Propensity score-based behavior attribution system, method, and computer program product

Abstract

The invention discloses a behavior attribution system and method based on tendency scores and a computer program product, and the system comprises a first screening module which screens users; the dividing module is used for dividing the users into an experimental group and a control group and only providing incentive operation for the users in the experimental group; the acquisition module is used for acquiring users which do not perform the first behavior in the control group to obtain a first user set; the determining module is used for determining tendency score distribution of the first user set; the second screening module is used for screening out a second user set from the experimental group according to the tendency score distribution; and the attribution module is used for attributing the ratio difference of the first behavior performed by the users in the first user set and the second user set in the same time period after the excitation operation is provided to be caused by the excitation operation. According to the method, the influence of other factors on attribution of the incentive operation can be eliminated, and the attribution accuracy of the user behavior aiming at each incentive operation is improved, so that the accuracy of identifying whether the platform data is safe or not is improved, and the platform data safety is guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of computer application technologies, and in particular, to a behavior attribution system, method, and computer program product based on propensity scores. Background Art

[0002] Currently, in order to promote the sales of their own goods or services, Internet service platforms (such as online shopping platforms, online car-hailing platforms, sharing platforms, maps, music, etc.) will provide some incentive operations to users, such as: red envelopes, coupons, free trials, giveaways, etc. Analyzing the impact of these incentive operations on users' purchase or rental behaviors, on the one hand, helps to more accurately identify abnormal situations that endanger data security in the platform, such as data theft, data modification, fraud, default, etc., and ensure the security of platform data; on the other hand, it can understand whether the investment in incentive operations by the platform has generated the expected returns, and guide the platform to adjust and optimize the incentive operations.

[0003] In practice, when an incentive operation is provided to a certain user and the user makes a purchase or rental, it cannot be directly attributed to the incentive operation, because it is possible that the user will also generate purchase or rental behaviors due to natural factors such as actual needs and personal preferences even without the incentive operation provided to the user.

[0004] In view of this, it is necessary to provide a method that can exclude the interference of other factors and attribute the purchase or rental behaviors of users to each incentive operation. Summary of the Invention

[0005] In view of this, the main purpose of the present invention is to propose a behavior attribution system, method, and computer program product based on propensity scores, in order to at least partially solve at least one of the above technical problems.

[0006] To solve the above technical problems, in a first aspect of the present invention, a behavior attribution system based on propensity scores is proposed. The system includes:

[0007] A first screening module for screening users who meet the incentive operation conditions;

[0008] A partitioning module for dividing the users into an experimental group and a control group, and only providing incentive operations to the users in the experimental group;

[0009] A collection module for collecting users in the control group who do not perform the first behavior to obtain a first user set; providing incentive operations to the users may cause the users to perform the first behavior;

[0010] A determination module for determining the propensity score distribution of the first user set;

[0011] A second screening module, configured to screen out a second user set from the experimental group according to the propensity score distribution of the first user set;

[0012] An attribution module, configured to attribute the ratio difference of the users in the first user set and the second user set performing the first behavior within the same time period after providing the incentive operation to the incentive operation.

[0013] According to a preferred embodiment of the present invention, the first screening module includes:

[0014] A sub-collection module, configured to collect user data;

[0015] A first sub-screening module, configured to screen out first target users with similar user data distributions;

[0016] A second sub-screening module, configured to screen out users who meet the incentive operation conditions from the first target users.

[0017] According to a preferred embodiment of the present invention, the determination module includes:

[0018] A first input module, configured to input the user data of each user in the first user set into a propensity model to obtain the propensity scores of the users in the first user set;

[0019] A sub-determination module, configured to determine the propensity score distribution of the first user set according to the propensity scores of the users in the first user set.

[0020] According to a preferred embodiment of the present invention, the second screening module includes:

[0021] A second input module, configured to input the user data of each user in the experimental group into a propensity model to obtain the propensity scores of the users in the experimental group;

[0022] A sub-screening module, configured to screen out users in the experimental group whose propensity scores are within the propensity score distribution of the first user set to obtain a second user set.

[0023] According to a preferred embodiment of the present invention, the system further includes:

[0024] An optimization module, configured to optimize the incentive strategy according to the probability of each incentive operation causing the first behavior of the user.

