CN112801690A

CN112801690A - Method and device for determining intervention characteristics

Info

Publication number: CN112801690A
Application number: CN202011624616.8A
Authority: CN
Inventors: 缪莹莹; 董越; 吴君涵; 吴卿仪
Original assignee: Beijing Didi Infinity Technology and Development Co Ltd
Current assignee: Beijing Didi Infinity Technology and Development Co Ltd
Priority date: 2020-12-31
Filing date: 2020-12-31
Publication date: 2021-05-14

Abstract

The application provides a method and a device for determining intervention characteristics, wherein the method for determining the intervention characteristics comprises the following steps: acquiring service information about multiple stages of a target dual-end service; carrying out causal analysis on the service information to select intervention characteristics having causal relation with order conversion indexes from the service information; aiming at each type of intervention characteristic, determining the influence degree of the type of intervention characteristic on service order conversion; and selecting a target intervention characteristic from the intervention characteristics according to the influence degree of the different types of intervention characteristics on the service order conversion. According to the determining method and device, the more accurate tendency characteristic influencing the user experience can be obtained.

Description

Method and device for determining intervention characteristics

Technical Field

The application relates to the technical field of double-end service, in particular to a method and a device for determining intervention characteristics.

Background

With the rapid development of internet technology, the dual-end service has been widely applied to people, whether the business value of the user can be retained to improve the transaction rate of the service order is one of the key factors for measuring whether the dual-end service can be stably developed, and the most direct quantification of the business value of the user is the user experience index, namely, the loyalty of the user. High quality user loyalty is beneficial to the facilitation of service orders.

Currently, the measure of the loyalty of the dual-end service is generally determined by using a reference factor having a correlation with the loyalty of the user, for example, when the loyalty of the user is the user retention rate, the user retention rate in the next month can be determined according to the number of active days of the user using the dual-end service in the current month. However, because reference factors influencing whether the user is loyal to the dual-end service are various, and a path of each reference factor influencing the loyalty of the user is extremely complex, the accuracy of the reference factors influencing the loyalty of the user, which are determined only according to the correlation, is not high, and in the dual-end service, the reference factors with low accuracy often cause the accuracy of a service provider recommended to a service requester to be reduced, so that the transaction rate of a service order is influenced.

Disclosure of Invention

In view of this, an object of the present application is to provide an intervention feature determining method and apparatus, which may quantitatively determine, in a cause-and-effect inference manner, an intervention feature having a cause-and-effect relationship with an order conversion index related to user experience, determine a tendency feature affecting the user experience, and screen, according to the tendency feature, service providers capable of providing a target dual-end service, so as to improve accuracy of service providers recommended to a service requester and improve a possibility of a deal of a service order.

The application mainly comprises the following aspects:

in a first aspect, an embodiment of the present application provides a method for determining an intervention feature, where the method for determining includes: acquiring service information of a plurality of stages of a target double-ended service, wherein the service information of the plurality of stages is used for representing service order execution conditions when a historical service order of the target double-ended service is executed to different stages; carrying out causal analysis on the service information to select an intervention feature having a causal relationship with an order conversion index from the service information; aiming at each type of intervention characteristic, determining the influence degree of the type of intervention characteristic on service order conversion; and selecting a target intervention characteristic from the intervention characteristics according to the influence degree of the different types of intervention characteristics on the service order conversion.

In one possible implementation, the target paired service is a travel service, and the determining method further includes:

in response to receiving a service order issued by a service requester for the travel service, screening out a service provider providing the travel service for the service requester by using the target intervention characteristic as a screening condition;

and pushing the screened service providers to the service requester.

In one possible embodiment, the target dual-end service is a takeaway delivery service, and the determining method further comprises:

in response to receiving a service request of a service requester for the takeout delivery service, screening out a service provider which provides the delivery service for the delivery service requester by using the target intervention characteristic as a screening condition;

and pushing the screened service providers to the service requester.

In a possible embodiment, the step of performing causal analysis on the service information to select an intervention feature having a causal relationship with an order conversion indicator from the service information includes:

constructing a causal relationship graph including a causal relationship representing between the service information and the order conversion index, wherein the causal relationship graph includes a plurality of nodes and a plurality of edges, two nodes having the causal relationship are connected through an edge, information corresponding to an end node for representing the edge is generated depending on information of a start node of the edge, and the plurality of nodes include nodes for representing the service information or representing the order conversion index;

acquiring service information corresponding to an initial node connected with the edge of a termination node corresponding to the order conversion index;

and determining the acquired service information as an intervention feature having a causal relationship with the order conversion index.

In a possible embodiment, the step of determining the degree of influence of the type of intervention feature on the conversion of the service order comprises:

determining a direct causal path of the order conversion index directly influenced by the intervention characteristics in the causal relationship graph;

determining an indirect causal path of the order conversion index indirectly influenced by the intervention characteristics in the causal relationship graph;

and determining the influence degree of the intervention characteristics on the service order conversion based on the weight of the edge corresponding to the direct causal path and the weight of the edge corresponding to the indirect causal path.

constructing a first initial linear model, a second initial linear model and a third initial linear model, wherein the first initial linear model represents the linear relation between the order conversion index and the corresponding constraint condition; the second initial linear model characterizes a linear relationship between the intervention feature and the corresponding constraint; the third initial linear model represents a linear relation among the intervention feature, the order conversion index and the corresponding constraint condition;

obtaining historical service information about multiple stages of a target dual-end service;

substituting the historical service information into the first initial linear model, the second initial linear model and the third initial linear model respectively to obtain a first linear model, a second linear model and a third linear model,

and determining the influence degree of the intervention characteristics on the conversion of the service order based on the model coefficients of the first linear model, the second linear model and the third linear model.

constructing a first initial nonlinear model and a second initial nonlinear model, wherein the first initial nonlinear model characterizes the nonlinear relationship between the intervention feature and the corresponding constraint condition; the second initial nonlinear model represents a nonlinear relation among the intervention feature, the corresponding constraint condition and the order conversion index;

respectively substituting the historical service information into the first initial nonlinear model and the second initial nonlinear model to obtain a first nonlinear model and a second nonlinear model;

determining the degree of influence of the type of intervention feature on the conversion of the service order based on the first nonlinear model and the second nonlinear model.

