CN111915156A - User value-based service push method, electronic device and storage medium - Google Patents

Abstract

The user value-based service push method, electronic device and storage medium provided by the present invention obtain user data within a preset time period, wherein the user data is used to represent the user's network data; the user is determined according to the user data. The performance index, wherein the performance index is used to characterize the business index and the user-related index; adopt a combination of subjective and objective methods to determine the comprehensive weight corresponding to the user's performance index; determine the user value according to the user's performance index and the comprehensive weight; Push business information corresponding to the user value to user terminals with different user values; that is, the present invention obtains user data of multiple dimensions, and determines multi-dimensional performance indicators according to the user data of multiple dimensions, and then according to the multi-dimensional user data. The user value is determined by the performance indicators, which realizes a comprehensive and accurate evaluation of user value, and effectively enhances the pertinence of pushing services to users with different values.

Description

User value-based service push method, electronic device and storage medium

technical field

The present invention relates to communication technology, in particular to a user value-based service push method, electronic device and storage medium.

Background technique

With the increasingly fierce competition in the telecommunications industry, how to maintain existing users, especially high-value existing users, has become an important issue related to the stability and long-term development of operators.

In the prior art, operators determine the value of each user according to the average revenue per user (Average Revenue Per User, ARPU), that is, the average revenue contribution per user, and then adopt different business strategies for users with different user values to improve user experience. , so as to achieve user retention.

However, the above method only considers the factor of the average income of users, resulting in an inaccurate and comprehensive evaluation of the user value, which in turn results in the less targeted services being pushed to users with different values.

SUMMARY OF THE INVENTION

In view of the above problems, the present invention provides a user value-based service push method, electronic device and storage medium.

In a first aspect, the present invention provides a service push method based on user value, comprising: acquiring user data within a preset time period, wherein the user data is used to represent network data of the user; determining according to the user data User performance indicators, wherein the performance indicators are used to represent business indicators and user-related indicators; a combination of subjective and objective methods is used to determine the comprehensive weight corresponding to the user's performance indicators; according to the user's performance indicators and the comprehensive weight, the user value is determined ; Push service information corresponding to the user value to user terminals with different user values.

In other optional implementation manners, the user data includes at least one of the following: user tariff data, user terminal data, call data, and Internet access data; the performance indicators include at least one of the following: profit indicators, service usage Indicators, loyalty indicators, consumption potential indicators, business potential indicators and user influence indicators.

In other optional implementation manners, the user tariff data includes income data and cost data; the user terminal data includes the region to which the user terminal belongs and the terminal type; the call data includes call duration and number of calls; the Internet access data Including the user's account opening time, Internet traffic, Internet access time, and the usage traffic and usage time of the application program installed on the user terminal; the determining the user's performance index according to the user data includes: according to the income data and cost data. Determine the profit index; determine the service usage index according to the average value of call duration, number of calls, Internet traffic and Internet access time; determine the loyalty index according to the user's account opening time; The consumption potential index is determined according to the traffic and the usage time; the service potential index is determined according to the call duration, the number of calls, the Internet traffic and the rate of change of the Internet time; the user influence index is determined according to the call duration and the number of calls between the user and other users.

In other optional implementation manners, the determining the consumption potential indicator according to the region to which the user terminal belongs, the terminal type, and the usage flow and usage duration of the application includes: according to the region to which the user terminal belongs and a first preset relationship comparison table Determine the regional consumption capacity, wherein the first preset relationship comparison table records the regional consumption capabilities corresponding to different regions; determine the terminal consumption capacity according to the terminal type and the second preset relationship comparison table, wherein the second preset relationship The relationship comparison table records the terminal consumption capacity corresponding to different terminal types; the application consumption capacity is determined according to the usage flow and usage duration of the application; the consumption potential indicator is determined according to the regional consumption capacity, terminal consumption capacity and application consumption capacity .

In other optional implementation manners, the determining the consumption capacity of the application according to the usage flow and usage duration of the application includes: obtaining the variation coefficient of the usage flow, usage duration and usage times of each application; The weight of the usage traffic, usage duration and usage times of the application; determine the score of each application according to the weight and the usage traffic, usage duration and usage times of each application; determine the user’s preferred application according to the score of each application program; determine the flow duration ratio corresponding to the user preference application program according to the third preset relationship comparison table, and determine the flow duration ratio as the consumption capacity of the application program; wherein, the third preset relationship comparison table records the following: The traffic duration ratio corresponding to different applications.

In other optional implementation manners, the determining the consumption potential indicator according to the regional consumption capacity, the terminal consumption capacity and the application consumption capacity includes: determining the regional consumption capacity and the terminal consumption capacity by adopting a combination of subjective and objective methods and the comprehensive weight of the consumption power of the application; the sum of the comprehensive weight of the regional consumption power, the terminal consumption power and the consumption power of the application program and the product of the regional consumption power, the terminal consumption power and the consumption power of the application program is determined as the consumption potential indicator .

In other optional implementation manners, the determining the user influence index according to the call duration and the number of calls between the user and other users includes: preprocessing the call data, and obtaining the preprocessed call data between the user and other users; The coefficient of variation weight method is used to calculate the call interaction heat between the user and other users; the node betweenness centrality value is calculated according to the graph mining algorithm, and the node betweenness centrality value is determined as the user influence index.

In other optional implementation manners, before determining the comprehensive weight corresponding to the performance index of the user by using the subjective and objective combination method, the method further includes: performing dimensionless processing on the user data.

In a second aspect, the present invention provides an electronic device, comprising: at least one processor and a memory; the memory stores computer-executable instructions; the at least one processor executes the computer-executable instructions stored in the memory, so that the at least one The processor performs the method as described in any of the preceding items.

In a third aspect, the present invention provides a readable storage medium, where computer-executable instructions are stored, and when a processor executes the computer-executable instructions, the method described in any preceding item is implemented.