[0025] To solve the above technical problems, a second aspect of the present invention provides a behavior attribution method based on propensity scores, and the method includes:

[0026] Screen out users who meet the incentive operation conditions;

[0027] Divide the users into an experimental group and a control group, and only provide incentive operations to the users in the experimental group;

[0028] Collect the users in the control group who do not perform the first behavior to obtain a first user set; Providing an incentive operation to a user may cause the user to perform the first behavior;

[0029] Determine the propensity score distribution of the first user set;

[0030] Screen out a second user set from the experimental group according to the propensity score distribution of the first user set;

[0031] Attribute the difference in the ratio of users performing the first behavior between the first user set and the second user set within the same time period after providing the incentive operation to the incentive operation.

[0032] According to a preferred embodiment of the present invention, the screening of users meeting the incentive operation conditions includes:

[0033] Collect user data;

[0034] Screen out first target users with similar user data distributions;

[0035] Screen out users meeting the incentive operation conditions from the first target users.

[0036] According to a preferred embodiment of the present invention, the determining the propensity score distribution of the first user set includes:

[0037] Input the user data of each user in the first user set into a propensity model to obtain the propensity scores of each user in the first user set;

[0038] Determine the propensity score distribution of the first user set according to the propensity scores of each user in the first user set.

[0039] According to a preferred embodiment of the present invention, the screening out a second user set from the experimental group according to the propensity score distribution of the first user set includes:

[0040] Input the user data of each user in the experimental group into a propensity model to obtain the propensity scores of each user in the experimental group;

[0041] Screen out the users in the experimental group whose propensity scores are within the propensity score distribution of the first user set to obtain a second user set.

[0042] According to a preferred embodiment of the present invention, the method further includes:

[0043] Optimize the incentive strategy according to the probability of each incentive operation causing the user's first behavior.

[0044] To solve the above technical problems, a third aspect of the present invention provides an electronic device, including:

[0045] a processor; and

[0046] a memory storing computer-executable instructions, which when executed cause the processor to execute the method described in any one of the above.

[0047] To solve the above technical problems, a fourth aspect of the present invention provides a computer program product, including a computer program, which when executed by a processor implements the method described in any one of the above.

[0048] In summary, the present invention divides users who meet the incentive operation conditions into an experimental group and a control group, and only provides incentive operations to the users in the experimental group. Then, all users in the control group do not have incentive operations. By collecting users in the control group who do not perform the first behavior, a first user set that is similar to the users in the experimental group and does not have incentives and does not perform the first behavior is obtained. Then, according to the propensity scores of the first user set, a second user set that matches the first user set is screened out from the experimental group. Thus, the users in the second user set and the users in the first user set are highly similar in other dimensional sub-user data except for whether incentive operations are provided, and can be considered the same user set. Therefore, the probability that the users in the first user set and the second user set perform the first behavior under the influence of natural factors is the same. Since the reasons for the users in the second user set to perform the first behavior include incentive operations and other factors (such as actual needs, personal preferences), and the reasons for the users in the first user set to perform the first behavior only include other factors, the difference in the ratio of the users in the first user set and the second user set who perform the first behavior within the same time period after providing incentive operations can be attributed to the incentive operations. Thus, the influence of other factors on the attribution of incentive operations is excluded, the accuracy of attributing user behavior to each incentive operation is improved, thereby improving the accuracy of identifying whether the platform data is secure, and ensuring the security of the platform data. At the same time, the incentive strategy is optimized to improve the incentive effect while saving platform costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to make the technical problems solved by the present invention, the technical means adopted, and the technical effects obtained more clear, the specific embodiments of the present invention will be described in detail below with reference to the drawings. It should be noted that the drawings described below are only the drawings of the exemplary embodiments of the present invention, and those skilled in the art can obtain the drawings of other embodiments without creative efforts.

[0050] Figure 1 It is a schematic structural framework diagram of a behavior attribution device based on propensity scores provided by an embodiment of the present invention;

[0051] Figure 2 is a schematic flowchart of a behavior attribution method based on propensity score provided by an embodiment of the present invention;

[0052] Figure 3 is a structural block diagram of an exemplary embodiment of an electronic device according to the present invention;

[0053] Figure 4 is a schematic diagram of an embodiment of a computer-readable medium of the present invention. Detailed implementation manners

[0054] On the premise of conforming to the technical concept of the present invention, the structures, performances, effects or other features described in a specific embodiment can be combined into one or more other embodiments in any suitable manner.

[0055] During the introduction of specific embodiments, the detailed descriptions of structures, performances, effects or other features are for those skilled in the art to fully understand the embodiments. However, it does not exclude that those skilled in the art can implement the present invention with technical solutions that do not include the above-mentioned structures, performances, effects or other features under specific circumstances. The figures in the drawings are only exemplary demonstrations, and do not mean that all the contents, operations and steps in the figures must be included in the solutions of the present invention, nor does it mean that the execution order must be as shown in the figures.

