CN112651582B - User category identification method and related equipment for product loss user - Google Patents
User category identification method and related equipment for product loss user Download PDFInfo
- Publication number
- CN112651582B CN112651582B CN201910959099.0A CN201910959099A CN112651582B CN 112651582 B CN112651582 B CN 112651582B CN 201910959099 A CN201910959099 A CN 201910959099A CN 112651582 B CN112651582 B CN 112651582B
- Authority
- CN
- China
- Prior art keywords
- user
- reflux
- value
- product loss
- level
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000010992 reflux Methods 0.000 claims abstract description 153
- 238000013210 evaluation model Methods 0.000 claims abstract description 39
- 238000012545 processing Methods 0.000 claims abstract description 22
- 238000007477 logistic regression Methods 0.000 claims description 20
- 238000004590 computer program Methods 0.000 claims description 9
- 238000005457 optimization Methods 0.000 claims description 6
- 238000007405 data analysis Methods 0.000 abstract description 5
- 238000005516 engineering process Methods 0.000 abstract description 5
- 239000002699 waste material Substances 0.000 abstract description 5
- 238000004891 communication Methods 0.000 description 10
- 238000004458 analytical method Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000000611 regression analysis Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007621 cluster analysis Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000011022 operating instruction Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
Landscapes
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Theoretical Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Data Mining & Analysis (AREA)
- Educational Administration (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application discloses a user category identification method, a system, electronic equipment and a computer readable storage medium of a product loss user, wherein the method comprises the following steps: acquiring user data of a product loss user; determining the value level of the product loss user according to consumption data in the user data; substituting a target attribute field in the user data into a pre-established reflux intention evaluation model, and calculating a reflux intention probability value of a product loss user; determining a reflux intention level corresponding to the reflux intention probability value; and determining the user category corresponding to the reflux intention level and the value level so as to recall the product loss user according to the target recall mode corresponding to the user category. The application introduces a data analysis processing technology to process the user data so as to adopt a corresponding reasonable targeted recall scheme aiming at different types of product loss users, thereby avoiding the waste of manpower recall cost and effectively guaranteeing the recall rate and recall efficiency.
Description
Technical Field
The present application relates to the field of data processing technologies, and in particular, to a method, a system, an electronic device, and a computer readable storage medium for identifying a user category of a product loss user.
Background
Data is a carrier for recording a large amount of information, and the importance of the data is not doubtful in the current network age and the large data age. For data processing and analysis, there are significant roles in business applications in many enterprises.
User data is a type of important data in enterprise operations. The method has important guiding significance in combining specific business requirements of enterprises and carrying out targeted processing and analysis on the user data, and can play a great role in recovering lost users of products. However, most of the prior art recall lost users indifferently, and the scientific and reasonable data processing analysis is not combined. In view of limited recall force of non-manual recall modes such as short messages, in-station messages, mails and the like, the recall rate of the prior art adopting the modes in a unified way is obviously lower; on the other hand, the prior art adopting the manual voice recall mode uniformly has the problems of great waste of manual recall cost and low efficiency.
In view of this, it has been a great need for a person skilled in the art to provide a solution to the above-mentioned technical problems.
Disclosure of Invention
The application aims to provide a user category identification method, a system, electronic equipment and a computer readable storage medium for a product loss user, so that scientific and reasonable guiding basis is provided for recall of the product loss user, and recall rate and recall efficiency are improved, and recall cost is reduced.
In order to solve the above technical problems, in a first aspect, the present application discloses a method for identifying a user category of a product loss user, including:
acquiring user data of the product loss user;
determining the value level of the product loss user according to consumption data in the user data;
substituting a target attribute field in the user data into a pre-established reflux intention evaluation model, and calculating a reflux intention probability value of the product loss user;
Determining a reflux intention level corresponding to the reflux intention probability value;
And determining the user category corresponding to the reflux intention level and the value level so as to recall the product loss user according to a target recall mode corresponding to the user category.
In a second aspect, the present application also discloses a user category identification system for a product loss user, including:
The data acquisition module is used for acquiring the user data of the product loss user;
The first processing module is used for determining the value level of the product loss user according to the consumption data in the user data;
the model building module is used for pre-building a reflux intention evaluation model of the reflux intention probability about the target attribute field;
The second processing module is used for substituting the target attribute field in the user data into the reflux intention evaluation model and calculating the reflux intention probability value of the product loss user; determining a reflux intention level corresponding to the reflux intention probability value;
And the category identification module is used for determining the user category corresponding to the reflux intention level and the value level so as to recall the product loss user according to a target recall mode corresponding to the user category.
In a third aspect, the present application also discloses an electronic device, including:
a memory for storing a computer program;
A processor for executing the computer program to implement the steps of any of the user category identification methods for product churn users as described above.
In a fourth aspect, the present application also discloses a computer readable storage medium having stored therein a computer program which when executed by a processor is adapted to carry out the steps of a user category identification method for any product loss user as described above.