The user value-based service push method, electronic device and storage medium provided by the present invention obtain user data within a preset time period, wherein the user data is used to represent the user's network data; the user is determined according to the user data. The performance index, wherein the performance index is used to characterize the business index and the user-related index; adopt a combination of subjective and objective methods to determine the comprehensive weight corresponding to the user's performance index; determine the user value according to the user's performance index and the comprehensive weight; Push business information corresponding to the user value to user terminals with different user values; that is, the present invention obtains user data of multiple dimensions, and determines multi-dimensional performance indicators according to the user data of multiple dimensions, and then according to the multi-dimensional user data. The user value is determined by the performance indicators, which realizes a comprehensive and accurate evaluation of user value, and effectively enhances the pertinence of pushing services to users with different values.

Description of drawings

1 is a schematic diagram of an application scenario on which the present invention is based;

2 is a schematic flowchart of a user value-based service push method provided by the present invention;

3 is a schematic diagram of the architecture of a user value evaluation system provided by the present invention;

4 is a schematic flowchart of another user value-based service push method provided by the present invention;

5 is a schematic diagram of a user call network provided by the present invention;

6 is a schematic structural diagram of a user value-based service push device provided by the present invention;

FIG. 7 is a schematic diagram of a hardware structure of an electronic device provided by the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the examples of the present invention clearer, the technical solutions in the examples of the present invention will be described clearly and completely below with reference to the accompanying drawings in the examples of the present invention.

With the increasingly fierce competition in the telecommunications industry, how to maintain existing users, especially high-value existing users, has become an important issue related to the stability and long-term development of operators.

In the prior art, telecom operators usually determine the value of each user according to the average revenue of users, that is, the average revenue contribution of each user, and then adopt different business strategies for users with different values to maintain high-value users.

However, the above method only takes into account the average income of users, resulting in an inaccurate and comprehensive evaluation of the user value, resulting in poor business targeting to users with different values.

In view of the above problems, the technical idea of the present invention is to realize comprehensive and accurate evaluation of user value by acquiring user data of multiple dimensions, thereby effectively enhancing the pertinence of pushing services to users with different values.

FIG. 1 is a schematic diagram of an application scenario provided by the present invention. As shown in FIG. 1 , an application scenario provided by the present invention includes several user terminals 1 and a server 2 whose user value is to be evaluated, wherein the server 2 can be used to obtain user User data of the terminal 1, and then execute the service push method of the following embodiments.

In a first aspect, an example of the present invention provides a user value-based service push method, and FIG. 2 is a schematic flowchart of a user value-based service push method provided by the present invention.

As shown in Figure 2, the user value-based service push method includes:

Step 101: Acquire user data within a preset time period.

Wherein, the user data is used to represent the network data of the user.

Optionally, the user data includes at least one of the following: user tariff data, user terminal data, call data and Internet access data.

In this embodiment, the preset time period may be a relatively long time period, such as 6 months, 8 months, 12 months, etc., which can be determined by those skilled in the art based on experience, which is not limited in the present invention. In addition, the acquired user data may be in multiple dimensions, such as user tariff data, user terminal data, call data, Internet access data, etc., wherein the user tariff data may include the monthly billing income of the user and the average data of various costs , for example, the inter-network settlement expenses, user arrears, user terminal subsidies, user discounts, electronic coupons and rights and interests, etc. allocated to each user; user terminal data may include the area to which the user terminal belongs, terminal type, etc.; call data may include Data such as the user's monthly call duration and number of calls; Internet access data may include the user's account opening time, the user's monthly Internet access time, Internet traffic, and Internet access times.

Step 102: Determine the performance index of the user according to the user data.

The performance indicators are used to represent service indicators and user-related indicators.

Optionally, the performance indicators include at least one of the following: profit indicators, business usage indicators, loyalty indicators, consumption potential indicators, business potential indicators, and user influence indicators.

Specifically, user data may be processed according to a preset user value evaluation system to obtain user performance indicators. FIG. 3 is a schematic diagram of the architecture of a user value evaluation system provided by the present invention. As shown in FIG. 3 , the user value system includes a three-layer architecture of a target layer, a performance indicator layer, and a factor indicator layer. Among them, the target layer of the first layer There is only one, that is, the evaluated user value result; the performance indicator layer of the second layer includes six performance indicators, namely the above-mentioned profit indicator, business usage indicator, loyalty indicator, consumption potential indicator, business potential indicator and user influence Power indicators, among which, profit indicators and business use indicators reflect the current value of users, and loyalty indicators, consumption potential indicators, business potential indicators and user influence indicators reflect the future value of users; and under each performance indicator, there are several A factor index affects the performance index of the second layer, and then affects the index of the first layer. The factor index here is the user data of multiple dimensions obtained in step 101, or the user data after simple processing. The data.

Continuing to refer to Figure 3, the second-level performance indicator "profit" includes two third-level factor indicators of "revenue" and "cost". Among them, "income" reflects the actual consumption level of users, which is measured by the user's average monthly income (ARPU); "cost" reflects the cost of user access to the network and enjoys the operator's investment, which is the user entry cost, infrastructure cost, and marketing cost allocated by cost. , business and management costs, settlement costs, depreciation, arrears, etc. Generally speaking, the larger the value of the "cost" factor indicator, the lower the user's current value.

The second-level performance indicator "service usage" reflects the actual usage of users' voice and traffic, including four third-level factor indicators of "call duration", "number of calls", "internet duration", and "internet traffic". Generally speaking, the higher the usage of voice or traffic, the higher the user's current value.

The second-tier performance indicator "loyalty" is the user's continuous purchase behavior of the operator's services based on the operator's recognition, experience satisfaction and emotional integration, including the third-tier factor of "user account opening time" The indicator, "user's account opening time" can reflect the indirect calculation and evaluation of the user's value in the future. The longer the user's account opening time, the higher the user's loyalty in the following period of time, and its value has an upward trend, and vice versa. is a downward trend.

