25,000 yuan.

In step 24, the marketing offer generating device 12 determines whether to generate an offer message according to the lifetime value data. When the marketing discount generating device 12 determines that the discount message needs to be generated, the process goes to step 25; and when the marketing discount generating device 12 determines that the discount message does not need to be generated, the process ends.

Referring to FIG. 3 , step 24 includes sub-steps 241 to 243 , and the sub-steps of step 24 are described below.

In step 241, the marketing discount generating device 12 obtains a pair of target threshold values corresponding to the target merchant terminal 101 from the threshold values according to the target merchant identification information.

In step 242, the marketing offer generating device 12 determines whether the target customer lifetime value is greater than the target threshold value. When the marketing offer generating device 12 determines that the target customer lifetime value is greater than the target threshold, it means that the offer needs to be generated and when the marketing offer generating device 12 determines that the target customer lifetime value is not greater than the target threshold value, it means that the offer message does not need to be generated.

In step 25, the marketing offer generating device 12 generates the offer message according to the target customer lifetime value and the multiple pieces of target merchant consumption data including the target merchant identification information in the consumption data.

Referring to FIG. 4 , step 25 includes sub-steps 251 and 252 , and the sub-steps of step 25 are described below.

In step 251, the marketing privilege generating device 12 obtains an average customer lifetime value of all customers of the target merchant 101 and a customer lifetime value of all customers of the target merchant 101 according to the target merchant consumption data standard deviation.

其中，x _CLV 為該目標顧客終身價值，μ _CLV 為該顧客終身價值平均，s _CLV 為該顧客終身價值標準差。 In step 252, the marketing offer generating device 12 generates the offer message according to the target customer lifetime value, the average customer lifetime value, and the customer lifetime value standard deviation. The discount message includes a discount number, and the discount number is calculated by the following formula:

Among them, x _CLV is the lifetime value of the target customer, μ _CLV is the average lifetime value of the customer, and s _CLV is the standard deviation of the customer's lifetime value.

It is worth noting that in other embodiments, the marketing offer generating device 12 The promotion information may be generated only based on the lifetime value data, but not limited thereto.

In step 26 , the marketing offer generating device 12 transmits the target customer information and the offer message to the push server 13 via the second communication network 102 .

In step 27, the push server 13 transmits the preferential message to the client corresponding to the target customer information according to the target customer information. It should be noted that, in this embodiment, the push server 13 transmits the discount message to the client via the first communication network 100 . In other embodiments, the push server 13 can also send a short message Sending the promotion message to the customer is not limited thereto.

To sum up, in the method and system for generating a personalized marketing offer of the present invention, the service server 11 obtains the target customer information according to the target customer characteristic information and the look-up table, and the marketing offer generating device 12 obtains the target customer information according to the target customer information and the identification information of the target merchant, obtain the target consumption data, and then obtain the lifetime value data according to the target consumption data, and after determining that the preferential message needs to be generated, generate the preferential message and transmit the target customer The information and the discount message are sent to the push server 13, so that the push server 13 transmits the discount message to the client terminal corresponding to the target customer information according to the target customer information, so as to improve the consumer satisfaction in the store. It can indeed achieve the purpose of the present invention.

However, the above are only examples of the present invention, and should not limit the scope of implementation of the present invention. Any simple equivalent changes and modifications made according to the scope of the patent application of the present invention and the contents of the patent specification are still included in the scope of the present invention. The scope of the invention patent Inside.

11: Service Server

12: Marketing Offer Generation Device

13: Push server

100: First Communication Network

101: Merchant side

102: Second Communication Network

Claims (16)