[0056] Refer to Figure 1 , Figure 1 is a schematic structural framework diagram of a behavior attribution system based on propensity score provided by the present invention. As Figure 1 shown, the system includes:

[0057] A first screening module 11, configured to screen users who meet the incentive operation conditions;

[0058] A partitioning module 12, configured to divide the users into an experimental group and a control group, and only provide incentive operations to the users in the experimental group;

[0059] A collection module 13, configured to collect users in the control group who do not perform the first behavior to obtain a first user set; providing incentive operations to the users may cause the users to perform the first behavior;

[0060] A determination module 14, configured to determine the propensity score distribution of the first user set;

[0061] A second screening module 15, configured to screen out a second user set from the experimental group according to the propensity score distribution of the first user set;

[0062] An attribution module 16, configured to attribute the ratio difference of the users in the first user set and the second user set who perform the first behavior within the same time period after providing the incentive operation to the incentive operation.

[0063] In a specific embodiment, the first screening module 11 includes:

[0064] A sub-collection module, configured to collect user data;

[0065] A first sub-screening module, configured to screen out first target users with similar user data distributions;

[0066] A second sub-screening module, configured to screen out users who meet the incentive operation conditions from the first target users.

[0067] The determination module 14 includes:

[0068] A first input module, configured to input the user data of each user in the first user set into a propensity model to obtain the propensity scores of each user in the first user set;

[0069] A sub-determination module, configured to determine the propensity score distribution of the first user set according to the propensity scores of each user in the first user set.

[0070] The second screening module 15 includes:

[0071] A second input module, configured to input the user data of each user in the experimental group into a propensity model to obtain the propensity scores of each user in the experimental group;

[0072] A sub-screening module, configured to screen out users in the experimental group whose propensity scores are within the propensity score distribution of the first user set to obtain a second user set.

[0073] Furthermore, the system may further include:

[0074] An optimization module, configured to optimize the incentive strategy according to the probability of each incentive operation causing the first behavior of the user.

[0075] Based on Figure 1 The behavior attribution system based on propensity scores described above, an embodiment of the present invention further provides a behavior attribution method based on propensity scores. As Figure 2 , the user action attribution method based on the propensity score includes:

[0076] S1. Screen out users who meet the incentive operation conditions;

[0077] In this embodiment, in order to maximize the effect of the incentive operation on users' purchase or rental of goods, it is necessary to select users who meet the incentive operation conditions from the users as the experimental group or the control group. Among them: the incentive operation can be an incentive method provided by the platform to users to promote the sale or rental of its goods or services, such as: issuing red envelopes, coupons, providing free trials, giving gifts, etc. The incentive operation conditions can be configured according to different incentive operations, for example: the condition for issuing red envelopes can be configured as: the user logs in and browses for one month continuously; the condition for providing free trials can be configured as: the user logs in and collects goods for one month continuously. And so on.

[0078] In a preferred example, this step may further screen users with similar distributions to ensure high similarity between users in the subsequent experimental group and the control group, thereby improving the accuracy of subsequent user set matching. This step may include:

[0079] S11, collecting user data;

[0080] The user data may be any user-related data that the user chooses to make public or has been desensitized, and may include sub-user data of one or more dimensions, and the sub-user data may be: user gender, age, region, education, occupation, fraud record, violation record, purchase or return record, or user communication record.

[0081] Among them, the purchase record refers to the record of the user purchasing goods on the platform. The return record refers to the record of whether the user returns the goods on time after renting the platform goods. The goods can be physical goods, virtual goods, services, etc., and the present invention does not make specific limitations. The user communication record can include the user's address book contacts, communication records and other communication-related information.

[0082] S12, screening out first target users with similar user data distribution;

[0083] Among them: user data distribution similarity means that the sub-user data distribution of all dimensions contained in the user data is similar. For example: user data includes: gender, fraud record, purchase record, then when the user's gender, fraud record, and purchase record are the same or similar (for example: gender: female, fraud record: none, purchase record: 2 to 3 times), their data distribution is similar and they can be used as the first target user.

[0084] S13: Filter out users who meet the incentive operation conditions from the first target users.

[0085] For example, the incentive operation is: issuing red envelopes, and the condition for issuing red envelopes is: the user logs in and browses for one month continuously. Then, in this step, users who log in and browse for one month continuously are screened out from the first target users.