The second-level performance indicator "consumption potential" reflects the ability of users to bring benefits to enterprises in the future time period. The greater the consumption potential of users, the higher their value. "Consumption potential" is measured by three third-level factor indicators: "regional consumption capacity", "terminal consumption capacity", and "application (Application, App) consumption capacity". Among them, "regional consumption capacity" reflects the consumption capacity of the user's region. Generally speaking, the consumption capacity of users in different regions is different. For example, the online consumption capacity of users in general cities is lower than that of users in large and medium cities; "Consumption ability" reflects the consumption ability of users who hold different types of terminals. For example, users who hold Apple terminals have strong consumption ability, and the consumption potential is large; "application consumption ability" reflects the users brought by different applications. Value, for example, some software apps will frequently connect to the network during use, occupying network resources for a long time, but they generate less traffic, then the software app is a low-value application for operators, while other software brings high user stickiness and high traffic consumption, which can bring great value to operators and belong to high-value applications.

The second-layer performance indicator "service potential" reflects the potential of users to use voice, traffic and other services in the future time period, including "call duration growth rate", "call times growth rate", "Internet time growth rate", "Internet traffic growth rate" 8 third-layer factor indicators: “Volatility Rate”, “Volatility Rate of Call Duration”, “Volatility Rate of Number of Calls”, “Volatility Rate of Internet Access Duration”, and “Volatility Rate of Internet Traffic”. Generally speaking, the greater the growth rate and the greater the volatility, the greater the user's business potential and the higher the user value.

The second-layer performance indicator "influence" reflects the influence and importance of the user in the communication network. The greater the influence of the user, the higher the importance to the operator, and the higher the value of the user. Optionally, "influence" is measured by graph pattern analysis, that is, using "node betweenness centrality" (third-level factor indicator) in graph theory.

Step 103: Determine the comprehensive weight corresponding to the user's performance index by adopting a combination of subjective and objective methods.

Specifically, set-valued method, analytic hierarchy process, eigenvalue method, ordinal relationship analysis method, etc. can be used to subjectively weight each performance index; Carry out objective weighting; use integration methods such as addition, multiplication, squared deviation method, and game theory method to carry out integrated calculation of subjective and objective weight coefficients, and finally determine the comprehensive weight corresponding to each performance index. In this step, subjective and objective influencing factors are taken into consideration, and an objective and accurate evaluation of user value is realized.

For example, first, the set-valued method is used to determine the subjective weight coefficient. Suppose there are n indicators in a set of indicators, denoted as {x ₁ , x ₂ ,...x _n }, and invite multiple (take 20 as an example) experts to select the most important m from the indicator set (m< n) indicators, forming 20 indicator sets, which are recorded as:

Expert 1: X ¹ = {m indicators selected by the 1st expert},

Expert 2: X ² = {m indicators selected by the second expert},

...

Expert 20: X ²⁰ = {m indicators selected by the 20th expert},

Then in these 20 index sets, the number of times that the index x _j (j∈(1,2,…n)) is selected is:

g(x _j )=Num(x _j ∈X ^k ,k∈(1,2,…,20)), then the subjective weight coefficient of the index x _j (j∈(1,2,…n)) can refer to the formula (1), as follows:

Then the objective weight coefficient is determined according to the entropy method. First, calculate the proportion of the jth index under the ith sample (assuming there are n' samples) according to formula (2), for example, take the obtained indexes corresponding to different users as different samples.

Among them, 0≤y _ij ≤1, and X _ij is the index value of the jth item of the ith sample.

Then calculate the information entropy of the jth index according to formula (3), as follows:

where k is Boltzmann's constant.

Then calculate the difference coefficient of the jth index according to formula (4), as follows:

d _j =1-e _j (4)

Then, the objective weight of each indicator is calculated according to formula (5), as follows:

Among them, 0≤W _j ≤1, m is the number of indicators.

最优的权重向量可归结为对公式中的m个线性组合系数α_l进行优化，优化目标是使W与各个W_l ^T的离差极小化。那么解决这个问题的模型可表示为：

根据矩阵的微分性质,可得出公式最优化的一阶导数条件为为：

公式对应的线性方程组为：Finally, the game theory method is used to determine the comprehensive weight coefficient. The details are as follows: write down the linear combination of m weight vectors W _l ^T as:

The optimal weight vector can be attributed to the optimization of m linear combination coefficients α _l in the formula, and the optimization goal is to minimize the dispersion between W and each W _l ^T. Then the model to solve this problem can be expressed as:

According to the differential properties of the matrix, it can be concluded that the first-order derivative conditions for the optimization of the formula are:

The system of linear equations corresponding to the formula is:

α_l＞0，即可求出综合权重向量W；W_l ^T表示第l项权重向量，W_q ^T表示各个W_l ^T中的第q个W_l ^T。To find the solution of the above equation α _l , substitute in

If α _l >0, the comprehensive weight vector W can be obtained; W _l ^T represents the l-th weight vector, and W _q ^T represents the q-th W _l ^T in each W _l ^T .

Step 104: Determine the user value according to the user's performance index and the comprehensive weight.

Specifically, the user value is determined by multiplying and adding each performance index and the comprehensive weight corresponding to each performance index.

Step 105: Push service information corresponding to the user value to user terminals with different user values.

Specifically, for users with different values, different services are pushed to the terminal, so as to achieve targeted maintenance of users.

As an optional implementation manner, before step 102, the method further includes: performing dimensionless processing on the user data.

Specifically, because the acquired user data includes multiple dimensions, and the user data of each dimension has different dimensions, in order to eliminate the influence of dimensions, it is necessary to perform dimensionless processing on the user data, also known as standardization processing. . Among them, the factor indicators that have a negative impact on user value, such as cost factor indicators, can be standardized by formula (6); the factor indicators that have a positive impact on user value can be standardized by formula (7).

