Recommendation strategy experiment method and device and computer equipment

Abstract

The embodiment of the invention provides a recommendation strategy experiment method and device and computer equipment, and the method comprises the steps: receiving a commodity recommendation request related to an online shop, and the commodity recommendation request carries a distribution parameter; the commodity recommendation request is responded, a target experiment is determined from a plurality of preset recommendation strategy experiments according to the distribution parameters, and the target experiment corresponds to a target recommendation strategy; and executing the target experiment to perform commodity recommendation based on the target recommendation strategy. The experiment efficiency can be effectively improved.

Description

Technical Field

[0001] The present application belongs to the field of data processing and relates to e-commerce information processing technology, and in particular to a recommendation strategy experimental method and its device and computer equipment. Background Art

[0002] At present, cross-border e-commerce is developing rapidly. When recommending products for various online stores, the same recommendation strategy has different recommendation effects in different online stores. Some recommendation effects are better, while others are worse. In order to find suitable recommendation strategies for online stores through experiments, the relevant technology usually selects the experimental configuration through the ABTest system. The ABTest system is an experimental system that can compare the effects of two different versions of systems, pages, advertisements, etc. The ABTest system is based on assumptions and can be in two independent experimental groups, for example, Group A receives version A and Group B receives version B, and then collects and analyzes the two sets of data to determine which version is more effective.

[0003] If you want to add a new experiment in the ABTest system, you need to upload the new experiment identifier and parse the experiment configuration in the code, and customize the recommendation strategy based on the parsed configuration content. This experimental method requires more development content, resulting in a large development workload and low experimental efficiency. Summary of the invention

[0004] The embodiments of the present application provide a recommendation strategy experiment method and its device, and computer equipment, which can solve the technical problems of low experimental efficiency and large development workload.

[0005] A first aspect of an embodiment of the present application provides a recommendation strategy experiment method, comprising: receiving a product recommendation request associated with an online store; determining a target experiment from a plurality of preset recommendation strategy experiments based on a diversion parameter in the product recommendation request, the target experiment corresponding to a target recommendation strategy; and executing the target experiment to make product recommendations based on the target recommendation strategy.

[0006] According to an embodiment of the present application, the method further includes: presetting the multiple recommendation strategy experiments, and configuring a corresponding recommendation strategy for each of the multiple recommendation strategy experiments.

[0007] According to an embodiment of the present application, the presetting of the multiple recommendation strategy experiments and configuring the corresponding recommendation strategy for each of the multiple recommendation strategy experiments includes: providing a first number of ordinary experiments, each of the first number of ordinary experiments including an experiment code and a configuration file required to execute the experiment code; if no request to add a custom experiment is received, taking each of the first number of ordinary experiments as a recommendation strategy experiment; or if a request to add a custom experiment is received, responding to the request to add a custom experiment, binding a second number of custom experiments to the specified online store, each of the second number of custom experiments including an experiment code and a configuration file required to execute the experiment code; and taking each of the first number of ordinary experiments as a recommendation strategy experiment, and taking each of the second number of custom experiments as a recommendation strategy experiment.

[0008] According to an embodiment of the present application, the configuration file of any common experiment includes the identification ID, traffic bucket ID, and recommended strategy of any common experiment; the configuration file corresponding to any custom experiment includes the ID, traffic bucket ID, recommended strategy, and the ID of the designated online store.

[0009] According to an embodiment of the present application, the online store runs in an independent station server of an e-commerce platform, and the diversion parameters include the ID of the online store and the ID of the user accessing the online store.

[0010] According to an embodiment of the present application, the target experiment is determined from a plurality of preset recommendation strategy experiments based on the diversion parameters in the product recommendation request, and the target experiment corresponds to the target recommendation strategy, including: determining whether the online store is bound to at least one custom experiment based on the ID of the online store; if the online store is not bound to any custom experiment, determining the target experiment from the first number of ordinary experiments based on the ID of the user; or if the online store is bound to at least one custom experiment, determining the target experiment from at least one custom experiment bound to the online store based on the ID of the user.

[0011] According to an embodiment of the present application, determining the target experiment from the first number of ordinary experiments based on the user's ID includes: calculating a hash code of the user's ID; obtaining a remainder by performing a modulo operation on the hash code and the number of all traffic buckets, and using the remainder as the ID of the target traffic bucket; and using the ordinary experiment corresponding to the ID of the target traffic bucket in the first number of ordinary experiments as the target experiment.