The user category identification method for the product loss user provided by the application comprises the following steps: acquiring user data of the product loss user; determining the value level of the product loss user according to consumption data in the user data; substituting a target attribute field in the user data into a pre-established reflux intention evaluation model, and calculating a reflux intention probability value of the product loss user; determining a reflux intention level corresponding to the reflux intention probability value; and determining the user category corresponding to the reflux intention level and the value level so as to recall the product loss user according to a target recall mode corresponding to the user category.
Therefore, the application introduces a data analysis processing technology to process the user data, determines the value level and the reflux intention level of the product loss user from the data level, and further can identify the user category of the product loss user and serve as the guiding basis of user recall so as to adopt a corresponding reasonable targeted recall scheme for the product loss users of different categories. The application not only avoids the waste of manpower recall cost, but also effectively ensures recall rate and recall efficiency. The user category identification system, the electronic equipment and the computer readable storage medium for the product loss user have the beneficial effects.
Drawings
In order to more clearly illustrate the technical solutions in the prior art and the embodiments of the present application, the following will briefly describe the drawings that need to be used in the description of the prior art and the embodiments of the present application. Of course, the following drawings related to embodiments of the present application are only a part of embodiments of the present application, and it will be obvious to those skilled in the art that other drawings can be obtained from the provided drawings without any inventive effort, and the obtained other drawings also fall within the scope of the present application.
Fig. 1 is a schematic diagram of a composition structure of an electronic device according to an embodiment of the present application;
fig. 2 is a flowchart of a method for identifying a user category of a product loss user according to an embodiment of the present application;
FIG. 3 is a diagram of a search interface for background user data of a game product according to an embodiment of the present application;
FIG. 4 is a detailed record page diagram of user data according to an embodiment of the present application;
fig. 5 is a flowchart of a method for identifying a user category of a product loss user according to an embodiment of the present application;
Fig. 6 is a flowchart of a method for identifying a user category of a product loss user according to an embodiment of the present application;
FIG. 7 is a diagram illustrating the structure of a variable of a reflow intent assessment model in an application scenario embodiment of the present disclosure;
Fig. 8 is a block diagram of a user category identification system for a product loss user according to an embodiment of the present application.
Detailed Description
The core of the application is to provide a user category identification method, a system, electronic equipment and a computer readable storage medium for a product loss user, so as to provide scientific and reasonable guiding basis for recall of the product loss user, and reduce recall cost while improving recall rate and recall efficiency.
In order to more clearly and completely describe the technical solutions in the embodiments of the present application, the technical solutions in the embodiments of the present application will be described below with reference to the accompanying drawings in the embodiments of the present application. It will be apparent that the described embodiments are only some, but not all, embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
User data is a type of important data in enterprise operations. The method has important guiding significance in combining specific business requirements of enterprises and carrying out targeted processing and analysis on the user data, and can play a great role in recovering lost users of products. However, in the prior art, the lost user is recalled indiscriminately, and scientific and reasonable data processing analysis is not combined, so that the efficiency is low, a large amount of manpower recall cost is wasted, or the recall condition is not ideal. In view of the above, the present application provides a method for identifying user categories of product loss users, which can effectively solve the above-mentioned problems.
For easy understanding, the following describes an electronic device to which the user category identification method of the product loss user of the present application is applicable, and in particular, reference may be made to fig. 1.
As can be seen in fig. 1, an electronic device 10 may include: a processor 11, a memory 12, a communication interface 13, an input unit 14 and a display 15 and a communication bus 16. The processor 11, the memory 12, the communication interface 13, the input unit 14, the display 15, all perform communication with each other via a communication bus 16.
In the embodiment of the present application, the processor 11 may be a central processing unit (Central Processing Unit, CPU), an asic, a dsp, an off-the-shelf programmable gate array, or other programmable logic device. The processor may call programs stored in the memory 12. Specifically, the processor may perform operations performed on the electronic device side in any of the user category identification method embodiments described below.
The memory 12 is used to store one or more programs, which may include program code including computer operating instructions, and in embodiments of the present application, at least one program for implementing a user category identification method described below is stored in the memory.
In one possible implementation, the memory 12 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, and applications required for at least one function (e.g., obtaining user data for a product churn user), etc.; the storage data area may store data created during use of the electronic device 10, such as value levels, reflux intent levels, and the like.
In addition, memory 12 may also include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device or other volatile solid-state storage device.
The communication interface 13 may be an interface of a communication module, such as an interface of a GSM module.
The application may also comprise a display 14 and an input unit 15 etc.
Of course, the structure of the electronic device shown in fig. 1 does not limit the electronic device in the embodiment of the present application, and the electronic device may include more or fewer components than those shown in fig. 1 or may combine some components in practical applications.
The electronic device 10 in fig. 1 may be a terminal (such as a PC) or a server with higher performance than a common terminal.
In the embodiment of the present application, the electronic device 10 may receive, according to the communication interface 13, user data of a product loss user sent by other external devices by using a network; user data of the product churn user may also be obtained through the own input unit 14 (e.g., keyboard, touch screen, voice input device, etc.).