[0086] In the above example, steps S12 and S13 are executed in sequence with step S12 being executed first and then step S13 for illustration purposes. In actual implementation, the execution order of steps S12 and S13 can be interchanged, and the present invention does not make specific limitations in this regard.

[0087] S2. Divide the users into an experimental group and a control group, and provide incentive operations only to the users in the experimental group.

[0088] Exemplarily, the users screened in step S1 can be divided into an experimental group and a control group according to a predetermined ratio. For example, 70% are in the experimental group and 30% are in the control group. Taking the incentive operation of sending red envelopes as an example, red envelopes are sent only to the users in the experimental group and not to the users in the control group.

[0089] S3. Collect the users in the control group who do not perform the first behavior to obtain a first user set.

[0090] The objective of the present invention is to exclude other natural factors and attribute the first behavior of users caused by each incentive operation. Among them, the first behavior may be caused by providing incentive operations to users, or may be caused by other natural factors (such as user preferences, user needs, etc.). The first behavior can be behaviors such as purchase, rental, use, etc. that the platform wants to promote.

[0091] After providing incentive operations only to the users in the experimental group in step S2, the ratio of the users in the experimental group who perform the first operation within a preset time (such as within one month after providing the incentive operation) and the ratio of the users in the control group who perform the first operation can be collected respectively. At the same time, the users in the control group who do not perform the first behavior within a preset time (such as within one month after providing the incentive operation) can be collected to obtain a first user set. Then, the users in the first user set will not perform the first behavior under the influence of natural factors.

[0092] S4. Determine the propensity score distribution of the first user set.

[0093] In this embodiment, the propensity score is a score obtained by evaluating users based on user data in various dimensions, that is, a user rating value obtained by reducing multi-dimensional user data to one dimension. By comparing the propensity scores between users, the similarity of user data in each dimension can be determined, and then the similarity between users can be determined. Among them, the propensity score distribution of the user set can be obtained by averaging or taking the median of the propensity scores of all users in the user set, or by selecting the minimum propensity score and the maximum propensity score to obtain a score interval. The propensity score of a single user can be calculated through a propensity model. Then this step may include:

[0094] S41. Input the user data of each user in the first user set into the propensity model to obtain the propensity scores of each user in the first user set.

[0095] In this embodiment, the propensity model can be obtained by training a logistic regression model.

[0096] S42. Determine the propensity score distribution of the first user set according to the propensity scores of each user in the first user set.

[0097] Exemplarily, the minimum propensity score and the maximum propensity score in the first user set can be selected, and the obtained score interval is used as the propensity score distribution of the first user set.

[0098] S5. Screen out the second user set from the experimental group according to the propensity score distribution of the first user set.

[0099] In this step, the second user set is screened out from the experimental group according to the propensity score distribution of the first user set. Then, the users in the second user set are similar to the users in the first user set in terms of sub-user data in each dimension. The only difference between them is that the users in the first user set do not provide incentive operations, while the users in the second user set provide incentive operations.

[0100] Exemplarily, this step may include:

[0101] S51. Input the user data of each user in the experimental group into the propensity model to obtain the propensity scores of each user in the experimental group.

[0102] Among them: The propensity model can be obtained by training a logistic regression model.

[0103] S52. Screen out the users in the experimental group whose propensity scores are within the propensity score distribution of the first user set to obtain the second user set.

[0104] If the propensity score distribution of the first user set is obtained by averaging or taking the median of the propensity scores of all users in the first user set, then screen out the users in the experimental group whose propensity scores are equal to the value obtained by averaging or taking the median to obtain the second user set.

[0105] If the propensity score distribution of the first user set is obtained by selecting the minimum propensity score and the maximum propensity score of the propensity scores of all users in the first user set, and the obtained score interval, then screen out the users in the experimental group whose propensity scores are within this interval to obtain the second user set.

[0106] S6. Attribute the ratio difference of the first behavior of the users in the first user set and the second user set within the same time period after providing the incentive operation to the incentive operation.

[0107] Exemplarily, the ratio Q1 of users in the first user set performing the first behavior within 6 months after providing the incentive operation and the ratio Q2 of users in the second user set performing the first behavior can be collected respectively. Then, the attribution of the first behavior of Q2 - Q1 is caused by the incentive operation.