[0012] According to an embodiment of the present application, determining the target experiment from at least one custom experiment bound to the online store based on the user's ID includes: calculating a hash code of the user's ID; obtaining a remainder by performing a modulo operation on the hash code and the number of all traffic buckets, and using the remainder as the ID of the target traffic bucket; and using the experiment corresponding to the ID of the target traffic bucket in at least one custom experiment bound to the online store as the target experiment.

[0013] A second aspect of an embodiment of the present application provides a recommendation strategy experiment device, including: a receiving module for receiving a product recommendation request associated with an online store; a response module for determining a target experiment from a plurality of preset recommendation strategy experiments based on a diversion parameter in the product recommendation request, the target experiment corresponding to a target recommendation strategy; and an execution module for executing the target experiment to make product recommendations based on the target recommendation strategy.

[0014] A third aspect of an embodiment of the present application provides a computer device, comprising: a memory, and a processor, wherein the processor executes computer-readable instructions stored in the memory to implement the recommendation strategy experimental method.

[0015] The recommendation strategy experiment method provided in the embodiment of the present application can effectively improve the experiment efficiency by determining the target experiment from multiple preset recommendation strategy experiments according to the diversion parameters in the product recommendation request associated with the online store when any user visits the online store, and executing the target experiment to recommend products to the user based on the target recommendation strategy corresponding to the target experiment. In addition, the recommendation strategy experiment method provided in the embodiment of the present application can also perform recommendation strategy experiments on online stores in combination with the store dimension, realize customized experiments on online stores with specific needs, and expand the diversity of recommendation strategy experiment requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 A network architecture adopted for an exemplary application scenario provided in an embodiment of the present application.

[0018] Figure 2 A schematic diagram of the flow chart of the recommended strategy experimental method provided in the embodiments of the present application.

[0019] Figure 3A flowchart of an extended embodiment of the recommendation strategy experimental method provided in the embodiment of the present application.

[0020] Figure 4 A flowchart of an experiment of preset multiple recommendation strategies provided in an embodiment of the present application.

[0021] Figure 5 This is an example of an experiment in which traffic buckets are allocated to multiple recommended strategies provided in an embodiment of the present application.

[0022] Figure 6 A schematic diagram of a flow chart of an experiment for determining a target based on diversion parameters provided in an embodiment of the present application.

[0023] Figure 7 A schematic diagram of a process for determining a target experiment from a common experiment based on diversion parameters provided in an embodiment of the present application.

[0024] Figure 8 A schematic diagram of a process for determining the target experiment from a custom experiment according to diversion parameters provided in an embodiment of the present application.

[0025] Fig. 9 A functional block diagram of a recommendation strategy experimental device provided in an embodiment of the present application.

[0026] Fig.10 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application is described in detail below with reference to the accompanying drawings and specific embodiments.

[0028] It should be noted that in this application, "at least one" means one or more, and "more than one" means two or more than two. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. The terms "first", "second", "third", "fourth", etc. (if any) in the specification, claims and drawings of this application are used to distinguish similar objects, rather than to describe a specific order or sequence.

[0029] In the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a concrete way. The following embodiments and features in the embodiments may be combined with each other without conflict.

[0030] See also Figure 1 , which is a network architecture adopted in an exemplary application scenario provided in an embodiment of the present application. Figure 1 The network architecture includes an application server 81 and multiple independent station servers 82 ( Figure 1 Only one independent station server 82) and a terminal device 83 are illustrated. The application server 81 can be used to deploy a computer program product (e.g., software code, computer-readable instructions, etc.) programmed and implemented according to the recommendation strategy experiment method provided in an embodiment of the present application, thereby providing a recommendation strategy experiment service. Each independent station server 82 can deploy an online store that provides e-commerce services. Users of the terminal device 83 can access the pages of the online store. The application server 81 can provide the user with a recommendation strategy experiment service based on the computer program product deployed by the application server 81 when the user accesses the product recommendation page of the online store.

[0031] Figure 1 The application scenario is only an example of platform deployment. In other exemplary embodiments, the computer program product implementing the recommendation strategy experiment method of the embodiment of the present application can also be run on any computer device with sufficient computing power (for example Fig.10 In the computer device shown in the figure), each step of the recommendation strategy experiment method is executed to provide the recommendation strategy experiment service.

[0032] Based on the principles disclosed above, the following description will be given by taking the computer program product running on the application server 81 as an example. Figure 2 As shown, it is a flow chart of the recommendation strategy experiment method provided in an embodiment of the present application. The recommendation strategy experiment method is applied to the application server 81. In an embodiment of the present application, the following steps are included:

[0033] Step S101: receiving a product recommendation request associated with an online store.