Among them, X _ij is the index value of the jth item of the ith sample, X′ _ij is the index value after standardization, X _max is the maximum value of the jth index, and X _min is the minimum value of the jth index. Then, the normalized X' _ij is substituted into the formula (1) as X _ij for subsequent processing.

The user value-based service push method provided by the embodiment of the present invention obtains user data within a preset time period, wherein the user data is used to represent the user's network data; the user's performance index is determined according to the user data, The performance indicators are used to represent business indicators and user-related indicators; a combination of subjective and objective methods is used to determine the comprehensive weight corresponding to the user's performance indicators; the user value is determined according to the user's performance indicators and the comprehensive weight; The corresponding user terminal pushes the business information corresponding to the user value; that is, the present invention obtains user data of multiple dimensions, determines multi-dimensional performance indicators according to the user data of multiple dimensions, and then determines the multi-dimensional performance indicators according to the multi-dimensional performance indicators. User value, realizes a comprehensive and accurate evaluation of user value, and effectively enhances the pertinence of pushing services to users of different values.

In conjunction with the foregoing embodiments, FIG. 4 is a schematic flowchart of another user value-based service push method provided by the present invention. As shown in FIG. 4 , the user value-based service push method includes:

Step 201: Obtain user data within a preset time period.

Wherein, the user tariff data includes income data and cost data; the user terminal data includes the region to which the user terminal belongs and the terminal type; the call data includes call duration and number of calls; And the usage traffic and usage time of the applications installed on the user terminal.

Step 202: Determine a profit index according to the revenue data and cost data.

In this step, the profit indicator can be obtained by obtaining the user's monthly revenue data and cost data, then calculating the monthly average revenue and monthly average cost, and then determining the profit indicator value according to the difference between the monthly average revenue and the monthly average cost.

Step 203: Determine the service usage indicator according to the average of the call duration, the number of calls, the Internet traffic and the Internet access time.

In this step, the service usage indicators can be obtained by obtaining the user's call duration, number of calls, Internet traffic and Internet time in each month, and then calculating the monthly average call time, monthly average number of calls, monthly average Internet traffic, and monthly average Internet time. According to the combination of subjective and objective methods, determine the respective comprehensive weights of call duration, number of calls, Internet traffic, and Internet access time. and, determined as the business usage indicator value.

Step 204: Determine the loyalty index according to the user's account opening duration.

In this step, for the loyalty index, the acquired user account opening duration of the user is determined as the user's loyalty index value.

Step 205: Determine the consumption potential indicator according to the region to which the user terminal belongs, the terminal type, and the usage flow and usage duration of the application.

In this step, the consumption potential index may be determined according to the region to which the user terminal belongs, the type of the terminal, and the usage flow and usage duration of the application. Optionally, the determining the consumption potential indicator according to the region to which the user terminal belongs, the terminal type, and the usage flow and usage duration of the application program includes: determining the regional consumption capacity according to the region to which the user terminal belongs and a first preset relationship comparison table, Wherein, the first preset relationship comparison table records the regional consumption capacity corresponding to different regions; the terminal consumption capacity is determined according to the terminal type and the second preset relationship comparison table, wherein the second preset relationship comparison table records the The terminal consumption capacity corresponding to different terminal types; the application consumption capacity is determined according to the usage flow and usage duration of the application; the consumption potential indicator is determined according to the regional consumption capacity, terminal consumption capacity and application consumption capacity.

Specifically, an expert scoring method can be used to obtain the first preset relationship comparison table, that is, experts evaluate and score the economic development levels of different regions, analyze and sort out the scoring results, and generate a regional consumption capacity scoring table, that is, the above-mentioned The first preset relationship comparison table; in this embodiment, after the region to which the user terminal belongs is obtained, the first preset relationship comparison table can be compared to determine the regional consumption capacity value of the region to which the user terminal belongs. Different evaluation values can also be given to different types of terminals according to the statistics of the average traffic of users corresponding to different types of terminals, that is, the above-mentioned second preset relationship comparison table; in this embodiment, after the user terminal type is obtained, By comparing the second preset relationship comparison table, the terminal consumption capacity value corresponding to the user terminal type is determined. It is also possible to obtain the traffic duration ratio corresponding to the application that the user prefers to use. The traffic duration ratio refers to the ratio of the traffic used by the application to the usage duration. The higher the traffic duration ratio, the greater the traffic consumed by the application per unit time. , which can bring more profits to operators, so users who prefer such applications are more valuable, and vice versa.

Optionally, determining the consumption capacity of the application according to the usage flow and usage duration of the application includes: obtaining the variation coefficient of the usage flow, usage duration, and usage times of each application; determining the usage flow of each application according to the variation coefficient , the weight of the usage time and the number of times of use; determine the score of each application according to the weight and the usage flow, usage time and number of times of each application; determine the user's preferred application according to the score of each application; Three preset relationship comparison tables determine the traffic duration ratio corresponding to the user's preferred application program, and determine the traffic duration ratio as the consumption capability of the application program; wherein, the third preset relationship comparison table records the correspondence of different application programs traffic time ratio.

Specifically, all kinds of apps include three attributes: duration of use, number of times of use, and flow of use. Because the dimensions of these three attributes are different, in order to eliminate the influence of different dimensions, the coefficient of variation of each attribute can be used to measure the value of each attribute. The degree of difference in values, as shown in Equation (8):

是用户j第i个属性的平均值。where v _j is the coefficient of variation of the j-th attribute of user v, σ _j is the standard deviation of the j-th attribute of user v,

is the average of the i-th attribute of user j.

Then calculate the weight of each attribute according to formula (9):

是用户v所有属性的变异系数之和。In the formula, W _j is the weight of the jth attribute of user v,

is the sum of the coefficients of variation of all attributes of user v.