In a possible embodiment, the step of selecting a target intervention feature from the intervention features according to the degree of influence of different classes of intervention features on the conversion of the service order includes:

sequencing the influence degrees of different types of intervention characteristics on service order conversion according to the size sequence;

and selecting the intervention characteristics corresponding to the influence degree under the specified ranking as target intervention characteristics.

In one possible embodiment, the order conversion index includes any one of the following: user retention rate, user satisfaction and user net recommendation value.

In a second aspect, an embodiment of the present application further provides an apparatus for determining an intervention feature, where the apparatus for determining includes:

the system comprises an acquisition module, a storage module and a processing module, wherein the acquisition module is used for acquiring service information of multiple stages of a target double-ended service, and the service information of the multiple stages is used for representing service order execution conditions when historical service orders of the target double-ended service are executed to different stages;

the analysis module is used for carrying out causal analysis on the service information so as to select intervention characteristics with causal relation with order conversion indexes from the service information;

the determining module is used for determining the influence degree of each type of intervention characteristic on the service order conversion;

and the screening module is used for selecting the target intervention characteristics from the intervention characteristics according to the influence degree of the different types of intervention characteristics on the service order conversion.

In a possible implementation, the determining means further includes:

the service provider determining module is used for responding to the received service order issued by the service requester for the travel service, taking the target intervention characteristics as screening conditions, and screening out the service provider providing the travel service for the service requester;

and the pushing module is used for pushing the screened service providers to the service requester.

In a possible implementation, the determining means further includes:

the service provider determining module is used for responding to a service request of a service requester for the takeout delivery service, and screening out a service provider which provides the delivery service for the delivery service requester by taking the target intervention characteristic as a screening condition;

In a possible implementation, the analysis module is specifically configured to:

In a possible implementation manner, the determining module is specifically configured to:

In a possible implementation, the screening module is specifically configured to:

In a third aspect, an embodiment of the present application further provides an electronic device, including: a processor, a memory and a bus, the memory storing machine-readable instructions executable by the processor, the processor and the memory communicating via the bus when the electronic device is running, the machine-readable instructions being executable by the processor to perform the steps of the method for determining an intervention feature as set forth in the first aspect or any one of the possible embodiments of the first aspect.

In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to perform the steps of the method for determining an intervention feature as described in the first aspect or any one of the possible implementation manners of the first aspect.

According to the method for determining the intervention characteristics, when the service information of multiple stages of the target double-end service is obtained, causal analysis is conducted on the service information, the intervention characteristics with the order conversion indexes having the causal relation are selected from the service information of the multiple stages, then the influence degree of the intervention characteristics on service order conversion is determined for each type of intervention characteristics, and finally the target intervention characteristics are selected from the intervention characteristics according to the influence degrees of different types of intervention characteristics on service order conversion, so that the tendency characteristics which influence the user experience more accurately can be obtained.

In order to make the aforementioned objects, features and advantages of the present application more comprehensible, preferred embodiments accompanied with figures are described in detail below.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained from the drawings without inventive effort.

FIG. 1 illustrates a flow chart of a method of determining an intervention feature provided by an embodiment of the present application;

FIG. 2 illustrates a flow chart of steps provided by an embodiment of the present application to select an intervention feature;

FIG. 3 illustrates an example of a causal graph provided by an embodiment of the present application;

FIG. 4 is a schematic structural diagram of an intervention feature determination apparatus provided in an embodiment of the present application;

fig. 5 shows a schematic structural diagram of an electronic device provided in an embodiment of the present application.

Detailed Description

To make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it should be understood that the drawings in the present application are for illustrative and descriptive purposes only and are not used to limit the scope of protection of the present application. Additionally, it should be understood that the schematic drawings are not necessarily drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of the present application. It should be understood that the operations of the flow diagrams may be performed out of order, and that steps without logical context may be performed in reverse order or concurrently. One skilled in the art, under the guidance of this application, may add one or more other operations to, or remove one or more operations from, the flowchart.

In addition, the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. The components of the embodiments of the present application, generally described and illustrated in the figures herein, can be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present application, presented in the accompanying drawings, is not intended to limit the scope of the claimed application, but is merely representative of selected embodiments of the application. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present application without making any creative effort, shall fall within the protection scope of the present application.

In order to enable a person skilled in the art to use the present disclosure, the following embodiments are given in connection with the specific application scenario "determination of an intervention feature". It will be apparent to those skilled in the art that the general principles defined herein may be applied to other embodiments and applications without departing from the spirit and scope of the application.

The following method, apparatus, electronic device or computer-readable storage medium in the embodiments of the present application may be applied to any scenario where intervention features need to be determined, and the embodiments of the present application do not limit specific application scenarios, and any scheme using the method and apparatus for determining intervention features provided in the embodiments of the present application is within the scope of protection of the present application.