[0034] In an application scenario provided in an embodiment of the present application, the online store can be run in an independent station server 82 of an e-commerce platform. The e-commerce platform can be a cross-border e-commerce platform, which allows each online store to run in a corresponding independent station server 82 to provide e-commerce services to public users. Public users can visit the page of each online store, browse product information and select corresponding products.

[0035] In one embodiment of the present application, when a user visits any online store of the e-commerce platform, the independent station server 82 running any online store will send a product recommendation request to the application server 81 providing the recommendation strategy experiment service.

[0036] In one embodiment of the present application, the product recommendation request carries a diversion parameter.

[0037] In an embodiment of the present application, the diversion parameter includes the user identification (userID) of the user.

[0038] In one embodiment of the present application, the diversion parameter may also include an identification (ID) of the online store. Each online store on the e-commerce platform corresponds to an ID that can uniquely identify the online store, so that different online stores on the same e-commerce platform can be distinguished according to the ID of the online store.

[0039] Step S102: determining a target experiment from a plurality of preset recommendation strategy experiments according to the diversion parameters in the product recommendation request, wherein the target experiment corresponds to a target recommendation strategy.

[0040] In one embodiment of the present application, each of the preset multiple recommendation strategy experiments corresponds to a configuration of a recommendation strategy. The multiple recommendation strategy experiments correspond to multiple recommendation strategies. The recommendation strategies corresponding to the preset multiple recommendation strategy experiments are different from each other. The recommendation strategy corresponding to the target experiment is used as the target recommendation strategy.

[0041] In one embodiment of the present application, the recommendation strategy mainly includes the recommendation service logic and the corresponding recommendation configuration. In one embodiment of the present application, the recommendation strategy experiment performs the recall of the product information of the products of the online store according to the business logic of the recommendation strategy, obtains the product information and score of each recalled product, merges the product information of all recalled products, and sorts all products according to the score to obtain a product recommendation list.

[0042] Step S103: Execute the target experiment to make product recommendations based on the target recommendation strategy.

[0043] In one embodiment of the present application, the application server 81 executes a target experiment and displays the product recommendation list obtained according to the target recommendation strategy on the graphical user interface corresponding to the online store, thereby making product recommendations for users.

[0044] In one embodiment of the present application, the recommendation strategy experiment method further includes: collecting user behavior data under the target recommendation strategy. User behavior data includes, but is not limited to: data of users browsing, collecting and / or purchasing products from the online store. User behavior data can be used for subsequent analysis of which recommendation strategy is the optimal recommendation strategy.

[0045] Since different users carry different diversion parameters in the product recommendation requests when they visit the online store, the target experiments determined from the preset multiple recommendation strategy experiments are also different. Therefore, when a large number of different users visit the online store, each of the preset multiple recommendation strategy experiments will be executed on the online store, and the user behavior data under the recommendation strategy corresponding to each recommendation strategy experiment will be collected. Furthermore, the embodiment of the present application can also determine the optimal recommendation strategy from the multiple recommendation strategies corresponding to the preset multiple recommendation strategy experiments based on the historical user behavior data under the recommendation strategy corresponding to each recommendation strategy experiment.

[0046] Reference Figure 3 , which is a flow chart of an extended embodiment of the recommendation strategy experiment method provided in the embodiment of the present application. The recommendation strategy experiment method, before receiving a product recommendation request associated with an online store in step S101, further includes:

[0047] Step S100: Preset a plurality of recommendation strategy experiments, and configure a corresponding recommendation strategy for each of the plurality of recommendation strategy experiments.

[0048] Based on any embodiment of this application, please refer to Figure 4 In the process shown, step S100 includes:

[0049] Step S1001: Provide a first number of common experiments, each of the first number of common experiments comprising an experiment code and a configuration file required to execute the experiment code.

[0050] In one embodiment of the present application, the configuration file of any common experiment includes the ID of the common experiment, the ID of the traffic bucket, and the recommended strategy. The recommended strategy of any common experiment refers to the recommended strategy configured for the common experiment.

[0051] In one embodiment, the value of the first number is a positive integer greater than or equal to 1. The recommendation strategies corresponding to the first number of common experiments are different from each other.

[0052] In one embodiment of the present application, the application server 81 allocates a preset number of traffic buckets (for example, 10 traffic buckets, that is, the number of all traffic buckets is 10) to the first number of ordinary experiments, each ordinary experiment is allocated to at least one traffic bucket, and the traffic buckets corresponding to the first number of ordinary experiments are different.