Then calculate the comprehensive score of each App according to formula (10):

Then, sort the comprehensive scores of each app from high to bottom, select the app with the highest ranking as the user's preferred app, and then determine the traffic duration ratio corresponding to the user's preferred app according to the user's preference app and the third preset relationship comparison table, where , the third preset relationship comparison table records the traffic duration ratios corresponding to different application programs.

Optionally, the method for obtaining the third preset relationship comparison table is as follows: grouping and summarizing the user's Internet access data by application (APP), counting the total traffic and total duration used by different APPs, and using the formula: application (APP) traffic Duration ratio = total duration of application (APP) traffic ÷ total duration of application (APP) Calculate the traffic duration ratio of each application (APP), and obtain the "application (APP) traffic duration ratio" code table, that is, the third preset relationship comparison table .

Finally, the consumption potential index is determined according to the acquired regional consumption ability, terminal consumption ability and application consumption ability. Optionally, the consumption potential is determined according to the regional consumption ability, terminal consumption ability and application consumption ability. Indicators, including: adopting a combination of subjective and objective methods to determine the comprehensive weight of the regional consumption capacity, terminal consumption capacity and application consumption capacity; The sum of the product of the consumption ability and the consumption ability of the application is determined as the consumption potential indicator. It should be noted that the method of combining subjective and objective is similar to that of the above-mentioned embodiment, and details are not repeated here.

Step 206: Determine the service potential indicator according to the call duration, the number of calls, the Internet traffic, and the rate of change of the Internet access time.

In this step, the service potential index may be determined according to the user's call duration, number of calls, Internet traffic, and the rate of change of the Internet time, where the rate of change may include a growth rate and a volatility rate. That is, the indicators that affect the business potential include "call duration growth rate", "call number growth rate", "Internet time growth rate", "Internet traffic growth rate", "call duration fluctuation rate", "call number fluctuation rate", "Internet access rate". 8 third-layer factor indicators such as duration volatility” and “Internet traffic volatility”. Among them, the growth rate indicators can fit the trend line by regression analysis of the user's historical data, and use the slope of the trend line to represent the growth rate of the corresponding indicator. When the slope is greater than 0, it indicates that the future trend of the indicator is a growth trend. The greater the growth rate, the higher the user's business potential, and the higher the potential value; when the slope is less than 0, it means that the future trend of the indicator is a downward trend, the greater the absolute value of the slope, the smaller the growth rate, the lower the user's business potential, the lower the potential value lower. Volatility indicators reflect the volatility of the indicator by calculating the coefficient of variation of the user's corresponding attributes. The higher the volatility, the higher the user's consumption potential and the higher the potential value.

Taking user v's call duration growth rate index processing as an example, Table 1 shows the monthly call duration data of user v from July 2019 to December 2019. The account period can be taken as the independent variable x and the call duration as the dependent variable Y, then the univariate linear regression equation is: Y _t =ax _t +b, where x _t represents the value of the independent variable in the t period; Y _t represents the value of the dependent variable in the t period ; a and b represent the parameters of the univariate linear regression equation, where a represents the slope and b represents the intercept.

Then, the slope a can be obtained according to formula (11), and the slope a is the growth rate of the call duration.

Table 1: Monthly call duration of user v

Account period Call duration (seconds) 201907 23089 201908 22985 201909 24653 201910 25764 201911 26398 201912 26544

Similarly, the growth rate of the number of calls, the growth rate of the online duration, and the growth rate of the online traffic can be calculated, which will not be repeated here.

Continue to take Table 1 as an example, calculate the call duration fluctuation rate of user v, calculate the variation coefficient of user v call duration according to formula (12), and determine the variation coefficient as the call duration volatility rate.

C.V＝(SD÷MN)×100％(12)

Among them, C.V represents the coefficient of variation of the call duration of user v; SD represents the standard deviation of the monthly call duration of user v; MN represents the average value of the monthly call duration of user v.

Similarly, the fluctuation rate of the number of calls of the user, the fluctuation rate of the duration of Internet access, and the fluctuation rate of Internet traffic can be calculated.

Finally, a comprehensive weighting method combining subjective and objective can be used to determine the "growth rate of call duration", "growth rate of number of calls", "growth rate of Internet access time", "growth rate of Internet traffic", "call duration fluctuation rate", "call time growth rate" The comprehensive weight of the eight third-layer factor indicators, including frequency volatility, Internet access duration volatility, and Internet traffic volatility, and the sum of their products is determined as the user's business potential indicator.

Step 207: Determine the user influence index according to the call duration and the number of calls between the user and other users.

Optionally, step 207 includes: preprocessing the call data, and obtaining the preprocessed call data between the user and other users; calculating the call interaction heat between the user and other users by using the coefficient of variation weighting method; calculating the node interface according to the graph mining algorithm. The number centrality value is determined, and the node betweenness centrality value is determined as the user influence index.

Specifically, the acquired original call data is firstly preprocessed to eliminate erroneous data, invalid data, etc., wherein the preprocessing rules include at least one of the following: eliminating call records whose peer number length does not meet the normal number of digits; Or remove call records whose peer number is a foreign number; remove call records whose peer number is a public number, business number, company switchboard (such as: 10086, 10010, 95588, 110, etc.); remove the call duration less than the specified duration (such as 3s) ) invalid call records; reject invalid call records whose number of calls is less than the specified number (for example, 3 times). The preprocessed call data between the user and other users is shown in Table 2.