The dual-end service refers to that the cloud service platform distributes a service order issued by a service requester to a specified service provider, and the specified service provider provides services for the service requester. Here, the dual end service may include, by way of example and not limitation, services provided in an application installed on any one of the terminal devices, such as a travel service (i.e., an online taxi-taking service), a takeout delivery service, and the like, and may also be a sub-service under services provided in an integrated application installed on any one of the terminal devices, such as a car pool service, a windward service, and the like under the travel service. Here, the term "service requester" in the present application may refer to an individual, entity or tool requesting or subscribing to dual-end service. The term "service provider" in this application refers to an individual, entity or tool that can provide a dual-end service. The terms "service request" and "service order" are used interchangeably herein to refer to a request initiated by a passenger, a service requester, a driver, a service provider, a supplier, or the like, or any combination thereof. Accepting the "service request" or "service order" may be a passenger, a service requester, a driver, a service provider, a supplier, or the like, or any combination thereof. The service request may be charged or free.

Generally, the execution flow of the dual-end service is as follows: the cloud service platform establishes an incidence relation between a service provider and a service requester, receives a service order request issued by the service requester to the cloud service platform, screens a preferred service provider capable of providing services for the service requester in response to the service order request, distributes the service order to the screened preferred service provider, and provides services for the service requester by the preferred service provider so as to complete the service order.

It is noted that, before the present application, in the dual end service field, in order to increase the transaction rate of service orders, a reference factor or a combination of multiple reference factors is generally selected as a screening condition to screen out a preferred service provider providing dual end service for the service requester. Generally, the screening condition is obtained through an evaluation system model built in the cloud service platform, for example, the evaluation system model may use the obtained factors having correlation with user loyalty (i.e., user experience) as reference factors, however, since the reference factors influencing whether the user is loyalty to the dual-end service are often diverse, and a path of each reference factor influencing the user loyalty is extremely complex, the accuracy of the reference factors influencing the user loyalty determined only according to the correlation is not high, and in the dual-end service, the reference factors with low accuracy often cause the accuracy of the preferred service provider recommended to the service requester to be reduced, thereby influencing the transaction rate of the service order.

In order to solve the above problem, in the embodiment of the application, when service information of multiple stages of a target dual-end service is acquired, causal analysis is performed on the service information of the multiple stages to select intervention features having a causal relationship with an order conversion index from the service information of the multiple stages, then, for each type of intervention features, an influence degree of the type of intervention features on service order conversion is determined, and finally, a target intervention feature is selected from the intervention features according to influence degrees of different types of intervention features on service order conversion. Therefore, the intervention characteristics influencing the user experience can be quantitatively analyzed in a cause and effect inference mode, the tendency characteristics influencing the user experience are determined, and more accurate intervention characteristics are obtained.

In order to solve the above problem, in the embodiment of the application, when service information of multiple stages of target dual-end service is acquired, causal analysis is performed on the service information, so that intervention features with causal relationships in order conversion indexes are selected from the service information of the multiple stages, then, for each type of intervention features, the influence degree of the type of intervention features on service order conversion is determined, and finally, target intervention features are selected from the intervention features according to the influence degrees of different types of intervention features on service order conversion, so that a more accurate tendency feature influencing user experience can be acquired.

For the convenience of understanding of the present application, the technical solutions provided in the present application will be described in detail below with reference to specific embodiments.

Fig. 1 is a flowchart of a method for determining an intervention feature provided in an embodiment of the present application, where a device for executing the method for determining an intervention feature may be a server or a cloud service platform, and the method for determining an intervention feature in the present application is described below with the cloud service platform as an execution subject. As shown in fig. 1, the method for determining an intervention feature provided in an embodiment of the present application includes the following steps:

s101: service information of a plurality of stages of the target double-ended service is obtained, wherein the service information of the plurality of stages is used for representing service order execution conditions when historical service orders of the target double-ended service are executed to different stages.

In a specific implementation, the cloud service platform may obtain service order execution conditions of historical service orders of the target double-ended service executed within a past predetermined time period at different stages of the service order being executed. Taking the target dual-end service as the online taxi taking service as an example, the cloud service platform may obtain service order execution conditions of certain stages (for example, an answering stage, a drive receiving stage and a drive sending stage) of a historical service order of the online taxi taking service, and specifically, service information of a plurality of stages may be characterized by data generated in the different stages. In this example, the service information of the multiple phases may be data generated by the service order at different phases being executed, for example, response duration data generated at a response phase, cancellation before response data, and the like; the pickup time length data generated in the pickup stage, the cancellation data after response and the like; detour duration data, driver service scoring data and the like generated in the driving sending stage.

S102: and carrying out causal analysis on the service information to select an intervention feature having a causal relationship with the order conversion index from the service information.

It should be noted that the order conversion index is an index related to service order conversion, and is used for embodying loyalty of the user to the dual-end service, where the service order conversion may represent a completion condition of the service order, such as service order transaction, service order cancellation, service order update, and the like, and the service order conversion may be used for measuring an acceptance degree of the user to the target dual-end service, for example, as many service orders as possible are completed to indicate that the user uses the dual-end service more, and the acceptance degree of the dual-end service is higher; more service orders cancelled indicates that the user has a low level of acceptance for double-ended service or that there is a defect in service at some stage of double-ended service.

Here, the order conversion index may include, but is not limited to, any one of the following: user retention rate, user satisfaction and user net recommendation value.