Accordingly, the processor 11 in the electronic device 10 may receive the user data from the communication interface 13 or the input unit 14 via the communication bus 16, and invoke the program stored in the memory 12 to determine the value level of the product loss user according to the consumption data in the user data; substituting a target attribute field in the user data into a pre-established reflux intention evaluation model, and calculating a reflux intention probability value of a product loss user; determining a reflux intention level corresponding to the reflux intention probability value; and further, the user category corresponding to the reflux intention level and the value level is determined, so that the aim of providing scientific and reasonable basis for the recall product loss user is fulfilled.
Referring to fig. 2, the embodiment of the application discloses a user category identification method of a product loss user, which mainly comprises the following steps:
S11: user data of a product loss user is obtained.
The product may be embodied as various types of software application products, such as games, video players, and the like. And once the continuous unregistered days of the user exceed a certain value, the lost days can be regarded as the product lost user. For example, taking a game product as an example, after more than 7 consecutive unregistered days, it can be considered a product loss user.
The account is established by the user, and various information data and activities under the product account form user data to be recorded, including personal information data, friend data, product use records, consumption data, customer service records and the like.
Each type of user data may further include a plurality of attribute fields, for example, the personal information material reflects social attribute information of the user, and may include attribute fields such as age, gender, job position, and location of the user; for another example, the customer service record reflects customer service attribute information of the user, and specifically may include attribute fields such as number of customer service orders, customer service problem resolvability, etc.; the product usage record reflects account attribute information of the user, and specifically may include attribute fields such as account level, loss days, and the like.
For example, referring to fig. 3, fig. 3 is a diagram of a search interface of background user data of a game product according to an embodiment of the present application, and specifically shows that the search result is user data with an account ID of 12345. After the account is selected and clicked, the user data detailed page of the user can be further checked, and particularly, see fig. 4.
S12: and determining the value level of the product loss user according to the consumption data in the user data.
Specifically, in the prior art, indiscriminate recall is performed on each product loss user, so that the pertinence is poor, and the recall quality is seriously influenced. Therefore, the embodiment of the application carries out category identification on the product loss users based on the user data so as to respectively and pointedly and effectively recall various product loss users.
The consumption data of the user is important data of the user during the use of the product, visually reflects the value and consumption capability of the user, and also reflects the loving degree and the backflow intention of the user for the product to a certain extent. Therefore, the application carries out value rating on the product loss user according to the consumption data of the user. It is readily understood that the more vigorous the user's consumption, the higher the corresponding value level. Wherein the consumption data includes, but is not limited to, attribute fields such as total consumption, number of consumption, etc.
In a simple embodiment, the product loss user may be classified into two value levels by using only one preset value threshold, and step S12 may specifically include: if the total consumption amount is higher than a preset value threshold, determining the value level of the product loss user as a first value level; if the total consumption amount is not higher than the preset value threshold, determining the value level of the product loss user as a second value level.
Of course, those skilled in the art may set a plurality of value thresholds to distinguish the product loss user into a plurality of value levels, which the present application is not limited to further.
S13: substituting the target attribute field in the user data into a pre-established reflux intention evaluation model, and calculating the reflux intention probability value of the product loss user.
The reflux intention probability value is the probability value that the lost user is successfully recalled. It should be noted that, in order to scientifically and reasonably evaluate the reflux intention of the product loss user, the application establishes a reflux intention evaluation model in advance, and based on the model, the reflux intention probability value of the product loss user can be determined after substituting the relevant user data of the product loss user.
Specifically, the reflux intention evaluation model is a mathematical model of reflux intention with respect to multivariate, and specifically can be established based on regression analysis, although other methods such as cluster analysis and the like can be employed. The multivariate target attribute fields are some attribute fields related to the user reflow intention in the user data, and collectively reflect the social attribute information, account attribute information or customer service attribute information of the user. As to which specific attribute fields can be used as the target attribute fields, it can be set by those skilled in the art according to specific characteristics of the product and actual application conditions of the product.
S14: a reflux intent level corresponding to the reflux intent probability value is determined.
The reflux intention probability value is based on a reflux intention evaluation result obtained from the data layer by data analysis, objectively reflects the possibility that the product loss user is successfully recalled, and can provide scientific and reasonable guidance opinion for the user recall. Based on the reflux intention probability value, the application can carry out level distinction on the recall possibility of the user so as to carry out reasonable and targeted recall on different types of product loss users respectively.
In a simple embodiment, the product loss user may be divided into two reflux intention levels by using only one preset probability threshold, and step S14 may specifically include: if the reflux intention probability value is higher than a preset probability threshold value, determining the reflux intention level of the product loss user as a first reflux intention level; and if the reflux intention probability value is not higher than the preset probability threshold value, determining the reflux intention level of the product loss user as a second reflux intention level.
Of course, those skilled in the art may set multiple probability thresholds to distinguish the product churn users into multiple levels of intent to reflow, which the present application does not further limit.
S15: and determining the user category corresponding to the reflux intention level and the value level so as to recall the product loss user according to the target recall mode corresponding to the user category.