[0053] For example, see Figure 5 As shown, assuming that the value of the first quantity is 5, it indicates that five common experiments are provided, and the corresponding IDs of the five common experiments are 100001, 100002, 100003, 100004, and 100005. The recommended strategies configured for the five common experiments are s1, s2, s3, s4, and s5, respectively, and the application server 81 allocates 10 traffic buckets to the five common experiments. For example, the traffic buckets allocated to the common experiment with ID 100001 have IDs 1 and 2; the traffic buckets allocated to the common experiment with ID 100002 have IDs 3 and 4; the traffic buckets allocated to the common experiment with ID 100003 have IDs 5 and 6; the traffic buckets allocated to the common experiment with ID 100004 have IDs 7 and 8; and the traffic buckets allocated to the common experiment with ID 100005 have IDs 9 and 0.

[0054] Step S1002: Determine whether a request for adding a custom experiment is received. If no request for adding a custom experiment is received, execute step S1003; if a request for adding a custom experiment is received, execute step S1004.

[0055] In one embodiment of the present application, in order to further expand the functionality of the present application, further custom experiments may be added in response to a developer's request to add a custom experiment.

[0056] For example, in Figure 5 An "add" button (not shown in the figure) is displayed on the interactive interface shown. When an input signal of clicking the "add" button is received, it can be determined that a request to add a custom experiment is received.

[0057] Step S1003: If no request for adding a custom experiment is received, each common experiment in the first number of common experiments is used as a recommended strategy experiment.

[0058] Still Figure 5 Take the five common experiments shown as an example, and use these five common experiments as recommended strategy experiments.

[0059] In one embodiment of the present application, the experimental code of each common experiment in the first number of common experiments can be deployed to the service container of the application server 81, and the configuration file of each common experiment can be deployed to other servers outside the application server 81. The data processing efficiency can be improved by separately deploying the experimental code and the configuration file of each common experiment. In addition, by deploying the experimental code of each common experiment to the service container, hot loading of the experimental code can be achieved when the experimental code needs to be executed.

[0060] For example, a service container may refer to a business code loader implemented by creating a custom classloader, which can load any business code jar package and expose an application programming interface (API) for the caller to choose to use a certain business logic.

[0061] Step S1004: If a request to add a custom experiment is received, respond to the request to add a custom experiment and bind a second number of custom experiments to the specified online store, wherein each custom experiment in the second number of custom experiments includes an experiment code and a configuration file required to execute the experiment code.

[0062] In one embodiment of the present application, the configuration file corresponding to any custom experiment bound to the specified online store includes the ID of the any custom experiment, the ID of the traffic bucket, the recommended strategy, and the ID of the specified online store. The recommended strategy of any custom experiment refers to the recommended strategy configured for the any custom experiment.

[0063] In an embodiment of the present application, the value of the second number may be a positive integer greater than or equal to 1, and the recommendation strategies corresponding to the second number of customized experiments are different from each other.

[0064] In one embodiment of the present application, the application server 81 allocates the preset number of traffic buckets to the second number of custom experiments bound to the designated online store. That is, the total number of traffic buckets allocated to the first number of common experiments is equal to the total number of traffic buckets allocated to the second number of custom experiments bound to the designated online store. In this embodiment, each custom experiment is allocated to at least one traffic bucket, and the traffic buckets corresponding to the second number of custom experiments are different from each other.

[0065] For example, see Figure 5As shown in the figure, three custom experiments are bound to the online store with ID 1677209962375, and the IDs of these three custom experiments are 10008, 10009, and 100010. Among them, the IDs of the traffic buckets assigned to the custom experiment with ID 10008 are 1, 2, 3, and 4; the IDs of the traffic buckets assigned to the custom experiment with ID 10009 are 5, 6, 7, 8, and 9; and the ID of the traffic bucket assigned to the custom experiment with ID 100010 is 0.

[0066] In an embodiment of the present application, if the second number is equal to 1, it indicates that only one custom experiment is bound to the designated online store, and the application server 81 allocates all traffic buckets to the custom experiment. Figure 5 As shown, the online store with ID 1670734596235 is bound to only one custom experiment, whose ID is 100011, and the application server 81 allocates all traffic buckets (e.g., 10 traffic buckets, whose IDs are 0-9) to the custom experiment with ID 100011. Compared with ordinary experiments, the custom experiment can refer to an experiment that optimizes a parameter of the recommendation strategy corresponding to the ordinary experiment. For example, for a store selling clothing, the clothing category of the store can be optimized.