In particular, the user retention rate may represent the probability that a user retains in the dual-end service, often reflecting the quality of the dual-end service and the ability to retain the user. For example, the user retention rate in the next month can be judged according to the active days of the user using the dual-session service in the current month, the high user retention rate indicates that the number of days using the dual-session service is more, so that the higher the quality of the dual-session service is reflected, the higher the capability of the user is, and the larger the deal of the corresponding service order is.

User satisfaction is used to measure user satisfaction with the interaction of dual-end service and dual-end service. User satisfaction with the dual end service may be obtained, for example, by a good rating of service orders historically completed by the service provider. The higher the satisfaction of the user is, the higher the quality of the double-end service is embodied, the higher the capability of reserving the user is, and the larger the deal of the corresponding service order is possible.

The user net recommendation value represents the likelihood that the user recommends a bi-polar service for others. For example, the net user recommendation value may be determined based on the success rate of the user sharing the installation link of the dual end service. The higher the net recommended value of the user is, the higher the quality of the double-ended service is embodied, and the higher the capability of the reserved user is, and the larger the deal of the corresponding service order is possible.

Further, the process of selecting an intervention feature having a causal relationship with the order conversion indicator will be explained below with reference to fig. 2, that is, the intervention feature having a causal relationship with the order conversion indicator is determined according to the following steps:

s201: constructing a causal graph comprising characterizing causal relationships between the service information and the order conversion indicators. The causal relationship graph comprises a plurality of nodes and a plurality of edges, two nodes with causal relationship are connected through the edges, information corresponding to the end nodes used for representing the edges is generated according to the information of the start nodes of the edges, and the nodes comprise nodes used for representing the service information or representing the order conversion indexes.

It should be noted that the causal graph including the characterization of the causal relationship between the service information and the order conversion index may be constructed in any conventional manner. For example, a causal graph is learned from data by constructing a bayesian network structure to establish a causal graph comprising causal relationships characterizing the service information and the order conversion indicators. The above-mentioned method is only exemplary, and the causal graph may also be established by other methods, and the present invention is not limited in any way herein.

S202: and acquiring service information corresponding to the starting node connected with the edge of the termination node corresponding to the order conversion index.

S203: and determining the acquired service information as an intervention feature having a causal relationship with the order conversion index.

Fig. 3 illustrates an example of a causal relationship graph provided in an embodiment of the present application, in which nodes corresponding to service information of multiple stages are a node a, a node B, a node C, a node D, a node E, and a node F, and a node corresponding to an order conversion indicator is a node Y, where an edge between two nodes indicates that a causal relationship exists between the two nodes, and a point of the edge indicates that information corresponding to an end node is generated depending on information of a start node of the edge, that is, the end node of the edge is affected by the start node of the edge. Referring to fig. 3, it can be seen that the node Y is a termination node, the nodes having causal relationship with the node Y are a node a, a node B, a node C and a node D, and the nodes having no causal relationship with the node Y are a node E and a node F. Accordingly, the service information corresponding to node a, node B, node C, and node D may be determined as an intervention feature having a causal relationship with the order conversion index, respectively.

Referring back to fig. 1, S103: and determining the influence degree of the type of the intervention characteristics on the service order conversion aiming at each type of the intervention characteristics.

In particular implementations, there are three ways to address the degree of impact of the intervention feature on the service order transformation.

In a first example, the degree of influence of this type of intervention feature on the conversion of a service order may be determined according to the following steps:

firstly, a direct causal path of the order conversion index directly influenced by the type of intervention feature in the causal graph is determined.

It should be noted that two nodes in the causal graph are connected through a directed edge, a start node in the two nodes directly affects a stop node, and a path connecting the edges of the two nodes is a direct causal path of the two nodes, that is, a direct causal path between information corresponding to the two nodes. For example, taking the case that the type of intervention feature corresponds to a node D in the causal graph shown in fig. 3, the node D as an initial node directly points to a termination node Y through an edge, the intervention feature corresponding to the node D directly affects the order conversion index, and a path connecting the node D to the edge of the node Y is a direct causal path from the node D to the node Y.

Then, an indirect causal path is determined in the causal graph, wherein the type of intervention feature indirectly affects the order conversion index.

It should be noted that, when two nodes in the causal graph are connected by at least two directed edges, a start node in the two nodes indirectly affects a stop node, and a path connecting the edges of the two nodes is an indirect causal path of the two nodes, that is, an indirect causal path of information corresponding to the two nodes. For example, taking the case that the type of intervention feature corresponds to a node D in the causal graph shown in fig. 3, where the node D as the start node directly points to a node a as the end node through an edge, and the node a as the start node directly points to a node Y as the end node through an edge, it can be seen that the intervention feature corresponding to the node D indirectly affects the order conversion index, and the paths connecting the node D to the edge of the node Y are the edge from the node D to the node a and the edge from the node a to the node Y, and then the edge from the node D to the node a and the edge from the node a to the node Y together serve as an indirect causal path from the node D to the node Y.

And finally, determining the influence degree of the intervention characteristics on the service order conversion based on the weight of the edge corresponding to the direct causal path and the weight of the edge corresponding to the indirect causal path.

It should be noted that the weight of the edge is the probability of the event corresponding to the information corresponding to the end node under the intervention of the information corresponding to the start node. For example, the weight of an edge between two nodes may be determined by quantifying data between the information to which the nodes correspond. Taking the target dual-end service as the online taxi taking service as an example, the information corresponding to the starting node is the 'pickup duration', the information corresponding to the terminating node is the 'user retention rate', and the 'user retention rate' Y is equal to the reference value Y of the user retention rate by estimating that the given 'pickup duration' is a₁The user then proceeds to determine the weight of the edge connecting the two nodes using the probability of the occurrence of the event of the online taxi-taking service.