Table 2: Preprocessed call data

User code Peer User Code number of calls call duration 1818121828677570 1818122128912190 74 5468 1815062534402560 1819082356396900 7 281 1815120701709010 1819022835212630 twenty two 3348 1816072224094570 1816122735554700 20 2432 1816070823165820 1817083148181250 98 4885 1817100496288730 1817111501333180 25 1050 1817080647324000 1819050740535940 14 786 1817032340996760 1819072351368350 7 189 1817032413802152 1819042852291920 8 601 1817032473056151 1817100249553430 53 16437

Then, the coefficient of variation weight method is used to calculate the call interaction heat between users. Specifically, the two attributes of call duration and number of calls in each pair of user call behavior data are comprehensively evaluated by the coefficient of variation weight method, and these two attributes are fitted. It is an attribute to measure the call interaction heat of each pair of users. The specific calculation process is similar to that of formulas (8), (9), and (10), that is, firstly, formula (8) is used to calculate the variation coefficients corresponding to the call duration and the number of calls, and then the calls are calculated according to formula (9). The weights of the two attributes of duration and number of calls are finally calculated according to formula (10), and the call interaction heat of each pair of users is calculated, as shown in Table 3.

Finally, a graph mining algorithm is used to calculate the node betweenness centrality value. Figure 5 is a schematic diagram of a user communication network provided by the present invention. As shown in Figure 5, the user communication network graph can be abstracted as a set V consisting of nodes, and edges The graph G=(V,E,W) composed of the set E of For a user, the edge between nodes represents the interaction relationship between users, the weight of the edge represents the call interaction between users, and then the degree of user influence is measured by the node betweenness centrality, where the node betweenness centrality refers to the interaction between users. Among all the shortest paths, the number of paths passing through the node accounts for the proportion of the total number of shortest paths, reflecting the role and influence of the node in the entire network.

Calculate the node betweenness centrality according to formula (13), as follows:

Among them, C _B (v) represents the betweenness centrality of node v; σ _st (v) represents the number of shortest paths from s to t through node v; σ _st represents the number of all shortest paths from node s to node t.

Table 3: User call interaction popularity data

User code Peer User Code Call interaction heat 1818121828677570 1818122128912190 0.0088 1815062534402560 1819082356396900 0.0004 1815120701709010 1819022835212630 0.0054 1816072224094570 1816122735554700 0.0039 1816070823165820 1817083148181250 0.0079 1817100496288730 1817111501333180 0.0017 1817080647324000 1819050740535940 0.0012 1817032340996760 1819072351368350 0.0002 1817032413802152 1819042852291920 0.0009 1817032473056151 1817100249553430 0.0265

Step 208: Determine the comprehensive weight corresponding to the user's performance index by adopting a combination of subjective and objective methods.

Step 209: Determine the user value according to the user's performance index and the comprehensive weight.

Step 210: Push service information corresponding to the user value to user terminals corresponding to different user values.

Steps 208 , 209 , and 210 in this embodiment are respectively similar to the implementations of steps 103 , 104 , and 105 in the foregoing embodiments, and will not be repeated here.

Different from the foregoing embodiments, the present embodiment further defines the manner of determining the user performance index. In this embodiment, the profit index is determined according to the income data and the cost data; the service usage index is determined according to the average value of the call duration, the number of calls, the total online traffic and the total online duration; the loyalty is determined according to the total online duration The consumption potential index is determined according to the region to which the user terminal belongs, the type of the terminal, and the Internet traffic and Internet access duration corresponding to the application; Potential indicators; user influence indicators are determined based on the duration and number of calls between users and other users.

For example, first select the users to be evaluated, such as extracting the existing users who are normally online in a certain province as of December 31, 2019 and are not newly developed in the month (excluding Internet cards, public beta, etc.); obtain the user data of the users to be evaluated, including " Attributes such as online time" and "area"; extract the user's billing income, call duration, number of calls, Internet access time, and Internet traffic data from July to December, respectively, and calculate the 6-month average of the corresponding indicators, if less than 6 months Take the monthly average online time; extract the average cost data of users from July to December (including: inter-network settlement expenses allocated to each user, user arrears, user terminal subsidies, user discounts, e-coupons and equity bonuses, etc. ); extract all the call details of the user in October, November and December; extract the terminal usage information of the user for 6 months; extract the usage time, usage times and usage traffic of various applications (APPs) used by the user within 3 months data.

After obtaining the user data of the dimension, the comprehensive weighting coefficient of the third layer and the second layer in the preset user value evaluation system is determined according to the comprehensive weighting method combining subjective and objective. Table 5 shows the performance of the second layer provided by the present invention. The indicator “consumption potential” (referred to as y4) and the third-level factor indicators “regional consumption capacity”, “terminal consumption capacity”, and “application consumption capacity” (respectively denoted as x1, x2, x3) are taken as examples, according to the above comprehensive integration Using the weight coefficient determined by the weighting method, the user value of the target layer is finally calculated, and service information is pushed to user terminals with different user values.

Table 5: Indicator Weight Coefficients

index Subjective Weight Coefficient objective weight factor Comprehensive integrated weight coefficient Regional spending power x1 0.32 0.45 0.38 Terminal spending power x2 0.48 0.4 0.44 App spending power x3 0.2 0.15 0.18 Consumption potential y4 0.12 0.14 0.13

The user value-based service push method provided by the embodiment of the present invention obtains user data within a preset time period; determines a profit index according to the income data and cost data; The service usage index is determined by the average value of the duration; the loyalty index is determined according to the total online duration; the consumption potential index is determined according to the area to which the user terminal belongs, the terminal type, and the internet traffic and internet duration corresponding to the application; The service potential index is determined according to the duration, the number of calls, the total Internet traffic, and the rate of change of the total Internet access time; the user influence index is determined according to the call time and the number of calls between the user and other users; the subjective and objective combination method is used to determine the corresponding performance index of the user According to the user's performance index and the comprehensive weight, the user value is determined; the business information corresponding to the user value is pushed to the user terminals corresponding to different user values; And determine multi-dimensional performance indicators based on multi-dimensional user data, and then determine user value according to multi-dimensional performance indicators, realize a comprehensive and accurate evaluation of user value, and effectively enhance the pertinence of pushing services to users with different values.