[0067] In an embodiment of the present application, the ABTest system can be used to create each experiment, including common experiments and custom experiments. The ABTest system is used to respond to input from developers and create the experimental code and corresponding configuration files for each experiment.

[0068] Step S1005: taking each common experiment in the first number of common experiments as a recommended strategy experiment, and taking each custom experiment in the second number of custom experiments bound to the designated online store as a recommended strategy experiment.

[0069] In one embodiment of the present application, the experimental code of each common experiment in the first number of common experiments and the experimental code of each custom experiment in the second number of custom experiments bound to the specified online store can be deployed to the service container of the application server 81, and the configuration file of each common experiment and the configuration file of each custom experiment can be deployed to other servers outside the application server 81. The data processing efficiency can be improved by deploying the experimental code and the configuration file separately. In addition, by deploying the experimental code to the service container, hot loading of the experimental code can be achieved when the experimental code needs to be executed.

[0070] It can be seen from the above process steps of the present application that by repeatedly executing steps S1004 and S1005, one or more custom experiments can be bound to multiple online stores. In one embodiment of the present application, the number of custom experiments bound to different online stores in the multiple online stores can be different. For example, see Figure 5 As shown, two custom experiments are bound to the online store with ID 1677209962373, and their corresponding IDs are 100006 and 100007 respectively.

[0071] In one embodiment of the present application, a configuration listener can also be set, which is used to monitor changes in the configuration files. When a change in any configuration file is detected, the experimental code corresponding to the configuration file is updated accordingly according to the change in the configuration file, thereby achieving that the configuration can take effect in real time after the configuration file is updated without re-issuing the code.

[0072] For example, when the application server 81 responds to the developer's input signal and swaps the IDs of the traffic buckets of two of the five common experiments, the IDs of the traffic buckets recorded in the configuration files corresponding to the two common experiments will change accordingly. Therefore, the application server 81 will update the values ​​of the traffic bucket IDs in the experimental codes corresponding to the two configuration files accordingly, thereby achieving real-time configuration effectiveness after the configuration files are updated without the need to republish the code.

[0073] For example, see Figure 5 As shown, assume that the traffic buckets 1 and 2 of the common experiment with ID 100001 are adjusted to the traffic buckets of the common experiment with ID 100002, and the traffic buckets 3 and 4 of the common experiment with ID 100002 are adjusted to the traffic buckets of the common experiment with ID 100001. Then the IDs of the traffic buckets recorded in the configuration files corresponding to the two common experiments will change accordingly. At this time, the values ​​of the IDs of the traffic buckets in the experimental codes corresponding to the two configuration files can be updated accordingly.

[0074] See also Figure 6 As shown, it is a schematic diagram of a process for determining a target experiment according to a diversion parameter provided in an embodiment of the present application, wherein step S102 includes:

[0075] Step S1021: Determine whether the online store is bound to at least one custom experiment according to the ID of the online store. If the online store is not bound to any custom experiment, execute step S1022; if the online store is bound to at least one custom experiment, execute step S1023.

[0076] As in the above embodiment, the configuration file of the custom experiment records the ID of the online store bound to the custom experiment. Therefore, by searching the configuration file of the custom experiment according to the ID of the online store, it can be determined whether there is a bound custom experiment in the online store.

[0077] Step S1022: Determine the target experiment from the first number of common experiments according to the ID of the user.

[0078] Step S1023: Determine the target experiment from at least one custom experiment bound to the online store according to the user ID.

[0079] See also Figure 7 As shown, it is a schematic diagram of a process for determining a target experiment from a common experiment according to a diversion parameter provided by an embodiment of the present application, wherein step S1022 includes:

[0080] Step S10221: Calculate the hash code of the user's ID.

[0081] Step S10222: obtain a remainder by performing a modulo operation on the hash code and the preset number (ie, the number of all traffic buckets), and use the remainder as the ID of the target traffic bucket.

[0082] Step S10223: The common experiment corresponding to the ID of the target traffic bucket among the first number of common experiments is used as the target experiment. For example, assuming the calculated remainder is 2, the common experiment with ID 100001 corresponding to the traffic bucket with ID 2 is determined as the target experiment.

[0083] Based on any embodiment of this application, please refer to Figure 8 As shown, it is a schematic diagram of a process for determining the target experiment from the custom experiment according to the diversion parameters provided by an embodiment of the present application, and the step S1023 includes:

[0084] Step S10231: Calculate the hash code of the user's ID.

[0085] Step S10232: obtain a remainder by performing a modulo operation on the hash code and the preset number, and use the remainder as the ID of the target traffic bucket.