It should be noted that after the direct causal path and the indirect causal path are obtained, the weight of the edge corresponding to the direct causal path and the weight of the edge corresponding to the indirect causal path may be directly aggregated, and the weight obtained after the aggregation is used as the influence degree of the type of intervention feature on the service order conversion. Here, since the weight of the edge is the probability of the event corresponding to the information corresponding to the end node under the intervention of the information corresponding to the start node, the weight obtained after the aggregation processing (i.e., the degree of influence of the type of intervention feature on the service order conversion) is also a quantized specific numerical value.

In a second example, the degree of influence of this type of intervention feature on the conversion of the service order is determined according to the following steps:

firstly, constructing a first initial nonlinear model and a second initial nonlinear model, wherein the first initial nonlinear model represents the nonlinear relation between the intervention feature and the corresponding constraint condition; the second initial nonlinear model characterizes a nonlinear relationship between the intervention feature, the corresponding constraint, and the order conversion indicator.

In a specific implementation, the first initial linear model, the second initial linear model and the third initial linear model are constructed as follows:

first initial linear model:

Y_i＝α_1i+β_1iX_i+ε_1i，

wherein, Y_iIndicating an order conversion index, α, for user i_1i、β_1iAnd ε_1iModel coefficients, X, representing a first initial linear model_iRepresenting constraints for user i, i representing the user.

Second initial linear model:

M_i＝α_2i+β_2iX_i+ε_2i，

wherein M is_iRepresenting an intervention feature, α, for user i_2i、β_2iAnd ε_2iModel coefficients, X, representing a second initial linear model_iRepresenting constraints for user i, i representing the user.

Third initial linear model:

Y_i＝α_3i+β_3iX_i+γM_i+ε_3i，

wherein, Y_iRepresenting an order conversion index, M, for user i_iRepresenting an intervention feature, α, for user i_3i、β_3iGamma and epsilon_3iModel coefficients, X, representing a third initial linear model_iRepresenting constraints for user i, i representing the user.

Then, historical service information is obtained for a plurality of phases of the target dual-end service.

And then, respectively substituting the historical service information into the first initial linear model, the second initial linear model and the third initial linear model to obtain a first linear model, a second linear model and a third linear model.

In a specific implementation, after the first initial linear model, the second initial linear model and the third initial linear model are constructed, historical service information of multiple stages of the target double-ended service can be obtained, the historical service information can include historical service information corresponding to each user in a target user group set using the target double-ended service, intervention features, order conversion indexes and constraint conditions are screened out from the historical service information of each user, for different service orders of any one user, the intervention feature order conversion indexes screened out from the different service orders and quantization values corresponding to the constraint conditions are respectively substituted into the first initial linear model, the second initial linear model and the third initial linear model, model coefficients of the first initial linear model, the second initial linear model and the third initial linear model for the user can be obtained by using a least square method, and respectively substituting the obtained model coefficients into the corresponding initial models to obtain real models of the first initial linear model, the second initial linear model and the third initial linear model, namely the first linear model, the second linear model and the third linear model.

Finally, the influence degree of the intervention characteristics on the conversion of the service order is determined based on the model coefficients of the first linear model, the second linear model and the third linear model.

In specific implementation, beta is obtained in the implementation process_1iAnd beta_3iIn the case of (2), β can be calculated_1i-β_3iThe influence degree of the intervention characteristic M on the service order conversion can be obtained aiming at the user i. Accordingly, the influence degree of the intervention feature M on the service order conversion for each user in the target user group set can be obtained in the same manner. And averaging the influence degrees of the intervention characteristics M on the service order conversion, which are obtained by aiming at all the users in the target user group set, so as to obtain the influence degree of the intervention characteristics M on the service order conversion, which is aiming at the target user group set.

In a third example, the degree of influence of this type of intervention feature on the conversion of the service order is determined according to the following steps:

In a specific implementation, the first initial nonlinear model and the second nonlinear model are constructed as follows:

first initial nonlinear model:

M_i(T_i)＝Pr(M_i＝1|T_i，X_i)＝logit^-1(α_4i+β_4iX_i+ε_4i)，

wherein M is_iRepresenting an intervention feature for user i, M_iIs a binary variable, M_iDenotes the intervention on user i, M_i0 denotes no intervention to user i, Pr (M)_i＝1|T_i，X_i) Expressed under the Pr function, the variable is T_i，X_iAbout M_iNon-linear model of (2), X_iRepresenting a constraint, T, for user i_iCharacteristic M indicating whether user i is intervened or not_iIntervention, T_iIs a binary variable, T_i1 denotes the user i intervened feature M_iIntervention, T_i0 denotes that user i is not subject to intervention feature M_iIntervention, i denotes user.

Second initial nonlinear model:

Y_i(T_i，M(T_i))＝Pr(Y_i＝1|T_i，M_i，X_i)＝log it^-1(α_5i+β_5iX_i+γM_i+ε_5i)，

wherein M is_iRepresenting an intervention feature for user i, M_iIs a binary variable, M_iDenotes the intervention on user i, M_i0 denotes no intervention to user i, Pr (Y)_i＝1|T_i，M_i，X_i) Expressed under the Pr function, the variable is T_i，X_i，M_iAbout Y_iNon-linear model of (2), X_iRepresenting a constraint, T, for user i_iCharacteristic M indicating whether user i is intervened or not_iIntervention, T_iIs a binary variable, T_i1 denotes the user i intervened feature M_iIntervention, T_i0 denotes that user i is not subject to intervention feature M_iIntervention, Y_iIndicates an order conversion index for user i, i indicating the user.