Specifically, when the category identification is performed on the product loss user, the method is specifically based on two level evaluation results: reflux intent level and value level. It is easy to understand that the reflux intent level directly reflects the ease with which the product loss user is successfully recalled, while the value level directly reflects the user value that the product loss user brings after being successfully recalled.
Based on the two level evaluation results, different product loss users can be finely classified, so that different modes of user recall can be further performed for the different product loss users. For example, product churn users may be categorized into four categories according to the aforementioned preset value threshold and preset probability threshold. And further, corresponding labels can be set for the product loss users of each category: at the same time, users belonging to the first valence level and the first reflux intention level can be identified as high-value easy-to-reflux; users belonging to both the first value level and the second intent level of reflow may be identified as high value, difficult to reflow; users belonging to both the second value level and the first reflux intent level may be identified as low value easy to reflux; users belonging to both the second value level and the second intent level of reflow may be identified as low value, difficult to reflow.
It should be noted that, common recall modes include a non-manual recall mode such as a short message, a tips message, an in-station message, an on-line activity, and a manual recall mode of manual voice. When different types of product loss users are recalled, based on the user type identification result, only the product loss users with important types tend to use a manual recall mode, and the balance between labor cost and recall rate is realized. For example, as one specific embodiment, the target recall mode corresponding to the first value level may include an artificial voice recall; the target recall mode corresponding to the second value level may not include an artificial voice recall.
The user category identification method for the product loss user provided by the embodiment of the application comprises the following steps: acquiring user data of a product loss user; determining the value level of the product loss user according to consumption data in the user data; substituting a target attribute field in the user data into a pre-established reflux intention evaluation model, and calculating a reflux intention probability value of a product loss user; determining a reflux intention level corresponding to the reflux intention probability value; and determining the user category corresponding to the reflux intention level and the value level so as to recall the product loss user according to the target recall mode corresponding to the user category.
Therefore, the application introduces a data analysis processing technology to process the user data, determines the value level and the reflux intention level of the product loss user from the data level, and further can identify the user category of the product loss user and serve as the guiding basis of user recall so as to adopt a corresponding reasonable targeted recall scheme for the product loss users of different categories. The application not only avoids the waste of manpower recall cost, but also effectively ensures recall rate and recall efficiency.
Based on the above, as a specific embodiment, the pre-established reflux intention evaluation model includes a plurality of reflux intention evaluation models corresponding to the respective value levels one by one; the step S13 may specifically be: substituting the target attribute fields of the product loss users with the same value level into the corresponding reflux intention evaluation model, and calculating the reflux intention probability value of the product loss users with the same value level.
Specifically, in this embodiment, corresponding reflux intention evaluation models may be built for the product loss users with different price levels, so as to obtain finer and more accurate reflux intention probability values.
The process of establishing the reflux intention evaluation model based on regression analysis will be described below. Referring to fig. 5, the embodiment of the application discloses a method for establishing a reflux intention evaluation model, which mainly comprises the following steps:
S21: a logistic regression model of the reflux intent probability with respect to the target attribute field is built.
Specifically, let variable p represent the probability of the intent to reflow, the value range of p is [0,1]; let the variable x i represent the quantized value of the target attribute field, and the range of x i is determined by the attribute of the target attribute field of the corresponding category and its quantization rule. Then, the probability formula for p=1 is:
Wherein x= [ X 1x2…]T;W＝[β1β2…]T;βi ] is the regression coefficient of the corresponding target attribute field. The logistic regression model (logistic model) of the reflow intent probability with respect to the target attribute field is:
Where ε is the random error.
S22: and obtaining a user data sample and a sample reflux result of the product loss user sample.
S23: and determining the parameter values in the logistic regression model according to the user data samples and the sample reflux result.
Specifically, in order to determine the parameter values in the logistic regression model, the application takes the product loss user samples as analysis objects, takes the user data and the reflux results of the product loss user samples as the user data samples and the sample reflux results respectively, and substitutes the user data and the sample reflux results into the logistic regression model to determine the value of each parameter. The parameter value may be determined by using various fitting methods, which is not limited in the present application.
S24: and calculating the relative influence value of each target attribute field on the reflux intention probability.
It should be noted that, in step S23, the logistic regression model is already initially established. Further, the embodiment may further perform optimization updating on the logistic regression model established in step S23 based on the above.
Since the number of target attribute fields contained in the initially established logistic regression model may be large, some of the target attribute fields that have a weak influence on the probability of reflow are not depleted. Thus, the present application can remove some unnecessary target attribute fields by calculating the relative impact value of each target attribute field on the reflux intent probability.
In one embodiment, the relative impact value may be specifically: the target attribute field corresponds to the ratio of the rate of change of the quantized value x i to the rate of change of the resulting reflow probability value p.
S25: and removing the target attribute field with the relative influence value lower than the preset influence threshold value.
It is easy to understand that, after the target attribute field with a lower relative influence value is removed, the corresponding regression coefficient is also removed.
S26: parameter values in the updated reflux intent assessment model are determined.