[0086] Step S10233: Use the experiment corresponding to the ID of the target traffic bucket in at least one custom experiment bound to the online store as the target experiment.

[0087] For example, assuming that a user with userID "A" currently visits the product recommendation page of an online store with ID 1677209962375, the independent station server 82 running the online store with ID 1677209962375 will be triggered to send a product recommendation request to the application server 81, and the product recommendation request includes the user's ID "A" and the ID of the online store, i.e., 1677209962375; taking the three custom experiments with IDs 10008, 10009, and 100010 bound to the online store with ID 1677209962375 as an example, assuming that the calculated ID of the target traffic bucket is 0, the custom experiment with ID 100010 corresponding to the traffic bucket with ID 0 will be determined as the target experiment. Thus, the application server 81 can execute the experimental code of the custom experiment with ID 100010 to provide the user with userID "A" with a product recommendation list obtained based on the recommendation strategy corresponding to the custom experiment, and collect the user behavior data with userID "A".

[0088] According to the above embodiments, it can be understood that the recommendation strategy experiment method of the present application determines whether there is a bound custom experiment in the online store according to the ID of the online store when any user visits the online store. When the online store is not bound to any custom experiment, the hash code is calculated according to the user's ID, and the ID of the target traffic bucket is obtained based on the hash code and the total number of traffic buckets, and a target experiment is determined from the first number of ordinary experiments according to the ID of the target traffic bucket, and the target experiment is executed to recommend products to the user based on the target recommendation strategy corresponding to the target experiment. Since the hash codes calculated by different user IDs are different, the IDs of the target traffic buckets calculated according to the total number of traffic buckets and different user IDs will also be different. For any online store, after multiple different users visit, each ordinary experiment in the first number of ordinary experiments can be executed on the online store, thereby determining which experiment in the first number of ordinary experiments corresponds to the recommendation strategy that is the optimal recommendation strategy for the online store based on the user behavior data collected under each ordinary experiment.

[0089] In addition, when the online store is bound to at least one custom experiment, the ID of the target traffic bucket is obtained according to the user ID and the total number of traffic buckets, and the target experiment is determined from the custom experiments bound to the online store according to the ID of the target traffic bucket, and the target experiment is executed to recommend products to the user based on the recommendation strategy corresponding to the target experiment. Similarly, since the hash codes calculated by different user IDs are different, the IDs of the target traffic buckets calculated according to the total number of traffic buckets and different user IDs will also be different. For any online store bound to the at least one custom experiment, after multiple different users visit, each custom experiment in the at least one custom experiment can be executed on the online store, thereby determining which experiment in the at least one custom experiment corresponds to the recommendation strategy that is the optimal recommendation strategy for the online store based on the user behavior data collected under each custom experiment. This solution can combine the store dimension to perform recommendation strategy experiments on online stores, realize custom experiments on online stores with specific needs, and make the user behavior data used to determine the best recommendation strategy more accurate. For example, after an online clothing store executes the above five ordinary experiments, it determines that the recommendation strategy s1 is the best. Assume that the developer wants to further confirm that under the recommendation strategy s1, the parameters of the recommendation strategy s1, such as clothing categories, are optimized from seasonal divisions to divisions by gender in each season. In this case, a custom experiment can be created based on the recommendation strategy after the service category optimization to understand the user behavior data under the recommendation strategy after the parameter optimization.

[0090] See also Fig. 9 , a recommendation strategy experiment device is provided to meet one of the purposes of the present application, which is a functional embodiment of the recommendation strategy experiment method of the present application. The recommendation strategy experiment device includes a receiving module 101, a response module 102 and an execution module 103, wherein: the receiving module 101 is used to receive a product recommendation request associated with an online store; the response module 102 is used to determine a target experiment from multiple preset recommendation strategy experiments according to the diversion parameters in the product recommendation request, and the target experiment corresponds to a target recommendation strategy; the execution module 103 is used to execute the target experiment to make product recommendations based on the target recommendation strategy.

[0091] On the basis of any embodiment of the present application, before the receiving module 101, it also includes: a preset module, which is used to preset multiple recommendation strategy experiments and configure a corresponding recommendation strategy for each of the multiple recommendation strategy experiments.