In a second aspect, an example of the present invention provides a user value-based service push device. FIG. 6 is a schematic structural diagram of a user value-based service push device provided by the present invention. As shown in FIG. 6 , the service push device includes: :

The data acquisition module 10 is used to acquire user data within a preset time period, wherein the user data is used to represent the network data of the user; the first processing module 20 is used to determine the performance index of the user according to the user data, The performance indicators are used to represent service indicators and user-related indicators; the second processing module 30 is used to determine the comprehensive weight corresponding to the user's performance indicators by adopting a combination of subjective and objective methods; the third processing module 40 is used to The performance index and the comprehensive weight determine the user value; the service push module 50 is configured to push service information corresponding to the user value to user terminals with different user values.

In other optional implementation manners, the user data includes at least one of the following: user tariff data, user terminal data, call data, and Internet access data; the performance indicators include at least one of the following: profit indicators, service usage Indicators, loyalty indicators, consumption potential indicators, business potential indicators and user influence indicators.

In other optional implementation manners, the user tariff data includes income data and cost data; the user terminal data includes the region to which the user terminal belongs and the terminal type; the call data includes call duration and number of calls; the Internet access data Including the user's account opening time, Internet traffic, Internet access time, and the usage traffic and usage time of the application installed on the user terminal; the first processing module 20 is specifically used for: determining the profit index according to the income data and cost data. ; Determine the service usage index according to the average value of call duration, number of calls, Internet traffic and Internet access time; Determine the loyalty index according to the user's account opening time; The consumption potential index is determined according to the duration; the service potential index is determined according to the call duration, the number of calls, the Internet traffic and the rate of change of the online duration; the user influence index is determined according to the call duration and the number of calls between the user and other users.

In other optional implementation manners, the first processing module 20 is specifically configured to: determine the regional consumption capacity according to the region to which the user terminal belongs and a first preset relationship comparison table, wherein the first preset relationship comparison table The regional consumption capabilities corresponding to different regions are recorded; the terminal consumption capabilities are determined according to the terminal type and the second preset relationship comparison table, wherein the second preset relationship comparison table records the terminal consumption capabilities corresponding to different terminal types; according to the application The usage flow and usage duration of the program determine the consumption capacity of the application; the consumption potential indicator is determined according to the regional consumption capacity, the terminal consumption capacity and the consumption capacity of the application.

In other optional embodiments, the first processing module 20 is specifically configured to: obtain the coefficient of variation of the usage flow, usage duration and usage times of each application; determine the usage flow, usage duration and The weight of the number of times of use; determine the score of each application program according to the weight and the usage flow, duration of use and the number of times of use of each application program; determine the user's preferred application program according to the score of each application program; compare according to the third preset relationship The table determines the traffic duration ratio corresponding to the user's preferred application, and determines the traffic duration ratio as the consumption capability of the application; wherein, the third preset relationship comparison table records the traffic duration ratio corresponding to different applications.

In other optional embodiments, the first processing module 20 is specifically configured to: determine the comprehensive weight of the regional consumption capacity, terminal consumption capacity and application consumption capacity by adopting a combination of subjective and objective methods; The sum of the product of the comprehensive weight of the ability and the consumption ability of the application and the product of the regional consumption ability, the terminal consumption ability and the consumption ability of the application is determined as the consumption potential indicator.

In other optional embodiments, the first processing module 20 is specifically configured to: preprocess the call data, and obtain the preprocessed call data between the user and other users; Call interaction heat; calculate the node betweenness centrality value according to the graph mining algorithm, and determine the node betweenness centrality value as the user influence index.

In other optional implementation manners, the first processing module 20 is specifically configured to: perform dimensionless processing on the user data before the comprehensive weight corresponding to the user's performance index is determined by the subjective and objective combination method.

Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the service push device described above and the corresponding beneficial effects can refer to the corresponding process in the foregoing method example, which will not be repeated here.

The user value-based service push device provided by the embodiment of the present invention uses a data acquisition module to acquire user data within a preset time period, wherein the user data is used to represent the user's network data; the first processing module, is used to determine the performance index of the user according to the user data, wherein the performance index is used to represent the service index and user-related index; the second processing module is used to determine the comprehensive weight corresponding to the user's performance index by adopting a combination of subjective and objective methods The third processing module is used to determine the user value according to the user's performance index and the comprehensive weight; the service push module is used to push the service information corresponding to the user value to the user terminals corresponding to different user values; namely The present invention realizes comprehensive and accurate evaluation of user value by obtaining user data of various dimensions, determining multi-dimensional performance indicators according to the user data of various dimensions, and then determining user value according to the multi-dimensional performance indicators, thereby effectively enhancing the user value. Push business targeting to users with different values.

In a third aspect, an example of the present invention provides an electronic device, and FIG. 7 is a schematic diagram of the hardware structure of an electronic device provided by the present invention, as shown in FIG. 7 , including:

At least one processor 701 and memory 702.

In a specific implementation process, at least one processor 701 executes the computer-executable instructions stored in the memory 702 , so that the at least one processor 701 executes the above user value-based service push method, wherein the processor 701 and the memory 702 pass through the bus 703 connect.

For the specific implementation process of the processor 701, reference may be made to the foregoing method embodiments, and the implementation principles and technical effects thereof are similar, and details are not described herein again in this embodiment.

In the above-mentioned embodiment shown in FIG. 7 , it should be understood that the processor may be a central processing unit (English: Central Processing Unit, CPU for short), or other general-purpose processors, digital signal processors (English: Digital Signal Processor) Signal Processor, referred to as DSP), application specific integrated circuit (English: Application Specific Integrated Circuit, referred to as: ASIC) and the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in conjunction with the invention can be directly embodied as executed by a hardware processor, or executed by a combination of hardware and software modules in the processor.

The memory may include high-speed RAM memory, and may also include non-volatile storage NVM, such as at least one disk memory.