[0108] The present invention screens out a second user set that matches the first user set from the experimental group according to the propensity scores of the first user set. Thus, except for whether the incentive operation is provided or not, the sub - user data in other dimensions of the users in the second user set are similar to those of the users in the first user set, and they can be considered as the same or similar user sets. Then, the probability of the users in the first user set and the second user set performing the first behavior under the influence of natural factors is the same. For example, within a preset time after providing the incentive operation (such as within one month after providing the incentive operation) in steps S4 and S5, the users in the first user set and the second user set will not perform the first behavior under the influence of natural factors, and outside the preset time after providing the incentive operation (such as outside one month after providing the incentive operation), the probability of the users in the first user set and the second user set performing the first behavior under the influence of natural factors is also the same. Since the reasons for the users in the second user set to perform the first behavior include the incentive operation and other factors (such as actual needs, personal preferences), and the reasons for the users in the first user set to perform the first behavior only include other factors, then the difference in the ratio of the users in the first user set and the second user set performing the first behavior within the same time period after providing the incentive operation can be attributed to the incentive operation, thereby excluding the influence of other factors on the attribution of the incentive operation, improving the accuracy of attributing each incentive operation to user behavior, and thus improving the accuracy of identifying whether the platform data is secure, ensuring the security of platform data; at the same time, optimizing the incentive strategy, improving the incentive effect while saving platform costs.

[0109] Furthermore, on the basis of step S6, the platform incentive strategy can be further optimized to improve the promotion effect and save costs. Then, the method may further include:

[0110] S7. Optimize the incentive strategy according to the probability of each incentive operation causing the first behavior of users.

[0111] For example: the proportion of incentive operations with a probability of causing the first behavior of users higher than the threshold in all incentive operations can be increased in the incentive strategy, the proportion of incentive operations with a probability of causing the first behavior of users less than the first threshold in all incentive operations can be decreased, and the incentive operations with a probability of causing the first behavior of users less than the second threshold can be removed.

[0112] Wherein: the threshold, the first threshold, and the second threshold decrease in sequence.

[0113] Those skilled in the art can understand that each module in the above device embodiments can be distributed in the device as described, or can be correspondingly changed and distributed in one or more devices different from the above embodiments. The modules of the above embodiments can be combined into one module, or can be further split into multiple sub-modules.

[0114] The following describes an embodiment of the electronic device of the present invention. This electronic device can be regarded as an implementation form in physical form of the above method and device embodiments of the present invention. For the details described in the embodiment of the electronic device of the present invention, they should be regarded as a supplement to the above method or device embodiments; for the details not disclosed in the embodiment of the electronic device of the present invention, they can be implemented with reference to the above method or device embodiments.

[0115] Figure 3 It is a structural block diagram of an exemplary embodiment of an electronic device according to the present invention. Figure 3 The electronic device shown is only an example and should not impose any limitation on the functions and usage scope of the embodiments of the present invention.

[0116] As Figure 3 shown, the electronic device 300 of this exemplary embodiment is presented in the form of a general-purpose data processing device. The components of the electronic device 300 may include, but are not limited to: at least one processing unit 310, at least one storage unit 320, a bus 330 connecting different electronic device components (including the storage unit 320 and the processing unit 310), a display unit 340, etc.

[0117] Among them, the storage unit 320 stores a computer-readable program, which can be the source program or the code of a read-only program. The program can be executed by the processing unit 310, so that the processing unit 310 executes the steps of various embodiments of the present invention. For example, the processing unit 310 can execute the steps as Figure 2 shown.

[0118] The bus 330 can represent one or more of several types of bus structures, including a storage unit bus or a storage unit controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any bus structure in a variety of bus structures.

[0119] The electronic device 300 can also communicate with one or more external devices 100 (such as a keyboard, a display, a network device, a Bluetooth device, etc.), enabling a user to interact with the electronic device 300 via these external devices 100, and / or enabling the electronic device 300 to communicate with one or more other data processing devices (such as a router, a modem, etc.). Such communication can be carried out through an input / output (I / O) interface 350, and can also be through a network adapter 360 with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network). The network adapter 360 can communicate with other modules of the electronic device 300 through a bus 330.

[0120] Figure 4 is a schematic diagram of an embodiment of a computer-readable medium of the present invention. As Figure 4 shown, the computer program can be stored on one or more computer-readable media. The computer-readable medium can be a readable signal medium or a readable storage medium. When the computer program is executed by one or more data processing devices, the computer-readable medium can implement the above method of the present invention, that is: screening users who meet the incentive operation conditions; dividing the users into an experimental group and a control group, and only providing incentive operations to the users in the experimental group; collecting users in the control group who do not perform the first behavior to obtain a first user set; the incentive operation provided to the users may cause the users to perform the first behavior; determining the propensity score distribution of the first user set; screening out a second user set from the experimental group according to the propensity score distribution of the first user set; attributing the ratio difference of the users performing the first behavior between the first user set and the second user set within the same time period after providing the incentive operation to the incentive operation.