And then, substituting the historical service information into the first initial nonlinear model and the second initial nonlinear model respectively to obtain a first nonlinear model and a second nonlinear model.

In a specific implementation, after the first initial nonlinear model and the second initial nonlinear model are constructed, historical service information about multiple stages of the target double-ended service can be obtained, the historical service information can include historical service information corresponding to each user in a target user group set using the target double-ended service, intervention features, order conversion indexes and constraint conditions are screened from the historical service information of each user, for different service orders of any one user, quantification values corresponding to the intervention feature order conversion indexes and the constraint conditions screened from the different service orders are respectively substituted into the first initial nonlinear model and the second initial nonlinear model, model coefficients of the first initial nonlinear model and the second initial nonlinear model for the user can be obtained by using a gradient descent method, the obtained model coefficients are respectively substituted into the corresponding models, the true models of the first initial nonlinear model and the second initial nonlinear model, that is, the first nonlinear model, the second nonlinear model and the third nonlinear model, can be obtained.

Finally, the influence degree of the type of intervention characteristic on the conversion of the service order is determined based on the first nonlinear model and the second nonlinear model.

In practice, the above-mentioned implementations have beenObtains M in the process_i(T_i) And Y_i(T_i，M(T_i) In the case of delta), delta can be calculated_i(T_i)＝Y_i(T_i，M(1))-Y_i(T_iM (0)) to obtain an intervention feature M for user i_iDegree of influence on service order conversion δ_i(T_i). Accordingly, the influence degree of the intervention feature M on the service order conversion for each user in the target user group set can be obtained in the same manner. And averaging the influence degrees of the intervention characteristics M on the service order conversion, which are acquired by aiming at all the users in the target user group set, so as to acquire the influence degree of the intervention characteristics M on the service order conversion.

Referring back to fig. 1, S104: and selecting a target intervention characteristic from the intervention characteristics according to the influence degree of the different types of intervention characteristics on the service order conversion.

In specific implementation, the influence degrees of different types of intervention characteristics on service order conversion can be sorted according to the size sequence, and then the intervention characteristics corresponding to the influence degrees under the specified ranking are selected as target intervention characteristics.

For example, the intervention feature corresponding to the first influence degree in the sorting of the influence degrees of different types of intervention features on the service order conversion according to the size order can be used as the target intervention feature.

In the embodiment of the application, when the service information of multiple stages of the target dual-end service is obtained, causal analysis is conducted on the service information, so that intervention characteristics with causal relation of order conversion indexes are selected from the service information of the multiple stages, then the influence degree of the intervention characteristics on service order conversion is determined for each type of intervention characteristics, and finally the target intervention characteristics are selected from the intervention characteristics according to the influence degree of different types of intervention characteristics on service order conversion, so that the tendency characteristics influencing user experience can be obtained more accurately.

Furthermore, after selecting the target intervention feature, in a preferred example of the present application, in addition to S101, S102, S103 and S104, after S104, in case the target dual end service is a travel service, the method of determining the intervention feature additionally comprises the following steps (not shown in fig. 1):

step (A): and in response to receiving a service order issued by a service requester for the travel service, screening out a service provider providing the travel service for the service requester by taking the target intervention characteristic as a screening condition.

In specific implementation, if the service requester sends a travel service request for travel service to the cloud service platform, after receiving the travel service request, the cloud service platform will screen the service providers capable of providing travel service by using the selected target intervention characteristics as screening conditions, and screen out the most preferred service providers capable of providing travel service for the service requester.

Step (B): and pushing the screened service providers to the service requester.

In implementation, the screened service provider may be pushed to the service requester to complete the service order.

For example, when the travel service is an online taxi taking service and the order conversion index is a user retention rate, the cloud service platform may acquire service information about multiple stages of the online taxi taking service, such as response duration data generated in a response stage, cancellation data before response, and the like; the pickup time length data generated in the pickup stage, the cancellation data after response and the like; detour duration data and driver service scoring data generated in the driving sending stage. The cloud service platform carries out causal analysis on the service information, selects an intervention characteristic 'response time length' and 'pickup time length' having a causal relationship with the 'user retention rate', determines the influence degree of the 'response time length' on the 'user retention rate' and the influence degree of the 'pickup time length' on the 'user retention rate', wherein the influence degree of the 'response time length' on the 'user retention rate' is greater than the influence degree of the 'pickup time length' on the 'user retention rate', and at the moment, the 'response time length' can be determined as a target interference characteristic. The terminal device (service requester) responds to a taxi taking interface of a travel application installed on the terminal device by a passenger to input information of a departure place, a destination, departure time, the number of travelers and the like to generate a service order, and sends the service order to the cloud service platform, the cloud service platform responds to the service order to screen a network appointment (namely, a service provider) capable of completing the service order, and in the screening process, the network appointment capable of completing the service order can be screened by taking the 'response time length' as one of screening conditions, for example, the network appointment with the shortest response time length can be taken as the most preferable network appointment providing travel service. Meanwhile, the cloud service platform can push the network appointment car with the shortest response time to the passenger to complete the travel order.

Furthermore, after selecting the target intervention feature, in another preferred example of the present application, the method for determining the intervention feature, in addition to comprising S101, S102, S103 and S104, additionally comprises, after S104, in case the target dual end service is a takeaway delivery service, the following steps (not shown in fig. 1):

step (a): and in response to receiving a service request of a service requester for the takeout delivery service, screening out a service provider which provides the delivery service for the delivery service requester by using the target intervention characteristic as a screening condition.