[0092] Based on any embodiment of the present application, the preset module includes: a providing submodule for providing a first number of common experiments, each of the first number of common experiments including an experiment code and a configuration file required to execute the experiment code; a judging submodule for judging whether a request to add a custom experiment is received; a first determining submodule for treating each of the first number of common experiments as a recommended strategy experiment if no request to add a custom experiment is received; a binding submodule for responding to a request to add a custom experiment if a request to add a custom experiment is received, and binding a second number of custom experiments to a specified online store, each of the second number of custom experiments including an experiment code and a configuration file required to execute the experiment code; and a second determining submodule for treating each of the first number of common experiments as a recommended strategy experiment, and treating each of the second number of custom experiments as a recommended strategy experiment.

[0093] Based on any embodiment of the present application, the diversion parameter includes the ID of the online store and the ID of the user accessing the online store. The response module 102 includes: a confirmation submodule, which is used to determine whether the online store is bound to at least one custom experiment according to the ID of the online store; a first target determination submodule, which is used to determine the target experiment from the first number of common experiments according to the ID of the user when the online store is not bound to any custom experiment; and a second target determination submodule, which is used to determine the target experiment from at least one custom experiment bound to the online store according to the ID of the user when the online store is bound to at least one custom experiment.

[0094] Based on any embodiment of the present application, the first target determination submodule includes: a first hash code calculation unit, used to calculate the hash code (hashCode) of the user's ID; a first remainder operation unit, used to obtain a remainder by performing a modulo operation on the hash code and the number of all traffic buckets, and use the remainder as the ID of the target traffic bucket; and a first target determination unit, used to use the ordinary experiment corresponding to the ID of the target traffic bucket among the first number of ordinary experiments as the target experiment.

[0095] Based on any embodiment of the present application, the second target determination submodule includes: a second hash code calculation unit, used to calculate the hash code of the user's ID; a second remainder operation unit, used to obtain a remainder by performing a modulo operation on the hash code and the number of all traffic buckets, and use the remainder as the ID of the target traffic bucket; and a second target determination unit, used to use the experiment corresponding to the ID of the target traffic bucket in at least one custom experiment bound to the online store as the target experiment.

[0096] Another embodiment of the present application also provides a computer device. Fig.10 As shown, in one embodiment of the present application, the computer device 100 includes, but is not limited to, a memory 1001, a processor 1002, and computer-readable instructions stored in the memory 1001 and executable on the processor 1002, such as a recommendation strategy experiment program.

[0097] Those skilled in the art will appreciate that the schematic diagram is merely an example of the computer device 100 and does not constitute a limitation on the computer device 100 , which may include more or fewer components than those shown in the diagram, or a combination of certain components, or different components. For example, the computer device 100 may also include input and output devices, network access devices, buses, etc.

[0098] The processor 1002 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor 1002 is the computing core and control center of the computer device 100, and uses various interfaces and lines to connect various parts of the entire computer device 100, and execute the operating system of the computer device 100 and various installed application programs, program codes, etc.

[0099] Exemplarily, the computer-readable instructions may be divided into one or more modules / sub-modules / units, one or more modules / sub-modules / units are stored in the memory 1001, and are executed by the processor 1002 to complete the present application. One or more modules / sub-modules / units may be a series of computer-readable instruction segments capable of completing a specific function, and the computer-readable instruction segments are used to describe the execution process of the computer-readable instructions in the computer device 100. For example, the computer-readable instructions may be divided into a receiving module 101, a response module 102, and an execution module 103.

[0100] The memory 1001 can be used to store computer-readable instructions and / or modules. The processor 1002 implements various functions of the computer device 100 by running or executing the computer-readable instructions and / or modules stored in the memory 1001 and calling the data stored in the memory 1001. The memory 1001 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the computer device 100, etc. The memory 1001 can include non-volatile and volatile memories, such as: a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other storage devices.

[0101] The memory 1001 may be an external memory and / or an internal memory of the computer device 100. Furthermore, the memory 1001 may be a memory in a physical form, such as a memory stick, a TF card (Trans-flash Card), and the like.

[0102] If the modules / units integrated in the computer device 100 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, and can also instruct the relevant hardware to complete through computer-readable instructions. The computer-readable instructions can be stored in a computer-readable storage medium, and when the computer-readable instructions are executed by the processor, the steps of the above-mentioned method embodiments can be implemented.

[0103] The computer-readable instructions include computer-readable instruction codes, which may be in source code form, object code form, executable files, or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer-readable instruction codes, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM).

[0104] Combination Figures 2 to 8 The memory 1001 in the computer device 100 stores computer-readable instructions, and the processor 1002 can execute the computer-readable instructions stored in the memory 1001 to implement the following Figures 2 to 8 The data verification method shown.