Since some expressions in the original logistic regression model are removed, in order to ensure model accuracy, the parameter values in the reflux intent assessment model need to be recalculated in order to update the reflux intent assessment model.
Referring to fig. 6, the embodiment of the application discloses a user category identification method of a product loss user, which mainly comprises the following steps:
s31: user data of a product loss user is obtained.
S32: judging whether a product loss user is a valid user or not according to the user data; if not, entering S33; if yes, the process proceeds to S34.
S33: recall the product loss user according to the target recall mode corresponding to the invalid user; proceed to S38.
Specifically, in view of the fact that a part of product loss users may be invalid users, the application also identifies the invalid users and sets a corresponding target recall mode. For inactive users, such as those selecting seal number, the recall modes of in-station messages, tips messages, online activities, etc., will not be accessible to the user, and thus, the present application may avoid using these modes in particular when recalling inactive users.
S34: determining the value level of the product loss user according to consumption data in the user data; the process advances to S35.
S35: substituting a target attribute field in the user data into a pre-established reflux intention evaluation model, and calculating a reflux intention probability value of a product loss user; and proceeds to S36.
S36: determining a reflux intention level corresponding to the reflux intention probability value; the process advances to S37.
S37: and determining the user category corresponding to the reflux intention level and the value level so as to recall the product loss user according to the target recall mode corresponding to the user category.
S38: and obtaining a reflux result of the product loss user.
S39: and merging the product loss user into a product loss user sample, and updating the reflux intention evaluation model according to the updated product loss user sample.
Specifically, in order to further perfect the reflux intention evaluation model and improve the accuracy of the recognition result, the user category recognition method provided by the embodiment can further monitor the subsequent actual reflux result after recall is performed on each product loss user, so that the product loss users are also taken into the product loss user sample, and the reflux intention evaluation model is further optimized and updated through enriching the sample capacity.
The technical scheme of the application will be described by taking the user category identification of a game product loss user as an example.
Firstly, a product loss user sample of the game product is obtained by collecting a large number of recall investigation records of the loss users of the game product. Then, according to a logistic regression analysis method, a reflux intention evaluation model of the game product is established according to the user data of the product loss user sample and the sample reflux result.
In the process of model establishment, the embodiment of the application specifically selects more than thirty target attribute fields in the user data as related variables to construct a logistic regression model, wherein the logistic regression model comprises user age, user gender, loss days, game grade, game war value, consumption total, game balance, user point, user punishment record, customer service order number, customer service problem resolvable property and the like. After the related variables are determined, fitting and determining the magnitude of each parameter value in the reflux intention evaluation model according to the final reflux result of each user in the product loss user sample.
In order to optimize the reflux intention evaluation model, the embodiment further calculates the relative influence of each target attribute field on the reflux intention probability respectively, and then deletes the target attribute field with lower relative influence, and only retains the target attribute field with obvious influence on the reflux intention probability in the model.
After screening and optimization, eight target attribute fields with significant influence are reserved, and in particular, see fig. 7, including: the user age and the user sex reflecting the user social attribute information; the consumption total amount, the game grade, the game war value and the loss days of the account attribute information are reflected; customer service order number reflecting customer service attribute information and customer service problem resolvable.
The quantized variables of the eight target attribute fields are sequentially marked as x i, and then the logistic regression model is changed into:
Logit(p)＝β0+β1x1+β2x2+β3x3+β4x4+β5x5+β6x6+β7x7+β8x8+ε.
And re-performing parameter fitting by using the product loss sample, and re-determining the parameter values in the model one by one. The resulting optimized reflux intent assessment model is shown below:
Logit(p)＝-6.8e-0.0512393x1+0.2660933x2-0.8x3+0.008774x4-0.8x5-0.0056803
x6+0.2077715x7+0.0529246x8-0.1334708。
And when the reflux intention evaluation model is built, the category identification can be carried out for the product loss user to be analyzed.
The system automatically loads user data of the product loss users, identifies effective users, extracts consumption data of the effective users, and determines value levels of the effective users according to preset value thresholds; and quantizing eight target attribute fields in the user data, then substituting the quantized target attribute fields into a backflow intention evaluation model, calculating backflow intention probability values of the effective users, and further determining backflow intention levels of the effective users.
In this embodiment, a preset value threshold with a size of 100000 is specifically set, and a product loss user is divided into two value levels; and a preset probability threshold value of 0.7 is set, and the product loss users are divided into two reflux intention levels. Thus, valid users can be identified as being in particular four categories.
The user category recognition result can be used as guiding data in the process of recalling lost users. As a specific example, embodiments of the present application implement recall of a product loss user with specific reference to table 1. The target recall patterns for different categories of product loss users are specifically shown in table 1.