In specific implementation, if the service request sends a trip service request for the takeout delivery service to the cloud service platform, after receiving the takeout delivery request, the cloud service platform will screen the service providers capable of providing the takeout delivery service by using the selected target intervention characteristics as screening conditions, and screen out the most preferred service providers capable of providing the takeout delivery service for the service request.

Step (b): and pushing the screened service providers to the service requester.

For example, when the travel service is a takeout delivery service and the order conversion index is a user retention rate, the cloud service platform may acquire service information about multiple stages of the takeout delivery service, for example, response duration data generated in a response stage, cancellation data before response, and the like; order receiving duration data generated in the order receiving stage, cancellation data after response, distance from a merchant and the like; the order sending duration data and the data of the service scores of the takeout personnel are generated in the order sending stage. The cloud service platform carries out causal analysis on the service information, selects intervention characteristics 'order receiving duration' and 'cancel after response' which have causal relation with 'user retention rate', determines the influence degree of the 'order receiving duration' on the 'user retention rate' and the influence degree of the 'cancel after response' on the 'user retention rate', wherein the influence degree of the 'order receiving duration' on the 'user retention rate' is larger than the influence degree of the 'cancel after response' on the 'user retention rate', and at the moment, the 'order receiving duration' can be determined as a target interference characteristic. The terminal device (service requester) pays an order to generate a service order in response to information such as a selected commodity, a delivery location and delivery time input by a user under a take-out order interface in a take-out application installed on the terminal device, and sends the service order to a cloud service platform, the cloud service platform screens a take-out distributor (i.e., a service provider) capable of completing the service order in response to the service order, and in the screening process, the take-out distributor capable of completing the service order can be screened by using the "take-out duration" as one of the screening conditions, for example, the take-out distributor with the shortest "take-out duration" can be used as the most preferable take-out distributor for providing travel services. Meanwhile, the cloud service platform can push the takeout deliverer with the shortest order taking time to the user to complete the travel order.

By the method, the intervention characteristic that the order conversion index related to the user experience has the causal relationship can be quantitatively determined in a causal inference mode, the tendency characteristic influencing the user experience is determined, and the service providers capable of providing the target double-ended service are screened according to the tendency characteristic, so that the accuracy of the service providers recommended to the service requester is improved, and the possibility of the deal of the service orders is improved.

Based on the same application concept, the embodiment of the present application further provides a device for determining an intervention feature corresponding to the method for determining an intervention feature provided by the above embodiment, and since the principle of solving the problem of the device in the embodiment of the present application is similar to the method for determining an intervention feature in the above embodiment of the present application, the implementation of the device may refer to the implementation of the method, and repeated details are omitted.

Fig. 4 shows a schematic structural diagram of an intervention feature determination apparatus provided in an embodiment of the present application.

As shown in fig. 4, the determination device 400 includes:

the obtaining module 401 obtains service information about multiple stages of the target dual-end service, where the service information of the multiple stages is used to characterize service order execution conditions when a historical service order of the target dual-end service is executed to different stages.

In particular implementations, the obtaining module 401 may obtain service order executions of historical service orders of the target double-ended service executed within a predetermined period of time in different stages of the service order being executed.

And the analysis module 402 is used for performing causal analysis on the service information so as to select an intervention feature having a causal relationship with the order conversion index from the service information.

In one possible implementation, the order conversion index may include, but is not limited to, any one of the following: user retention rate, user satisfaction and user net recommendation value.

In a possible implementation, the analysis module 402 is specifically configured to:

and constructing a causal relationship graph comprising a causal relationship representing between the service information and the order conversion index, wherein the causal relationship graph comprises a plurality of nodes and a plurality of edges, two nodes with causal relationship are connected through an edge, the point of the edge is generated according to the information of the starting node of the edge, and the point of the edge points to the end node for representing the edge corresponds to the information which is generated according to the information of the starting node of the edge, and the plurality of nodes comprise nodes for representing the service information or the order conversion index.

And acquiring service information corresponding to the starting node connected with the edge of the termination node corresponding to the order conversion index.

The determining module 403 determines, for each type of intervention feature, a degree of influence of the type of intervention feature on the service order conversion.

In a possible implementation manner, the determining module 403 is specifically configured to:

and determining a direct causal path of the type of intervention characteristic in the causal graph directly influencing the order conversion index.

It should be noted that two nodes in the causal graph are connected through a directed edge, a start node in the two nodes directly affects a stop node, and a path connecting the edges of the two nodes is a direct causal path of the two nodes, that is, a direct causal path between information corresponding to the two nodes.

And determining an indirect causal path of the order conversion index indirectly influenced by the intervention characteristics in the causal relationship graph.

It should be noted that, when two nodes in the causal graph are connected by at least two directed edges, a start node in the two nodes indirectly affects a stop node, and a path connecting the edges of the two nodes is an indirect causal path of the two nodes, that is, an indirect causal path of information corresponding to the two nodes.