[0121] The present invention also provides a computer program product, including a computer program, which implements Figure 2 the above method when executed by a processor.

[0122] In summary, the present invention can be implemented by a method, a device, a system, an electronic device or a computer-readable medium that can execute a computer program. Some or all functions of the present invention can be implemented by using a general data processing device such as a microprocessor or a digital signal processor (DSP) in practice.

[0123] The specific embodiments described above further elaborate on the purpose, technical solution and beneficial effects of the present invention. It should be understood that the present invention is not inherently related to any specific computer, virtual device or electronic device, and various general devices can also implement the present invention. The above are only specific embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A behavioral attribution system based on propensity scores, characterized in that: The system comprises: The first screening module is used to screen users who meet the incentive operation conditions; A division module, used to divide the users into an experimental group and a control group, and provide incentive operations only to the users in the experimental group; A collection module is used to collect users in the control group who do not perform the first behavior to obtain a first user set; after providing incentive operations to the users, the users may be caused to perform the first behavior; A determination module, configured to determine a propensity score distribution of the first set of users; A second screening module, configured to screen out a second set of users from the experimental group according to the propensity score distribution of the first set of users; The attribution module is used to attribute the difference in the ratio of the first behavior performed by users in the first user set and the second user set in the same time period after the incentive operation is provided to the incentive operation.

2. The system according to claim 1, characterized in that The first screening module comprises: Sub-collection module, used for collecting user data; A first sub-screening module, used to screen out first target users with similar user data distribution; The second sub-screening module is used to screen out users who meet the incentive operation conditions from the first target users.

3. The system according to claim 1, characterized in that The determination module comprises: A first input module, used to input user data of each user in the first user set into the propensity model to obtain a propensity score of each user in the first user set; The sub-determination module is used to determine the propensity score distribution of the first user set according to the propensity score of each user in the first user set.

4. The system according to claim 1, characterized in that The second screening module comprises: The second input module is used to input the user data of each user in the experimental group into the propensity model to obtain the propensity score of each user in the experimental group; The sub-screening module is used to screen out users in the experimental group whose propensity scores are within the propensity score distribution of the first user set to obtain a second user set.

5. The system according to claim 1, characterized in that The system further comprises: The optimization module is used to optimize the incentive strategy according to the probability of each incentive operation causing the user's first behavior.

6. A behavioral attribution method based on propensity score, characterized in that: The method comprises: Screen users who meet the incentive operation conditions; Dividing the users into an experimental group and a control group, and providing incentive operations only to the users in the experimental group; Collect users who do not perform the first behavior in the control group to obtain a first user set; provide incentive operations to the users, which may cause the users to perform the first behavior; determining a propensity score distribution for the first set of users; Selecting a second set of users from the experimental group based on the propensity score distribution of the first set of users; The difference in the ratio of the first behavior performed by users in the first user set and the second user set in the same time period after the incentive operation is provided is attributed to being caused by the incentive operation.

7. The method according to claim 6, characterized in that The screening of users who meet the incentive operation conditions includes: Collect user data; Filter out the first target users with similar user data distribution; Users who meet the incentive operation conditions are screened out from the first target users.

8. The method according to claim 6, characterized in that Determining the propensity score distribution of the first set of users includes: Inputting user data of each user in the first user set into the propensity model to obtain a propensity score of each user in the first user set; The propensity score distribution of the first user set is determined according to the propensity score of each user in the first user set.

9. The method according to claim 6, characterized in that The step of selecting the second user set from the experimental group according to the propensity score distribution of the first user set includes: Input the user data of each user in the experimental group into the propensity model to obtain the propensity score of each user in the experimental group; The users in the experimental group whose propensity scores are within the propensity score distribution of the first user set are screened out to obtain a second user set.

10. The method according to claim 6, characterized in that The method further comprises: Optimize the incentive strategy based on the probability that each incentive operation will cause the user's first behavior.

11. An electronic device, comprising: processor; as well as A memory storing computer executable instructions which, when executed, cause the processor to perform a method according to any one of claims 6 to 10.

12. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 6 to 10 is implemented.

Images