TABLE 1
Referring to fig. 8, an embodiment of the present application discloses a user category identification system for a product loss user, which mainly includes:
a data acquisition module 41, configured to acquire user data of a product loss user;
a first processing module 42, configured to determine a value level of the product loss user according to consumption data in the user data;
a model building module 43, configured to pre-build a reflow intention evaluation model of the reflow intention probability with respect to the target attribute field;
A second processing module 44, configured to substitute a target attribute field in the user data into the reflux intention evaluation model, and calculate a reflux intention probability value of the product loss user; determining a reflux intention level corresponding to the reflux intention probability value;
The category identifying module 45 is configured to determine a category of the user corresponding to both the reflux intention level and the value level, so as to recall the product loss user according to a target recall mode corresponding to the category of the user.
Therefore, the user category identification system of the product loss user disclosed by the embodiment of the application introduces a data analysis processing technology to process the user data, determines the value level and the reflux intention level of the product loss user from the data level, and further can identify the user category of the product loss user and serve as a guide basis for user recall so as to adopt a corresponding reasonable targeted recall scheme for different types of product loss users. The application not only avoids the waste of manpower recall cost, but also effectively ensures recall rate and recall efficiency.
For the specific content of the user category identification system of the product loss user, reference may be made to the detailed description of the user category identification method of the product loss user, which is not repeated herein.
Based on the foregoing, in the user category identification system of the product loss user disclosed in the embodiment of the present application, in a specific implementation manner, the model building module 43 is specifically configured to: a plurality of reflux intention evaluation models which are respectively in one-to-one correspondence with the value levels are established in advance;
the second processing module 44 is specifically configured to: substituting the target attribute fields of the product loss users with the same value level into the corresponding reflux intention evaluation model, and calculating the reflux intention probability value of the product loss users with the same value level.
Based on the foregoing, in the user category identification system of the product loss user disclosed in the embodiment of the present application, in a specific implementation manner, the model building module 43 specifically includes:
the model building unit is used for building a logistic regression model of the reflux intention probability about the target attribute field;
the sample acquisition unit is used for acquiring a user data sample and a sample reflux result of a product loss user sample;
And the parameter determining unit is used for determining the parameter value in the logistic regression model according to the user data sample and the sample reflux result.
Based on the foregoing, as a specific implementation manner, in the user category identification system of the product loss user disclosed in the embodiment of the present application, the model building module 43 further includes:
The model optimization unit is used for calculating the relative influence value of each target attribute field on the reflux intention probability; removing target attribute fields with relative influence values lower than a preset influence threshold value; parameter values in the updated reflux intent assessment model are determined.
Based on the foregoing, as a specific implementation manner, in the user category identification system of the product loss user disclosed in the embodiment of the present application, the model optimization unit is further configured to:
After the product loss user is recalled according to the target recall mode corresponding to the user category, a reflux result of the product loss user is obtained; incorporating the product loss user into a product loss user sample; and updating the reflux intention evaluation model according to the updated product loss user sample.
Based on the foregoing, as a specific implementation manner, in the user category identification system of the product loss user disclosed in the embodiment of the present application, the first processing module 42 is specifically configured to:
If the total consumption amount is higher than a preset value threshold, determining the value level of the product loss user as a first value level; the target recall mode corresponding to the first value level comprises manual voice recall;
if the total consumption amount is not higher than the preset value threshold, determining the value level of the product loss user as a second value level; the target recall mode corresponding to the second value level does not include manual voice recall.
Based on the foregoing, as a specific implementation manner, in the user category identification system of the product loss user disclosed in the embodiment of the present application, the data obtaining module 41 is further configured to:
After obtaining the user data of the product loss user, judging whether the product loss user is a valid user or not according to the user data; and after the product loss user is judged to be the invalid user, the product loss user is recalled according to the target recall mode corresponding to the invalid user.
Further, the embodiment of the application also discloses a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is used for realizing the steps of the user category identification method of any product loss user when being executed by a processor.
For the specific content of the computer-readable storage medium, reference may be made to the foregoing detailed description of the user category identification method for the product loss user, which is not repeated here.
In the application, each embodiment is described in a progressive manner, and each embodiment is mainly used for illustrating the difference from other embodiments, and the same similar parts among the embodiments are mutually referred. For the apparatus disclosed in the examples, since it corresponds to the method disclosed in the examples, the description is relatively simple, and the relevant points are referred to in the description of the method section.
It should also be noted that in this document, relational terms such as "first" and "second" are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Furthermore, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The technical scheme provided by the application is described in detail. The principles and embodiments of the present application have been described herein with reference to specific examples, the description of which is intended only to facilitate an understanding of the method of the present application and its core ideas. It should be noted that it will be apparent to those skilled in the art that the present application may be modified and practiced without departing from the spirit of the present application.
Claims (16)
1. A method for identifying a user category of a product loss user, comprising:
acquiring user data of the product loss user;
Determining the value level of the product loss user according to consumption data in the user data; the value levels include a first value level and a second value level;
substituting a target attribute field in the user data into a pre-established reflux intention evaluation model, and calculating a reflux intention probability value of the product loss user;
Determining a reflux intention level corresponding to the reflux intention probability value; the reflow intent level includes a first reflow intent level and a second reflow intent level;
Determining a user category corresponding to both the reflux intention level and the value level, so as to recall the product loss user according to a target recall mode corresponding to the user category; wherein users belonging to both the first value level and the first intent level of reflow are identified as high value easy-reflow; users belonging to both the first value level and the second intent level are identified as high value hard to reflow; at the same time, users belonging to the second value level and the first reflux intention level are marked as low-value easy-to-reflux; users belonging to both the second value level and the second intent level of reflow are identified as low value, difficult to reflow.