The bus may be an Industry Standard Architecture (Industry Standard Architecture, ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (Extended Industry Standard Architecture, EISA) bus, or the like. The bus can be divided into address bus, data bus, control bus and so on. For convenience of representation, the buses in the drawings of the present application are not limited to only one bus or one type of bus.

In a fourth aspect, the present invention also provides a readable storage medium, where computer-executable instructions are stored in the readable storage medium, and when the processor executes the computer-executable instructions, the above user value-based service push method is implemented.

The above-mentioned readable storage medium may be implemented by any type of volatile or non-volatile storage device or combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM) , Erasable Programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), Magnetic Memory, Flash Memory, Magnetic Disk or Optical Disk. A readable storage medium can be any available medium that can be accessed by a general purpose or special purpose computer.

An exemplary readable storage medium is coupled to the processor such that the processor can read information from, and write information to, the readable storage medium. Of course, the readable storage medium can also be an integral part of the processor. The processor and the readable storage medium may be located in application specific integrated circuits (Application Specific Integrated Circuits, ASIC for short). Of course, the processor and the readable storage medium may also exist in the device as discrete components.

Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments may be completed by program instructions related to hardware. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the steps including the above method embodiments are executed; and the foregoing storage medium includes: ROM, RAM, magnetic disk or optical disk and other media that can store program codes.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, but not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: The technical solutions described in the foregoing embodiments can still be modified, or some or all of the technical features thereof can be equivalently replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the technical solutions of the embodiments of the present invention. scope.

Claims (10)

1. a service push method based on user value, is characterized in that, comprises: acquiring user data within a preset time period, wherein the user data is used to characterize the user's network data; Determine user performance indicators according to the user data, wherein the performance indicators are used to represent service indicators and user-related indicators; The combination of subjective and objective methods is used to determine the comprehensive weights corresponding to the user's performance indicators; Determine the user value according to the user's performance index and the comprehensive weight; The service information corresponding to the user value is pushed to user terminals with different user values. 2. The method according to claim 1, wherein the user data comprises at least one of the following: user tariff data, user terminal data, call data and Internet access data; The performance indicators include at least one of the following: profit indicators, business usage indicators, loyalty indicators, consumption potential indicators, business potential indicators, and user influence indicators. 3. The method according to claim 2, wherein the user tariff data includes revenue data and cost data; the user terminal data includes a region to which the user terminal belongs and a terminal type; the call data includes call duration and call time The number of times of Internet access; the Internet access data includes the user account opening time, Internet access traffic, Internet access time, and the usage traffic and usage time of the application program installed on the user terminal; the determination of the user's performance indicators according to the user data includes: Determine profit indicators based on the revenue data and cost data; Determine the service usage indicator according to the average value of call duration, number of calls, Internet traffic and Internet access time; Determine the loyalty index according to the user's account opening time; Determine the consumption potential indicator according to the region to which the user terminal belongs, the type of the terminal, and the usage flow and usage duration of the application; Determine the business potential indicator according to the call duration, the number of calls, the Internet traffic and the rate of change of the Internet access time; Determine the user influence indicator based on the duration and number of calls between the user and other users. 4. The method according to claim 3, wherein determining the consumption potential index according to the area to which the user terminal belongs, the terminal type, and the usage flow and usage duration of the application, comprising: Determine the regional consumption capacity according to the region to which the user terminal belongs and a first preset relationship comparison table, wherein the first preset relationship comparison table records the regional consumption capabilities corresponding to different regions; Determine the terminal consumption capability according to the terminal type and a second preset relationship comparison table, wherein the second preset relationship comparison table records the terminal consumption capabilities corresponding to different terminal types; Determine the consumption capacity of the application according to the usage traffic and usage time of the application; The consumption potential indicator is determined according to the regional consumption capacity, terminal consumption capacity and application consumption capacity. 5. The method according to claim 4, wherein the determining the consumption capability of the application according to the usage flow and the usage duration of the application comprises: Obtain the coefficient of variation of the usage flow, usage duration and usage times of each application; Determine the weight of the usage flow, usage duration and usage times of each application according to the coefficient of variation; Determine the score of each application according to the weight and the usage flow, usage duration and usage times of each application; Determine the user's preferred application according to the score of each application; Determine the traffic duration ratio corresponding to the user's preferred application program according to the third preset relationship comparison table, and determine the traffic duration ratio as the consumption capacity of the application program; wherein, the third preset relationship comparison table records different applications The traffic duration ratio corresponding to the program. 6. The method according to claim 4 or 5, wherein the determining the consumption potential indicator according to the regional consumption capacity, terminal consumption capacity and application consumption capacity comprises: Determine the comprehensive weight of the regional consumption capacity, terminal consumption capacity and application consumption capacity by adopting the combination of subjective and objective methods; Determine the consumption potential indicator as the sum of the comprehensive weight of the regional consumption capacity, the terminal consumption capacity and the application consumption capacity and the product of the regional consumption capacity, the terminal consumption capacity and the application consumption capacity. 7. The method according to any one of claims 3-5, wherein the determining the user influence index according to the call duration and the number of calls between the user and other users, comprising: Preprocess the call data to obtain the preprocessed call data between the user and other users; The coefficient of variation weight method is used to calculate the call interaction between the user and other users; The node betweenness centrality value is calculated according to the graph mining algorithm, and the node betweenness centrality value is determined as the user influence index. 8. The method according to any one of claims 1-5, characterized in that before using the subjective and objective combination method to determine the comprehensive weight corresponding to the user's performance index, the method further comprises: Perform dimensionless processing on the user data. 9. An electronic device, comprising: at least one processor and a memory; the memory stores computer-executable instructions; The at least one processor executes computer-implemented instructions stored in the memory, causing the at least one processor to perform the method of any one of claims 1-8. 10. A readable storage medium, characterized in that, computer-executable instructions are stored in the readable storage medium, and when a processor executes the computer-executable instructions, the computer-executable instructions according to any one of claims 1 to 8 are implemented. method.

Images