It should be noted that the weight of the edge is the probability of the event corresponding to the information corresponding to the end node under the intervention of the information corresponding to the start node. After the direct causal path and the indirect causal path are obtained, the weight of the edge corresponding to the direct causal path and the weight of the edge corresponding to the indirect causal path may be directly aggregated, and the weight obtained after aggregation is used as the influence degree of the type of intervention feature on the service order conversion. Here, since the weight of the edge is the probability of the event corresponding to the information corresponding to the end node under the intervention of the information corresponding to the start node, the weight obtained after the aggregation processing (i.e., the degree of influence of the type of intervention feature on the service order conversion) is also a quantized specific numerical value.

constructing a first initial linear model, a second initial linear model and a third initial linear model, wherein the first initial linear model represents the linear relation between the order conversion index and the corresponding constraint condition; the second initial linear model characterizes a linear relationship between the intervention feature and the corresponding constraint; the third initial linear model characterizes a linear relationship between the intervention feature, the order conversion indicator, and the corresponding constraint.

And the screening module 404 selects a target intervention feature from the intervention features according to the influence degree of the different types of intervention features on the service order conversion.

In a possible implementation manner, the screening module 404 is specifically configured to:

In a possible implementation, the determining apparatus 400 further includes:

In the embodiment of the application, when the service information of multiple stages of the target dual-end service is obtained, causal analysis is performed on the service information, so that intervention characteristics with causal relations of order conversion indexes are selected from the service information of the multiple stages, then the influence degree of the intervention characteristics on service order conversion is determined for each type of intervention characteristics, and finally the target intervention characteristics are selected from the intervention characteristics according to the influence degree of different types of intervention characteristics on service order conversion, so that the tendency characteristics influencing user experience can be obtained more accurately.

Based on the same application concept, referring to fig. 5, a schematic structural diagram of an electronic device 500 provided in the embodiment of the present application includes: a processor 510, a memory 520 and a bus 530, the memory 520 storing machine-readable instructions executable by the processor 510, the processor 510 and the memory 520 communicating via the bus 530 when the electronic device 500 is operating, the machine-readable instructions being executable by the processor 510 to perform the steps of the method of determining an intervention feature as described in any of the above embodiments.

In particular, the machine readable instructions, when executed by the processor 510, may perform the following:

acquiring service information of a plurality of stages of a target double-ended service, wherein the service information of the plurality of stages is used for representing service order execution conditions when a historical service order of the target double-ended service is executed to different stages;

carrying out causal analysis on the service information to select an intervention feature having a causal relationship with an order conversion index from the service information;

aiming at each type of intervention characteristic, determining the influence degree of the type of intervention characteristic on service order conversion;

and selecting a target intervention characteristic from the intervention characteristics according to the influence degree of the different types of intervention characteristics on the service order conversion.

Further, the machine readable instructions, when executed by the processor 510, may perform the following:

and pushing the screened service providers to the service requester.

Based on the same application concept, the embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and the computer program is executed by a processor to perform the steps of the method for determining the intervention feature provided by the above embodiment.

Specifically, the storage medium can be a general storage medium, such as a mobile disk, a hard disk, or the like, and when a computer program on the storage medium is executed, the method for determining the intervention feature can be executed, the intervention feature affecting the user experience is quantitatively analyzed in a cause-and-effect inference manner, a tendency feature affecting the user experience is determined, and an operation strategy of the target dual-session service is guided according to the tendency feature, so as to achieve benign development of the target dual-session service.

It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.

In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and there may be other divisions when actually implemented, and for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of devices or units through some communication interfaces, and may be in an electrical, mechanical or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit.

The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a non-volatile computer-readable storage medium executable by a processor. Based on such understanding, the technical solution of the present application or portions thereof that substantially contribute to the prior art may be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.

Finally, it should be noted that: the above-mentioned embodiments are only specific embodiments of the present application, and are used for illustrating the technical solutions of the present application, but not limiting the same, and the scope of the present application is not limited thereto, and although the present application is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: any person skilled in the art can modify or easily conceive the technical solutions described in the foregoing embodiments or equivalent substitutes for some technical features within the technical scope disclosed in the present application; such modifications, changes or substitutions do not depart from the spirit and scope of the exemplary embodiments of the present application, and are intended to be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A method for determining an intervention feature, comprising:

2. The determination method of claim 1, wherein the target paired end service is a travel service, the determination method further comprising:

and pushing the screened service providers to the service requester.

3. The method of determining of claim 1, wherein the target dual end service is a takeaway delivery service, the method of determining further comprising:

and pushing the screened service providers to the service requester.

4. The method of claim 1, wherein the step of causally analyzing the service information to select an intervention feature from the service information having a causal relationship to an order conversion indicator comprises:

5. The method of determining of claim 4, wherein the step of determining the extent of the impact of the type of intervention feature on the conversion of the service order comprises:

6. The method of determining of claim 1, wherein the step of determining the extent of the impact of the type of intervention feature on the conversion of the service order comprises:

7. The method of determining of claim 1, wherein the step of determining the extent of the impact of the type of intervention feature on the conversion of the service order comprises:

8. The method of claim 1, wherein the step of selecting a target intervention feature from the intervention features based on the extent to which the different classes of intervention features affect the conversion of the service order comprises:

9. The determination method of claim 1, wherein the order conversion index comprises any one of: user retention rate, user satisfaction and user net recommendation value.

10. An intervention feature determination apparatus, comprising:

11. An electronic device, comprising: a processor, a memory and a bus, the memory storing machine-readable instructions executable by the processor, the processor and the memory communicating over the bus when the electronic device is operating, the processor executing the machine-readable instructions to perform the steps of the method of any one of claims 1 to 9.

12. A computer-readable storage medium, having stored thereon a computer program which, when being executed by a processor, is adapted to carry out the steps of the method according to any one of claims 1 to 9.

13. A computer program product comprising a computer program or instructions, characterized in that the computer program or instructions, when executed by a processor, implement the steps of the method of any one of claims 1 to 9.