A method for generating personalized marketing privileges is executed by a marketing privilege generating system, the marketing privilege generating system is connected to a plurality of merchants via a communication network, and the marketing privilege generating system stores a lookup table of customer characteristic information to customer information, a plurality of respectively corresponding to the threshold value of the corresponding merchant side and multiple pieces of consumption data, the lookup table includes multiple pieces of customer characteristic information respectively related to the customers, and a plurality of pieces of customer characteristic information respectively corresponding to the corresponding customer characteristic information and corresponding to the multiple customer terminals. information, each consumption data includes a customer information, a consumption time, a consumption amount, and a merchant identification information, the personalized marketing discount generating method includes: (A) after receiving a target merchant from one of the merchants After a target customer characteristic information and a target merchant identification information related to the target merchant terminal on the end, a target customer information corresponding to the target customer characteristic information is obtained according to the target customer characteristic information and the look-up table; (B) according to the target customer information and the identification information of the target merchant, from the consumption data, obtain multiple pieces of target consumption data including the target customer information and the identification information of the target merchant; (C) according to the target consumption data, obtain an item including a predicted consumption frequency, a predicted active time, a target customer lifetime value, and a lifetime value data of a predicted consumption amount; (D) according to the lifetime value data to determine whether a discount message needs to be generated, step (D) includes the following sub-steps: (D) -1) Obtain a pair of target thresholds corresponding to the target merchant from the thresholds according to the target merchant identification information, and (D-2) determine whether the lifetime value of the target customer of the lifetime value data is greater than the target threshold, When it is determined that the target customer lifetime value of the lifetime value data is greater than the target threshold value, the discount message needs to be generated; (E) when it is determined that the discount message needs to be generated, according to the lifetime value data and the consumption data including To generate the discount message, step (E) includes the following sub-steps: (E-1) according to the target merchant's consumption data to obtain a data related to all customers of the target merchant. The average customer lifetime value, and a standard deviation of the customer lifetime value related to all customers of the target merchant, (E-2) According to the target customer lifetime value, the average customer lifetime value and the customer lifetime value standard deviation, generate the discount message, the discount message includes a discount number, and the discount number is calculated by the following formula:

Among them, x _CLV is the lifetime value of the target customer, μ _CLV is the average lifetime value of the customer, and s _CLV is the standard deviation of the customer’s lifetime value; and (F) according to the target customer information, transmit the preferential message to the corresponding target customer information. customer side. The method for generating personalized marketing privileges according to claim 1, wherein, in step (C), the predicted consumption frequency of the lifetime value data is obtained according to a predicted consumption frequency model, and the predicted consumption frequency model is Parson allocation, The formula is as follows:

Among them, Y _i is the consumption times of customer i , λ _i is the average consumption times of customer i in a preset unit time, τ _i is the time point when customer i is lost, T _i is a specific time point, τ _i > T For customer i at Ti has not been lost _. The method for generating a personalized marketing preference according to claim 1, wherein, in step (C), the predicted consumption frequency of the lifetime value data is obtained according to a predicted consumption frequency model, and the predicted consumption frequency model is gamma allocation, The formula is as follows:

Among them, λ _i is the average consumption times of customer i within a preset unit time, Γ(.) is a gamma function, and γ and α are preset parameters. The method for generating a personalized marketing preference according to claim 1, wherein, in step (C), the predicted active time of the lifetime value data is obtained according to a predicted active time model, and the predicted active time model is index allocation, as follows Mode:

Among them, τ _i is the time point when customer i churns, and μ _i is the average churn rate of customer i within a preset unit time. The method for generating personalized marketing offers as claimed in claim 1, wherein, in step (C), the predicted active time of the lifetime value data is obtained according to a predicted active time model, and the predicted active time model is gamma allocation, The formula is as follows:

Among them, μ _i is the average churn rate of customer i within a preset unit time, Γ(.) is a gamma function, and s and β are preset parameters. The method for generating a personalized marketing preference according to claim 1, wherein, in step (C), the predicted consumption amount of the lifetime value data is obtained according to a predicted consumption amount model, and the predicted consumption amount model is gamma allocation, The formula is as follows:

Among them, m _i is the average consumption amount of customer i in a preset unit time, Γ(.) is the gamma function, p is the shape number of gamma distribution, and v is the scale parameter of gamma distribution. The method for generating a personalized marketing preference according to claim 1, wherein, in step (C), the predicted consumption amount of the lifetime value data is obtained according to a predicted consumption amount model, and the predicted consumption amount model is gamma allocation, The formula is as follows:

Among them, θ _i is the expected value of customer i in a preset unit time, Γ(.) is a gamma function, and q and r are preset parameters. The method for generating personalized marketing privileges according to claim 1, wherein, in step (C), the predicted consumption amount of the lifetime value data is obtained according to a predicted consumption amount model, and the predicted consumption amount model is inverse beta allocation , as follows:

Wherein, M _i is the average consumption amount of customer i in a preset unit time, p is the total consumption amount of customer i in the preset unit time, and q and r are preset parameters. A system for generating personalized marketing privileges, connecting a plurality of merchant terminals via a first communication network, comprising: a service server storing a look-up table of customer characteristic information to customer information, the look-up table includes a plurality of transactions related to the Customer characteristic information of the customer, and a plurality of customer information respectively corresponding to the corresponding customer characteristic information and corresponding to a plurality of customer terminals, and connecting the merchant terminals through the first communication network, after receiving a target from the merchant terminals After a target customer characteristic information on the merchant side and a target merchant identification information related to the target merchant side, obtain a target customer information corresponding to the target customer characteristic information according to the target customer characteristic information and the look-up table; a marketing discount generating device, Stores a plurality of consumption data and a plurality of threshold values corresponding to the corresponding merchants. Each consumption data includes a customer information, a consumption time, a consumption amount, and a merchant identification information, and communicates with the service server through a second communication network. connected to the server, receive the target customer information and the target merchant identification information from the service server, and obtain a number of items including the target customer information and the target merchant identification information from the consumption data according to the target customer information and the target merchant identification information. The target merchant identifies target consumption data of the information, and according to the target consumption data, obtains a lifetime value data including a predicted consumption frequency, a predicted active time, a target customer lifetime value, and a predicted consumption amount, and according to the The lifetime value data determines whether a discount message needs to be generated, obtains a pair of target thresholds corresponding to the target merchant from the thresholds according to the target merchant identification information, and determines whether the target customer's lifetime value is greater than the target threshold. When it is determined that the target customer's lifetime value is greater than the target threshold value, the discount message needs to be generated. When it is determined that the discount message needs to be generated, according to the lifetime value and the consumption data including the target merchant's identification information for multiple target merchants' consumption The data generates the preferential message, and the marketing preferential generating device first obtains an average customer lifetime value related to all customers of the target merchant and a customer lifetime value related to all customers of the target merchant according to the consumption data of the target merchant. standard deviation, and then generate the preferential message according to the target customer lifetime value, the average customer lifetime value and the standard deviation of the customer lifetime value, the preferential message includes a discount number, and the discount number is calculated by the following formula:

Wherein, x _CLV is the lifetime value of the target customer, μ _CLV is the average lifetime value of the customer, s _CLV is the standard deviation of the lifetime value of the customer; and a push server connected to the marketing privilege generating device via the second communication network , after receiving the target customer information and the preferential message from the marketing preferential generating device, transmit the preferential message to the client corresponding to the target customer information via the first communication network according to the target customer information. The personalized marketing preference generating system according to claim 9, wherein the marketing preference generating device obtains the predicted consumption frequency of the lifetime value data according to a predicted consumption frequency model, and the predicted consumption frequency model is Parson allocation, as follows: :

Among them, Y _i is the consumption times of customer i , λ _i is the average consumption times of customer i in a preset unit time, τ _i is the time point when customer i is lost, T _i is a specific time point, τ _i > T For customer i at Ti has not been lost _. The personalized marketing preference generating system according to claim 9, wherein the marketing preference generating device obtains the predicted consumption frequency of the lifetime value data according to a predicted consumption frequency model, and the predicted consumption frequency model is gamma distribution, as follows: :

Among them, λ _i is the average consumption times of customer i within a preset unit time, Γ(.) is a gamma function, and γ and α are preset parameters. The personalized marketing preference generating system according to claim 9, wherein the marketing preference generating device obtains the predicted active time of the lifetime value data according to a predicted active time model, and the predicted active time model is index allocation, as follows:

Among them, τ _i is the time point when customer i churns, and μ _i is the average churn rate of customer i within a preset unit time. The system for generating personalized marketing benefits according to claim 9, wherein the device for generating marketing benefits obtains the predicted active time of the lifetime value data according to a predicted active time model, and the predicted active time model is gamma distribution, as follows: :

Among them, μ _i is the average churn rate of customer i within a preset unit time, Γ(.) is a gamma function, and s and β are preset parameters. The personalized marketing discount generating system according to claim 9, wherein the marketing discount generating device obtains the predicted consumption amount of the lifetime value data according to a predicted consumption amount model, and the predicted consumption amount model is gamma allocation, as follows: :

Among them, m _i is the average consumption amount of customer i in a preset unit time, Γ(.) is the gamma function, p is the shape number of gamma distribution, and v is the scale parameter of gamma distribution. The personalized marketing discount generating system according to claim 9, wherein the marketing discount generating device obtains the predicted consumption amount of the lifetime value data according to a predicted consumption amount model, and the predicted consumption amount model is gamma allocation, as follows: :

Among them, θ _i is the expected value of customer i in a preset unit time, Γ(.) is a gamma function, and q and r are preset parameters. The personalized marketing discount generating system according to claim 9, wherein the marketing discount generating device obtains the predicted consumption amount of the lifetime value data according to a predicted consumption amount model, and the predicted consumption amount model is inverse beta allocation, as follows Mode:

Wherein, M _i is the average consumption amount of customer i in a preset unit time, p is the total consumption amount of customer i in the preset unit time, and q and r are preset parameters.

Images