2. The user category identification method according to claim 1, wherein the pre-established back flow intention assessment model includes a plurality of back flow intention assessment models respectively corresponding to the respective value levels one to one;
Substituting the target attribute field in the user data into a pre-established reflux intention evaluation model, and calculating the reflux intention probability value of the product loss user comprises the following steps:
Substituting the target attribute fields of the product loss users with the same value level into the corresponding reflux intention evaluation model, and calculating the reflux intention probability value of the product loss users with the same value level.
3. The user category identification method according to claim 1, wherein the reflow intent assessment model is specifically built by:
Establishing a logistic regression model of the reflux intention probability with respect to the target attribute field;
obtaining a user data sample and a sample reflux result of a product loss user sample;
And determining parameter values in the logistic regression model according to the user data samples and the sample reflow result.
4. A user category identification method as claimed in claim 3, further comprising, after said determining parameter values in the logistic regression model:
Calculating the relative influence value of each target attribute field on the reflux intention probability;
Removing target attribute fields with relative influence values lower than a preset influence threshold value;
Parameter values in the updated reflux intent assessment model are determined.
5. The method of claim 4, further comprising, after said recalling said product-loss user in a targeted recall manner corresponding to said user category:
obtaining a reflux result of the product loss user;
Incorporating the product loss user into the product loss user sample;
And updating the reflux intention evaluation model according to the updated product loss user sample.
6. The method according to any one of claims 1 to 5, wherein determining the value level of the product churn user based on consumption data in the user data comprises:
If the total consumption amount is higher than a preset value threshold, determining the value level of the product loss user as a first value level; the target recall mode corresponding to the first valence level comprises manual voice recall;
If the total consumption amount is not higher than the preset value threshold, determining the value level of the product loss user as a second value level; the target recall mode corresponding to the second value level does not comprise manual voice recall.
7. The method of claim 6, further comprising, after said obtaining the user data of the product churn user:
judging whether the product loss user is an effective user or not according to the user data;
if yes, executing the step of determining the value level of the product loss user according to the consumption data in the user data;
If not, the product loss user is recalled according to the target recall mode corresponding to the invalid user.
8. A user category identification system for a product churn user, comprising:
The data acquisition module is used for acquiring the user data of the product loss user;
The first processing module is used for determining the value level of the product loss user according to the consumption data in the user data; the value levels include a first value level and a second value level;
the model building module is used for pre-building a reflux intention evaluation model of the reflux intention probability about the target attribute field;
The second processing module is used for substituting the target attribute field in the user data into the reflux intention evaluation model and calculating the reflux intention probability value of the product loss user; determining a reflux intention level corresponding to the reflux intention probability value; the reflow intent level includes a first reflow intent level and a second reflow intent level;
The category identification module is used for determining a user category corresponding to the reflux intention level and the value level so as to recall the product loss user according to a target recall mode corresponding to the user category; wherein users belonging to both the first value level and the first intent level of reflow are identified as high value easy-reflow; users belonging to both the first value level and the second intent level are identified as high value hard to reflow; at the same time, users belonging to the second value level and the first reflux intention level are marked as low-value easy-to-reflux; users belonging to both the second value level and the second intent level of reflow are identified as low value, difficult to reflow.
9. The system of claim 8, wherein the modeling is specifically configured to: a plurality of reflux intention evaluation models which are respectively in one-to-one correspondence with the value levels are established in advance;
The second processing module is specifically configured to: substituting the target attribute fields of the product loss users with the same value level into the corresponding reflux intention evaluation model, and calculating the reflux intention probability value of the product loss users with the same value level.
10. The system of claim 8, wherein the model building module specifically comprises:
the model building unit is used for building a logistic regression model of the reflux intention probability about the target attribute field;
the sample acquisition unit is used for acquiring a user data sample and a sample reflux result of a product loss user sample;
And the parameter determining unit is used for determining the parameter value in the logistic regression model according to the user data sample and the sample reflux result.
11. The system of claim 10, wherein the modeling module further comprises:
The model optimization unit is used for calculating the relative influence value of each target attribute field on the reflux intention probability; removing target attribute fields with relative influence values lower than a preset influence threshold value; parameter values in the updated reflux intent assessment model are determined.
12. The system according to claim 11, wherein the model optimization unit is further configured to:
After the product loss user is recalled according to the target recall mode corresponding to the user category, a reflux result of the product loss user is obtained; incorporating the product loss user into a product loss user sample; and updating the reflux intention evaluation model according to the updated product loss user sample.