[0105] Specifically, the specific implementation method of the processor 1002 for the above-mentioned computer readable instructions can refer to FIG. Figures 2 to 8 The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0106] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of modules is only a logical function division, and there may be other division methods in actual implementation.

[0107] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0108] In addition, each functional module in each embodiment of the present application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0109] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present application is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present application. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0110] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices can also be implemented by one unit or device through software or hardware. The words first, second, etc. are used to indicate names, and do not indicate any specific order.

[0111] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present application and are not intended to limit it. Although the present application has been described in detail with reference to the preferred embodiments, a person of ordinary skill in the art should understand that the technical solution of the present application may be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present application.

Claims

1. A recommendation strategy experimental method, It is characterized in that The method comprises: receiving a product recommendation request associated with an online store; Determining a target experiment from a plurality of preset recommendation strategy experiments according to the diversion parameters in the product recommendation request, wherein the target experiment corresponds to a target recommendation strategy; and The target experiment is executed to make product recommendations based on the target recommendation strategy.

2. The recommendation strategy experimental method according to claim 1, It is characterized in that The method further comprises: The plurality of recommendation strategy experiments are preset, and a corresponding recommendation strategy is configured for each of the plurality of recommendation strategy experiments.

3. The recommendation strategy experimental method according to claim 2, It is characterized in that The presetting the plurality of recommendation strategy experiments and configuring a corresponding recommendation strategy for each of the plurality of recommendation strategy experiments comprises: Providing a first number of common experiments, each common experiment in the first number of common experiments includes an experiment code and a configuration file required to execute the experiment code; If no request for adding a custom experiment is received, each common experiment in the first number of common experiments is used as a recommended strategy experiment; or If a request to add a custom experiment is received, in response to the request to add a custom experiment, a second number of custom experiments are bound to the specified online store, each of the second number of custom experiments including an experiment code and a configuration file required to execute the experiment code; and each ordinary experiment in the first number of ordinary experiments is used as a recommended strategy experiment, and each custom experiment in the second number of custom experiments is used as a recommended strategy experiment.

4. The recommendation strategy experimental method according to claim 3, It is characterized in that The configuration file of any common experiment includes the identification ID, traffic bucket ID, and recommended strategy of the common experiment; The configuration file corresponding to any custom experiment includes the ID of the custom experiment, the ID of the traffic bucket, the recommendation strategy, and the ID of the designated online store.

5. The recommendation strategy experimental method according to claim 4, It is characterized in that The online store runs on an independent website server of an e-commerce platform, and the diversion parameters include the ID of the online store and the ID of a user accessing the online store.

6. The recommendation strategy experimental method according to claim 5, It is characterized in that The step of determining a target experiment from a plurality of preset recommendation strategy experiments according to the diversion parameters in the product recommendation request, wherein the target experiment corresponds to a target recommendation strategy including: Determining, according to the ID of the online store, whether the online store is bound to at least one custom experiment; If the online store is not bound to any custom experiment, determining the target experiment from the first number of common experiments according to the user ID; or If the online store is bound to at least one custom experiment, the target experiment is determined from the at least one custom experiment bound to the online store according to the ID of the user.

7. The recommendation strategy experimental method according to claim 6, It is characterized in that Determining the target experiment from the first number of common experiments according to the user ID includes: Calculate a hash code for the user's ID; A remainder is obtained by performing a modulo operation on the hash code and the number of all traffic buckets, and the remainder is used as the ID of the target traffic bucket; and The common experiment corresponding to the ID of the target traffic bucket among the first number of common experiments is used as the target experiment.

8. The recommendation strategy experimental method according to claim 6, It is characterized in that Determining the target experiment from at least one custom experiment bound to the online store according to the user ID includes: Calculate a hash code for the user's ID; A remainder is obtained by performing a modulo operation on the hash code and the number of all traffic buckets, and the remainder is used as the ID of the target traffic bucket; and The experiment corresponding to the ID of the target traffic bucket in at least one custom experiment bound to the online store is used as the target experiment.

9. A recommendation strategy experimental device, It is characterized in that include: A receiving module, used to receive a product recommendation request associated with an online store; A response module, configured to determine a target experiment from a plurality of preset recommendation strategy experiments according to the diversion parameters in the product recommendation request, wherein the target experiment corresponds to a target recommendation strategy; and An execution module is used to execute the target experiment to make product recommendations based on the target recommendation strategy.

10. A computer device, It is characterized in that include: Memory, and A processor, wherein the processor executes computer-readable instructions stored in the memory to implement the recommendation strategy experimental method according to any one of claims 1 to 8.

Images