13. The system according to any one of claims 8 to 12, wherein the first processing module is specifically configured to:
If the total consumption amount is higher than a preset value threshold, determining the value level of the product loss user as a first value level; the target recall mode corresponding to the first value level comprises manual voice recall;
if the total consumption amount is not higher than the preset value threshold, determining the value level of the product loss user as a second value level; the target recall mode corresponding to the second value level does not include manual voice recall.
14. The system of claim 13, wherein the data acquisition module is further configured to:
After obtaining the user data of the product loss user, judging whether the product loss user is a valid user or not according to the user data; and after the product loss user is judged to be the invalid user, the product loss user is recalled according to the target recall mode corresponding to the invalid user.
15. An electronic device, comprising:
a memory for storing a computer program;
A processor for executing the computer program to implement the steps of the user category identification method of a product churn user as claimed in any one of claims 1 to 7.
16. A computer readable storage medium, wherein a computer program is stored in the computer readable storage medium, which computer program, when being executed by a processor, is adapted to carry out the steps of the user category identification method of a product loss user according to any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910959099.0A CN112651582B (en) | 2019-10-10 | 2019-10-10 | User category identification method and related equipment for product loss user |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910959099.0A CN112651582B (en) | 2019-10-10 | 2019-10-10 | User category identification method and related equipment for product loss user |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112651582A CN112651582A (en) | 2021-04-13 |
| CN112651582B true CN112651582B (en) | 2024-04-26 |
Family
ID=75342870
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910959099.0A Active CN112651582B (en) | 2019-10-10 | 2019-10-10 | User category identification method and related equipment for product loss user |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112651582B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113343089A (en) * | 2021-06-11 | 2021-09-03 | 北京完美赤金科技有限公司 | User recall method, device and equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107833063A (en) * | 2017-11-30 | 2018-03-23 | 上海科瓴医疗科技有限公司 | Pharmacy member is lost in early warning and intelligent interfering system and method |
| CN109461023A (en) * | 2018-10-12 | 2019-03-12 | 中国平安人寿保险股份有限公司 | It is lost user and retrieves method and device, electronic equipment, storage medium |
| CN109583651A (en) * | 2018-12-03 | 2019-04-05 | 焦点科技股份有限公司 | A kind of method and apparatus for insuring electric business platform user attrition prediction |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190182059A1 (en) * | 2017-12-12 | 2019-06-13 | Facebook, Inc. | Utilizing machine learning from exposed and non-exposed user recall to improve digital content distribution |
-
2019
- 2019-10-10 CN CN201910959099.0A patent/CN112651582B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107833063A (en) * | 2017-11-30 | 2018-03-23 | 上海科瓴医疗科技有限公司 | Pharmacy member is lost in early warning and intelligent interfering system and method |
| CN109461023A (en) * | 2018-10-12 | 2019-03-12 | 中国平安人寿保险股份有限公司 | It is lost user and retrieves method and device, electronic equipment, storage medium |
| CN109583651A (en) * | 2018-12-03 | 2019-04-05 | 焦点科技股份有限公司 | A kind of method and apparatus for insuring electric business platform user attrition prediction |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112651582A (en) | 2021-04-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110837931B (en) | Customer churn prediction method, device and storage medium | |
| Fryges et al. | Exports and profitability: first evidence for German manufacturing firms | |
| CN111461180A (en) | Sample classification method and device, computer equipment and storage medium | |
| CN111915156B (en) | Service pushing method based on user value, electronic equipment and storage medium | |
| CN113643070A (en) | Intelligent information pushing method and system based on big data | |
| CN112651582B (en) | User category identification method and related equipment for product loss user | |
| CN107977855B (en) | Method and device for managing user information | |
| Wilson et al. | Mobility, inequality, and beliefs about distribution and redistribution | |
| Adão et al. | Fast and slow technological transitions | |
| Abel-Koch | Firm size and the choice of export mode | |
| CN111222769A (en) | Annual report data quality evaluation method and device, electronic equipment and storage medium | |
| CN111882113A (en) | Enterprise mobile banking user prediction method and device | |
| CN113254463A (en) | List updating method, system and medium | |
| CN110690987B (en) | Account information management method, device and equipment | |
| JP2011186855A (en) | Question recommending device, method and program | |
| CN114693243A (en) | Enterprise platform management method, system and storage medium | |
| CN113435857A (en) | Data analysis method and device for applicants | |
| Huang et al. | Exporting firm dynamics and productivity growth: Evidence from China | |
| Goswami et al. | Technological efforts, firm ownership and productivity: A study of information technology service firms in India | |
| Chen | Information asymmetry and dynamic sourcing: Evidence from Chinese firms | |
| Darden | A COMMENT ON | |
| CN109754290B (en) | Game data processing method and device | |
| Weng et al. | Outward foreign direct investment and product quality of domestic productions: an empirical investigation | |
| CN117726360B (en) | Client behavior driven industrial product sales analysis method and related components | |
| Csiszár | Digital Intensity and Technology Adoption in European Enterprises: Comparative Analysis Using NRI and DESI 2023 Data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 40042608 Country of ref document: HK |
|
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |