Resource determination method and device, equipment, storage medium and program product

Abstract

The invention relates to a resource determination method and device, equipment, a storage medium and a program product. The method comprises the following steps: acquiring activity configuration information of a target online car-hailing platform and historical performance data of each registered driver; for each registered driver, determining a target prediction model corresponding to the registered driver, inputting the activity configuration information and the historical performance data of the registered driver into the target prediction model, and determining the online probability and the possible order completion type of the registered driver according to the output of the target prediction model; determining the number of drivers corresponding to different order completion types according to the online probability of each registered driver and the possible order completion type; and determining activity resources of the target online car-hailing platform according to the number of drivers corresponding to different order completion types and the reward reference values corresponding to different order completion types. By adopting the method, the accuracy of resource determination can be improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular, to a method, apparatus, device, storage medium, and program product for resource determination. Background Art

[0002] With the continuous expansion of the scale of the online car-hailing market, more and more enterprises have entered the online car-hailing market to compete for market share. Each online car-hailing platform attracts drivers to provide services for the platform by launching activities and distributing bonuses. For the platform, distributing bonuses to registered drivers who complete the activity tasks is an important resource for platform operation. It is very necessary to determine the activity resources before launching the activity.

[0003] In the prior art, time series analysis is usually used to determine the resources of the platform's current activity. For example, based on historical similar activity resources, models such as time average and exponential smoothing are used to predict the resources of this activity. However, the prior art has the problem of low accuracy. Summary of the Invention

[0004] Based on this, in view of the above technical problems, it is necessary to provide a method, apparatus, device, storage medium, and program product for resource determination that can improve accuracy.

[0005] In a first aspect, this application provides a method for resource determination, including:

[0006] Obtain the activity configuration information of the target online car-hailing platform and the historical performance data of each registered driver; for each registered driver, determine the target prediction model corresponding to the registered driver, and input the activity configuration information and the historical performance data of the registered driver into the target prediction model, and determine the online probability and possible order completion types of the registered driver according to the output of the target prediction model; according to the online probability and possible order completion types of each registered driver, determine the number of drivers corresponding to different order completion types; determine the activity resources of the target online car-hailing platform according to the number of drivers corresponding to different order completion types and the reward reference values corresponding to different order completion types.

[0007] In one embodiment, the target prediction model includes an online probability model and an order completion type model. Inputting the activity configuration information and the historical performance data of the registered driver into the target prediction model, and determining the online probability and possible order completion types of the registered driver according to the output of the target prediction model includes:

[0008] Input the activity configuration information and the historical performance data of the registered drivers into the online probability model, and determine the online probability of the registered drivers according to the output of the online probability model; input the activity configuration information and the historical performance data of the registered drivers into the order completion type model, and determine the possible order completion types of the registered drivers according to the output of the order completion type model.

[0009] In one embodiment, determining the target prediction model corresponding to the registered driver includes:

[0010] Obtain the type of the registered driver; determine the target prediction model from the preset model list according to the type of the registered driver.

[0011] In one embodiment, obtaining the type of the registered driver includes:

[0012] Determine the performance type of the registered driver according to the identification information of the registered driver; determine the service platform type of the registered driver according to the service platform information of the registered driver, and the service platform type includes a single service platform and a multi-service platform; determine the type of the registered driver according to the performance type of the registered driver and the service platform type of the registered driver.

[0013] In one embodiment, the preset model list includes a single-platform model list and a multi-platform model list. Determining the target prediction model from the preset model list according to the type of the registered driver includes:

[0014] If the service platform type of the registered driver is a single service platform type, determine the target prediction model corresponding to the online car-hailing platform served by the registered driver and the performance type of the registered driver from the single-platform model list. The single-platform model list is used to store the corresponding relationships between different online car-hailing platforms and performance types and prediction models; if the service platform type of the registered driver is a multi-service platform type, obtain the target platform combination corresponding to the registered driver. The target platform combination includes multiple online car-hailing platforms served by the registered driver. Determine the sub-list of prediction models corresponding to the target platform combination and the performance type of the registered driver from the multi-platform model list, and determine the target prediction model corresponding to the target online car-hailing platform from the sub-list of prediction models. The multi-platform model list is used to store the corresponding relationships between different platform combinations and performance types and sub-lists of prediction models, and the sub-list of prediction models is used to store the corresponding relationships between different online car-hailing platforms and prediction models.

[0015] In one embodiment, determining the number of drivers corresponding to different order completion types according to the online probability and order completion type of each registered driver includes:

[0016] For each registered driver, determine the probability estimate of the possible order completion types of the registered driver, and determine the expected value of the driver quantity contribution of the registered driver to the order completion type according to the probability estimate and the online probability; for each order completion type, determine the driver quantity corresponding to the order completion type according to the sum of the expected values of the driver quantity contributions of each registered driver to the order completion type.

[0017] In one embodiment, determine the activity resources of the target online car-hailing platform according to the driver quantities corresponding to different order completion types and the reward reference values corresponding to different order completion types, including:

[0018] For each order completion type, determine the reward resources of the order completion type according to the driver quantity and the reward reference value; determine the activity resources of the target online car-hailing platform according to the sum of the reward resources of each order completion type.

[0019] In a second aspect, the present application also provides a resource determination device, including:

[0020] An acquisition module, configured to acquire the activity configuration information of the target online car-hailing platform and the historical performance data of each registered driver;

[0021] A first determination module, configured to, for each registered driver, determine the target prediction model corresponding to the registered driver, input the activity configuration information and the historical performance data of the registered driver into the target prediction model, and determine the online probability and the possible order completion types of the registered driver according to the target prediction model;

[0022] A second determination module, configured to determine the driver quantities corresponding to different order completion types according to the online probabilities and the possible order completion types of each registered driver;

[0023] A third determination module, configured to determine the activity resources of the target online car-hailing platform according to the driver quantities corresponding to different order completion types and the reward reference values corresponding to different order completion types.

[0024] In one embodiment, the target prediction model includes an online probability model and an order completion type model. The first determination module is specifically configured to input the activity configuration information and the historical performance data of the registered driver into the online probability model, and determine the online probability of the registered driver according to the output of the online probability model; input the activity configuration information and the historical performance data of the registered driver into the order completion type model, and determine the possible order completion types of the registered driver according to the output of the order completion type model.

[0025] In one embodiment, the first determination module is specifically configured to obtain the type of the registered driver; determine the target prediction model from the preset model list according to the type of the registered driver.

[0026] In one embodiment, the first determination module is specifically configured to determine the performance type of the registered driver according to the identification information of the registered driver; determine the service platform type of the registered driver according to the service platform information of the registered driver, where the service platform type includes a single service platform and a multi-service platform; and determine the type of the registered driver according to the performance type of the registered driver and the service platform type of the registered driver.

[0027] In one embodiment, the preset model list includes a single-platform model list and a multi-platform model list. The first determination module is specifically configured to, if the service platform type of the registered driver is a single service platform type, determine a target prediction model corresponding to the online car-hailing platform served by the registered driver and the performance type of the registered driver from the single-platform model list, where the single-platform model list is used to store the corresponding relationships between different online car-hailing platforms and performance types and prediction models; if the service platform type of the registered driver is a multi-service platform type, obtain the target platform combination corresponding to the registered driver, where the target platform combination includes multiple online car-hailing platforms served by the registered driver, determine a sub-list of prediction models corresponding to the target platform combination and the performance type of the registered driver from the multi-platform model list, and determine a target prediction model corresponding to the target online car-hailing platform from the sub-list of prediction models, where the multi-platform model list is used to store the corresponding relationships between different platform combinations and performance types and sub-lists of prediction models, and the sub-list of prediction models is used to store the corresponding relationships between different online car-hailing platforms and prediction models.

[0028] In one embodiment, the second determination module is specifically configured to, for each registered driver, determine the probability estimate of the possible order completion type of the registered driver, and determine the expected value of the driver quantity contribution of the registered driver to the order completion type according to the probability estimate and the online probability; for each order completion type, determine the driver quantity corresponding to the order completion type according to the sum of the expected values of the driver quantity contributions of each registered driver to the order completion type.

[0029] In one embodiment, the third determination module is specifically configured to, for each order completion type, determine the reward resources of the order completion type according to the driver quantity and the reward reference value; determine the activity resources of the target online car-hailing platform according to the sum of the reward resources of each order completion type.

[0030] In a third aspect, the present application further provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the method described in any one of the first aspects above is implemented.

[0031] In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method described in any one of the first aspects above is implemented.

[0032] Fifth aspect, the present application also provides a computer program product, including a computer program which, when executed by a processor, implements the method described in any one of the above first aspects.

[0033] For the above resource determination method, device, equipment, storage medium and program product, by obtaining the activity configuration information of the target online car-hailing platform and the historical performance data of each registered driver, then, for each registered driver, determining the target prediction model corresponding to the registered driver, and inputting the activity configuration information and the historical performance data of the registered driver into the target prediction model, determining the online probability of the registered driver and the possible order completion types according to the output of the target prediction model, and then, according to the online probability of each registered driver and the possible order completion types, determining the number of drivers corresponding to different order completion types, and finally, determining the activity resources of the target online car-hailing platform according to the number of drivers corresponding to different order completion types and the reward reference values corresponding to different order completion types. In this way, different types of registered drivers correspond to different target prediction models, and the online probability of each registered driver and the possible order completion types can be accurately determined, and then the number of drivers corresponding to different order completion types can be determined, so as to determine the activity resources according to the number of drivers and the corresponding reward values. Through multi-type and hierarchical determination, the impact of the accurate activity configuration on the service behaviors of different types of registered drivers can be obtained, thereby improving the accuracy of activity resource determination. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required to be used in the description of the embodiments of the present application or related technologies. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained according to these drawings.

[0035] Figure 1 It is a schematic flowchart of the resource determination method in an embodiment;

[0036] Figure 2 It is a schematic flowchart of the steps of determining the online probability of a registered driver and the possible order completion types in an embodiment;

[0037] Figure 3 It is a schematic flowchart of the steps of determining the target prediction model in an embodiment;

[0038] Figure 4 It is a schematic flowchart of the steps of obtaining the type of a registered driver in an embodiment;

[0039] Figure 5 It is a schematic diagram of the result of driver classification in an embodiment;

[0040] Figure 6 It is a schematic flow chart of the steps for determining the target prediction model in another embodiment;

[0041] Figure 7 It is a schematic flow chart of the steps for determining the number of drivers corresponding to different order completion types in one embodiment;

[0042] Figure 8 It is a schematic flow chart of determining activity resources in one embodiment;

[0043] Figure 9 It is a schematic flow chart of the resource determination method in another embodiment;

[0044] Figure 10 It is a schematic framework diagram of the resource determination method in one embodiment;

[0045] Figure 11 It is a structural block diagram of the resource determination device in one embodiment;

[0046] Figure 12 It is an internal structure diagram of a computer device in one embodiment. Detailed implementation manners

[0047] In order to make the objectives, technical solutions and advantages of the present application more clear and understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0048] With the continuous expansion of the online car-hailing market scale, more and more enterprises enter the online car-hailing market to compete for market share. Each online car-hailing platform attracts drivers to provide services for the platform by launching activities and awarding bonuses. For the platform, awarding bonuses to registered drivers who complete the activity tasks is an important resource for platform operation. It is very necessary to determine the activity resources before launching the activity, where the activity resources are the activity costs.

[0049] In the prior art, the platform's activity resources are usually determined by time series analysis. For example, based on historical similar activity resources, models such as time average and exponential smoothing are used to predict the activity resources this time. However, the prior art has the problem of low accuracy.

[0050] In view of this, the present application provides a resource determination method that can improve the accuracy of resource determination. The resource determination method provided by the embodiments of the present application may be executed by a resource determination device, which may be implemented in a software, hardware, or a combination of software and hardware manner. It may be embedded in the processor of a computer device in a hardware form or be independent of it, or be stored in the memory of the computer device in a software form. In the following method embodiments, the execution subject is taken as a computer device for illustration. The computer device may be a server or a desktop computer. The embodiments of the present application do not limit the specific type of the computer device.

[0051] In an exemplary embodiment, as Figure 1 shown, a resource determination method is provided, including the following steps 101 to 104. Wherein:

[0052] S101, obtain the activity configuration information of the target online car-hailing platform and the historical performance data of each registered driver.

[0053] Optionally, the target online car-hailing platform may be the online car-hailing platform that will determine activity resources through the method of the embodiments of the present application.

[0054] Optionally, the activity configuration information may be the activity information set by the target online car-hailing operator. For example, the values of the bonus that registered drivers can obtain for completing different order volumes and the activity validity period, etc.

[0055] Optionally, the historical performance data of the registered driver may include the historical performance data of the registered driver in the past week, past month, or past year. The embodiments of the present application do not limit the time range of the historical performance data. The historical performance data may include data such as the daily order completion volume of the driver, the maximum order completion volume, the online duration, the number of consecutive days the driver is on duty, and the number of days the driver is not on duty.

[0056] Optionally, the activity configuration information may be obtained from a database or a configuration file, and the historical performance data of each registered driver may be obtained from a database or a data warehouse. The database, configuration file, or data warehouse may include the activity configuration information and the historical performance data of multiple online car-hailing platforms, or may only include the activity configuration information and the historical performance data of the target online car-hailing platform. The embodiments of the present application do not limit the manner of obtaining the activity configuration information of the target online car-hailing platform and the historical performance data.

[0057] S102, for each registered driver, determine the target prediction model corresponding to the registered driver, and input the activity configuration information and the historical performance data of the registered driver into the target prediction model. Determine the online probability of the registered driver and the possible order completion types according to the output of the target prediction model.

[0058] Optionally, registered drivers can be classified into different types, and the corresponding target prediction models for different types of registered drivers may be different. When classifying registered drivers, it can be achieved through data such as the age, region, registration time, order completion volume, and / or service platform of the registered drivers.

[0059] Optionally, the target prediction model can be implemented based on a recurrent neural network or deep learning, and is trained using the historical performance data and historical activity configuration information of all registered drivers in the registered driver type corresponding to the prediction model as the training set.

[0060] Optionally, the target prediction model can be a model that can output multiple categorical results, that is, by inputting the activity configuration information and the historical performance data of the registered driver into the target prediction model, the target prediction model can output the online probability of the registered driver and the possible order completion types; it can also include multiple sub-models. By inputting the activity configuration information and the historical performance data of the registered driver into a target prediction sub-model, the online probability of the registered driver can be output. By inputting the activity configuration information and the historical performance data of the registered driver into another target prediction sub-model, the possible order completion types of the registered driver can be output.

[0061] Optionally, the online probability can be a variable between 0 and 1, which is used to represent the possibility of the registered driver going online to receive orders; the order completion types can be multiple types set according to the order completion quantity. For example, the order completion types can include low, medium, high, or more, and the possible order completion type can be the order completion type with the highest probability estimate.

[0062] S103. Determine the number of drivers corresponding to different order completion types according to the online probability and possible order completion types of each registered driver.

[0063] In a possible implementation manner, for each registered driver, after determining the online probability, it can be determined whether the registered driver will go online according to the online probability and a preset probability threshold. If it is determined that the registered driver will go online, then the number of drivers corresponding to the order completion type of this registered driver is incremented by 1.

[0064] In another possible implementation manner, for each registered driver, calculate the online probability of the registered driver and the probability estimate corresponding to the possible order completion type, determine the expected value of the contribution of the registered driver to the number of drivers for different order completion types, and then calculate the sum of the expected values of the contribution of the number of drivers under each order completion type to obtain the corresponding number of drivers.

[0065] S104. Determine the activity resources of the target online car-hailing platform according to the number of drivers corresponding to different order completion types and the reward reference values corresponding to different order completion types.

[0066] Among them, the activity resources can be the activity costs.

[0067] Optionally, for each order completion type, the corresponding reward reference value can be obtained by taking the average, median, or weighted sum of the reward amounts corresponding to multiple historical activities of this order completion type. The embodiments of the present application do not limit the method for determining the reward reference values corresponding to different order completion types.

[0068] Optionally, for each order completion type, after determining the corresponding driver quantity and reward reference value, the activity resources for this order completion type can be determined, and then based on the activity resources corresponding to each order completion type, the activity resources of the target online car-hailing platform can be determined. The activity resources can be the reward amounts distributed by the target online car-hailing platform to registered drivers after the registered drivers complete the required order quantities for the activities.

[0069] The above resource determination method obtains the activity configuration information of the target online car-hailing platform and the historical performance data of each registered driver. Then, for each registered driver, it determines the target prediction model corresponding to the registered driver, inputs the activity configuration information and the historical performance data of the registered driver into the target prediction model, determines the online probability and possible order completion types of the registered driver according to the output of the target prediction model, and then determines the driver quantities corresponding to different order completion types according to the online probabilities and possible order completion types of each registered driver. Finally, it determines the activity resources of the target online car-hailing platform according to the driver quantities corresponding to different order completion types and the reward reference values corresponding to different order completion types. In this way, different types of registered drivers correspond to different target prediction models, which can accurately determine the online probabilities and possible order types that each registered driver can complete, and then determine the driver quantities corresponding to different order completion types, so as to determine the activity resources according to the driver quantities and the corresponding reward values. Through multi-type and hierarchical determination, the impact of the accurate activity configuration on the service behaviors of different types of registered drivers can be obtained, thereby improving the accuracy of activity resource determination.

[0070] In an exemplary embodiment, as Figure 2 shown, optionally, the target prediction model includes an online probability model and an order completion type model. Inputting the activity configuration information and the historical performance data of the registered driver into the target prediction model and determining the online probability and possible order completion types of the registered driver according to the output of the target prediction model includes the following steps 201 to 202. Among them:

[0071] S201, input the activity configuration information and the historical performance data of the registered driver into the online probability model, and determine the online probability of the registered driver according to the output of the online probability model.

[0072] Optionally, the historical performance data of all registered drivers in this type can be determined according to the type of registered drivers, and the initial online probability model can be trained with the historical performance data and the historical activity configuration information as training data until the loss function converges, and then the online probability model can be obtained.

[0073] Optionally, when training the initial online probability model, the difference between the prediction of the model and the actual label can be measured based on the loss function. The loss function can be the cross-entropy loss plus the "L2" regularization term, which can improve the accuracy of the online probability model and effectively prevent the model from overfitting.

[0074] Optionally, when determining the online probability of a registered driver according to the online probability model, the activity configuration information of this activity and the historical performance data of this registered driver can be used as inputs to obtain the online probability.

[0075] Optionally, in the online probability model, preprocessing and feature engineering can be performed on the input data. The preprocessing can include handling missing values, duplicate values, outliers, etc., and feature extraction can be performed on the preprocessed data.

[0076] Optionally, the online probability model can be implemented based on logistic regression, can be other binary classification algorithms, or can also be implemented based on methods such as gradient boosting decision trees or Bayesian networks. The specific type of the online probability model is not limited in the embodiments of this application.

[0077] S202, input the activity configuration information and the historical performance data of the registered driver into the order completion type model, and determine the possible order completion types of the registered driver according to the output of the order completion type model.

[0078] Optionally, the historical performance data of all registered drivers in this type can be determined according to the type of registered drivers, and the historical performance data after going online can be screened out from the historical performance data. The initial order completion type model can be trained with the historical performance data after going online and the historical activity configuration information as training data until the loss function converges, and then the order completion type model can be obtained.

[0079] Optionally, the activity configuration information of this activity and the historical performance data of this registered driver after going online can be used as inputs, and the order completion type of this registered driver type after going online can be determined through the order completion type model.

[0080] Optionally, the order completion type model can be implemented based on the random forest method, or can also be based on methods such as gradient boosting decision trees or Bayesian networks. This is not limited in the embodiments of this application.

[0081] The above-mentioned activity configuration information and the historical performance data of registered drivers are input into the online probability model, and the online probability of registered drivers is determined according to the output of the online probability model. The activity configuration information and the historical performance data of registered drivers are input into the order completion type model, and the possible order completion types of registered drivers are determined according to the output of the order completion type model. In this way, different online probability models and order completion type models can be determined according to the types of registered drivers, and the online probability and order completion types of registered drivers can be determined accurately and specifically according to driver behavior characteristics.

[0082] In an exemplary embodiment, as Figure 3 shown, optionally, determining the target prediction model corresponding to the registered driver includes the following steps 301 to 302. Among them:

[0083] S301, obtain the type of the registered driver.

[0084] In a possible implementation manner, the historical performance data of the registered driver can be input into a preset driver classification model, and the type of the registered driver is determined according to the output of the driver classification model.

[0085] In another possible implementation manner, as Figure 4 shown, obtaining the type of the registered driver may include the following steps 401 to 403. Among them:

[0086] S401, determine the performance type of the registered driver according to the identification information of the registered driver.

[0087] Optionally, the performance type of the registered driver can be predetermined. For example, according to the service performance dimension of the registered driver, the performance type is divided into five types, order1 - order5, and stored as the attribute information of the registered driver in the registered driver information table of the target online car-hailing platform. The performance type of the registered driver can be determined according to the identification information of the registered driver.

[0088] Optionally, the identification information of the registered driver can be the registration account, mobile phone number or ID number of the registered driver, that is, the identification information of the registered driver corresponds one-to-one with the registered driver.

[0089] In a possible implementation manner, the performance type of the registered driver can be obtained by clustering the registered drivers on the target online car-hailing platform into multiple performance types according to two dimensions of the average daily order completion volume and the maximum order completion volume of all registered drivers using a clustering method.

[0090] Optionally, the clustering method may include the Kmeans clustering method or the Gaussian mixture clustering method, etc.

[0091] Optionally, higher-order moment information such as the variance or skewness of the daily order completion volume of all registered drivers can also be used as the basis for clustering to cluster the registered drivers.

[0092] In another possible implementation, the questionnaire survey results of registered drivers can be obtained, and the performance types of registered drivers can be determined based on the questionnaire survey results.

[0093] In another possible implementation, the historical performance data of registered drivers can be evaluated according to an evaluation method to determine the performance types of registered drivers based on the evaluation results.

[0094] S402. Determine the service platform type of the registered driver according to the service platform information of the registered driver.

[0095] Among them, the service platform types include single-service platforms and multi-service platforms.

[0096] In one possible implementation, the registered driver information of each online car-hailing platform can be obtained to determine the service platform information of the registered driver.

[0097] In another possible implementation, the service platform information of the registered driver can be filled in by the registered driver, and the service platform information can be determined from the registered driver information table of the target online car-hailing platform according to the identification information of the registered driver.

[0098] Optionally, when the service platform information of the registered driver is one online car-hailing platform, the service platform type of the registered driver can be determined as a single-service platform; when the service platform information of the registered driver is two or more, the service platform type of the registered driver can be determined as a multi-service platform.

[0099] It can be understood that when determining the service platform type of the registered driver, the specific service platform can be determined. For example, taking 5 online car-hailing platforms A - E as an example, the service platform type of the registered driver can be a single-service platform A or a multi-service platform AC.

[0100] S403. Determine the type of the registered driver according to the performance type of the registered driver and the service platform type of the registered driver.

[0101] Optionally, the type of the registered driver can be determined according to the combination of the performance type and the service platform type of the registered driver, and the combination method is not limited.

[0102] Exemplarily, if the performance type of the registered driver is order1 and the service platform type is AC, the type of the registered driver can be order1 - AC.

[0103] Exemplarily, such as Figure 5As shown, taking 5 ride-hailing platforms with 5 performance types, namely ride-hailing platform 1 - ride-hailing platform 5, and a multi-service platform serving two ride-hailing platforms as an example, 5 single-service platform types and 10 multi-service platforms can be obtained. According to the combination of the performance type of registered drivers and the service platform type, 5×(5 + 10)=75 driver types can be determined.

[0104] S302. Determine a target prediction model from a preset model list according to the type of registered driver.

[0105] Optionally, multiple prediction models can be stored in the preset model list. The preset model list is used to store the corresponding relationships between different prediction models and the types of registered drivers. At least one prediction model corresponds to the type of a registered driver. For example, when the performance type of a registered driver is order2 and the service platform type of the registered driver is the multi-service platform type AC, two prediction models A22 and C22 can be corresponding. The A22 prediction model takes platform A as the main platform. Therefore, the A22 prediction model can be trained with the historical performance data of registered drivers with all performance types of order2 on platforms A and C and the historical activity configuration information of platform A, or can be trained with the historical performance data of registered drivers with all performance types of order2 on platform A and the historical activity configuration information of platform A, or can also be to select a part of the historical performance data from the historical performance data of registered drivers with all performance types of order2 on platforms A and C according to a certain ratio, where the proportion of the historical performance data of platform A is greater than that of platform C, and then train the selected part of the historical performance data and the historical activity configuration information of platform A.

[0106] In a possible implementation manner, as Figure 6 shown, the preset model list includes a single-platform model list and a multi-platform model list. Determining a target prediction model from the preset model list according to the type of registered driver includes the following steps 601 to step 602. Among them:

[0107] S601. If the service platform type of the registered driver is a single-service platform type, determine a target prediction model corresponding to the ride-hailing platform served by the registered driver and the performance type of the registered driver from the single-platform model list.

[0108] Among them, the single-platform model list is used to store the corresponding relationships between different ride-hailing platforms, performance types, and prediction models.

[0109] Optionally, the single-platform model list can be stored in a database, Redis, or local cache. A Structured Query Language (SQL) statement can be generated based on the online car-hailing platforms registered by the driver service and the performance type of the registered driver, and the target prediction model can be determined according to the query result.

[0110] Exemplarily, a possible format of the single-platform model list is shown in Table 1:

[0111] Table 1

[0112]

[0113] S602. If the service platform type of the registered driver is a multi-service platform type, obtain the target platform combination corresponding to the registered driver. The target platform combination includes multiple online car-hailing platforms served by the registered driver. Determine a sub-list of prediction models corresponding to the target platform combination and the performance type of the registered driver from the multi-platform model list, and determine the target prediction model corresponding to the target online car-hailing platform from the sub-list of prediction models.

[0114] Among them, the multi-platform model list is used to store the corresponding relationships between different platform combinations and performance types and sub-lists of prediction models, and the sub-lists of prediction models are used to store the corresponding relationships between different online car-hailing platforms and prediction models.

[0115] Optionally, similar to the method for determining the target prediction model for the above single-service platform type, the multi-platform model list can be stored in a database, Redis, or local cache. Similarly, a Structured Query Language (SQL) statement can be generated based on the target platform combination of the registered driver and the performance type of the registered driver, and the target prediction model can be determined according to the query result.

[0116] Optionally, the target platform combination can be generated according to a preset combination order based on the multiple online car-hailing platforms served by the registered driver.

[0117] Exemplarily, a possible format of the multi-platform model list is shown in Table 2:

[0118] Table 2

[0119]

[0120] In another possible implementation, the preset model list stores the correspondence between different registered driver types and prediction models. The prediction model can be obtained from the preset model list according to the type of the registered driver. When the number of obtained prediction models is 1, this prediction model is used as the target prediction model. When the number of obtained prediction models is more than one, the target prediction sub-model is determined from the prediction models according to the target online car-hailing platform. A possible format of the preset model list is shown in Table 3:

[0121] Table 3

[0122]

[0123] By obtaining the type of the registered driver and determining the target prediction model from the preset model list according to the type of the registered driver as described above, in this way, a multi-category and multi-level target prediction model can be utilized, and a refined and accurate calibration fitting and explanatory description of the impact of the activity configuration on the service behavior of the registered driver can be obtained. At the same time, the prediction model effectively learns from the performance data of the registered drivers of the corresponding type, can identify the behavior logic of the registered drivers, and improve the accuracy of prediction.

[0124] In an exemplary embodiment, as Figure 7 shown, optionally, according to the online probability and order completion type of each registered driver, determine the number of drivers corresponding to different order completion types, including the following steps 701 to step 702. Wherein:

[0125] S701. For each registered driver, determine the probability estimate of the possible order completion types of the registered driver, and determine the expected value of the contribution of the registered driver to the number of drivers of the order completion type according to the probability estimate and the online probability.

[0126] Optionally, when the order completion type model determines the possible order completion types of the registered driver, it also outputs the probability estimate corresponding to the order completion type.

[0127] Optionally, the probability estimate and the online probability can be multiplied, and the multiplication result is determined as the expected value of the contribution of the registered driver to the number of drivers of the order completion type.

[0128] S702. For each order completion type, determine the number of drivers corresponding to the order completion type according to the sum of the expected values of the contribution of each registered driver to the number of drivers of the order completion type.

[0129] Optionally, for an order completion type, the expected value of the contribution of the registered drivers corresponding to the order completion type can be obtained, and the expected values of the contributions of multiple drivers are summed to determine the number of drivers corresponding to the order completion type. In this way, the number of drivers corresponding to each order completion type can be obtained.

[0130] For each registered driver, the probability estimate of the possible order completion types of the registered driver is determined. According to the probability estimate and the online probability, the expected value of the driver quantity contribution of the registered driver to the order completion type is determined. For each order completion type, according to the sum of the expected values of the driver quantity contributions of each registered driver to the order completion type, the driver quantity corresponding to the order completion type can be determined, and the driver quantities corresponding to different order types can be determined.

[0131] In an exemplary embodiment, as Figure 8 shown, optionally, according to the driver quantity corresponding to different order completion types and the reward reference value corresponding to different order completion types, the activity resources of the target online car-hailing platform are determined, including the following steps 801 to step 802. Wherein:

[0132] S801, for each order completion type, determine the reward resources of the order completion type according to the driver quantity and the reward reference value.

[0133] Among them, the reward resources can be the reward costs.

[0134] Optionally, after determining the driver quantity corresponding to an order completion type and the corresponding reward reference value, the reward resources generated under this order completion type can be determined according to the product of the driver quantity and the reward reference value.

[0135] Optionally, the reward resources can be the reward amounts sent by the target online car-hailing platform to the registered drivers after the registered drivers complete the order quantities corresponding to the order completion types.

[0136] S802, determine the activity resources of the target online car-hailing platform according to the sum of the reward resources of each order completion type.

[0137] Optionally, the reward resources of each order completion type can be summed up, and the sum value can be used as the activity resources of the target online car-hailing platform.

[0138] Optionally, after determining the activity resources, the activity resources can be output to the operation personnel of the target online car-hailing platform, so that the operation personnel can adjust the activity configuration information according to the activity resources and the activity budget.

[0139] For each order completion type above, the reward resources of the order completion type are determined according to the driver quantity and the reward reference value, and the activity resources of the target online car-hailing platform are determined according to the sum of the reward resources of each order completion type. The activity resources can be determined according to the accurate driver quantity corresponding to the order completion type and the corresponding reward reference value. Since the driver quantity corresponding to the order completion type is determined by selecting different prediction models according to the types of registered drivers, the activity effect can be accurately calibrated, and the determined driver quantity has high accuracy.

[0140] As an alternative implementation, as Figure 9 shown, the resource determination method provided in the embodiments of the present application may include the following specific steps:

[0141] S901, Obtain the activity configuration information of the target online car-hailing platform and the historical performance data of each registered driver.

[0142] S902, For each registered driver, determine the performance type of the registered driver according to the identification information of the registered driver.

[0143] S903, For each registered driver, determine the service platform type of the registered driver according to the service platform information of the registered driver.

[0144] Among them, the service platform type includes a single service platform and a multi-service platform.

[0145] S904, For each registered driver, determine the type of the registered driver according to the performance type of the registered driver and the service platform type of the registered driver.

[0146] S905, For each registered driver, if the service platform type of the registered driver is a single service platform type, determine the target prediction model corresponding to the online car-hailing platform served by the registered driver and the performance type of the registered driver from the single-platform model list.

[0147] Among them, the single-platform model list is used to store the corresponding relationships between different online car-hailing platforms and performance types and prediction models.

[0148] S906, For each registered driver, if the service platform type of the registered driver is a multi-service platform type, obtain the target platform combination corresponding to the registered driver. The target platform combination includes multiple online car-hailing platforms served by the registered driver. Determine the sub-list of prediction models corresponding to the target platform combination and the performance type of the registered driver from the multi-platform model list, and determine the target prediction model corresponding to the target online car-hailing platform from the sub-list of prediction models.

[0149] Among them, the multi-platform model list is used to store the corresponding relationships between different platform combinations and performance types and sub-lists of prediction models, and the sub-list of prediction models is used to store the corresponding relationships between different online car-hailing platforms and prediction models.

[0150] The target prediction model includes an online probability model and an order completion type model.

[0151] S907, For each registered driver, input the activity configuration information and the historical performance data of the registered driver into the online probability model, and determine the online probability of the registered driver according to the output of the online probability model.

[0152] S908. For each registered driver, input the activity configuration information and the historical performance data of the registered driver into an order completion type model, and determine the possible order completion types of the registered driver according to the output of the order completion type model.

[0153] S909. For each registered driver, determine the probability estimate of the possible order completion types of the registered driver, and determine the expected value of the driver quantity contribution of the registered driver to the order completion type according to the probability estimate and the online probability.

[0154] S910. For each order completion type, determine the driver quantity corresponding to the order completion type according to the sum of the expected values of the driver quantity contributions of each registered driver to the order completion type.

[0155] S911. For each order completion type, determine the reward resources of the order completion type according to the driver quantity and the reward reference value.

[0156] S912. Determine the activity resources of the target online car-hailing platform according to the sum of the reward resources of each order completion type.

[0157] Optionally, as Figure 10 shown, it is the framework of the resource determination method of the embodiment of the present application. Taking the online probability model as a logistic regression model and the order completion type model as a random forest model as an example, determine the corresponding online probability model and order completion type model according to the type of the registered driver, so as to determine the driver quantity corresponding to each order completion type, and then determine the activity resources of the target online car-hailing platform according to the driver quantity corresponding to different order completion types and the reward reference values corresponding to different order completion types.

[0158] It should be understood that although each step in the flowcharts involved in the above-described embodiments is sequentially shown according to the indication of the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same moment, but can be executed at different moments. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0159] Based on the same inventive concept, an embodiment of the present application further provides a resource determination device for implementing the above-mentioned resource determination method. The solution provided by this device for solving problems is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the resource determination device provided below can refer to the limitations on the resource determination method in the foregoing, and will not be repeated here.

[0160] In an exemplary embodiment, as Figure 11 shown, a resource determination device 1100 is provided, including: an acquisition module 1101, a first determination module 1102, a second determination module 1103, and a third determination module 1104, where:

[0161] The acquisition module 1101 is configured to acquire the activity configuration information of the target online car-hailing platform and the historical performance data of each registered driver.

[0162] The first determination module 1102 is configured to, for each registered driver, determine the target prediction model corresponding to the registered driver, input the activity configuration information and the historical performance data of the registered driver into the target prediction model, and determine the online probability of the registered driver and the possible order completion types according to the target prediction model.

[0163] The second determination module 1103 is configured to determine the number of drivers corresponding to different order completion types according to the online probabilities of the registered drivers and the possible order completion types.

[0164] The third determination module 1104 is configured to determine the activity resources of the target online car-hailing platform according to the number of drivers corresponding to different order completion types and the reward reference values corresponding to different order completion types.

[0165] In one embodiment, the target prediction model includes an online probability model and an order completion type model. The first determination module 1102 is specifically configured to input the activity configuration information and the historical performance data of the registered driver into the online probability model, and determine the online probability of the registered driver according to the output of the online probability model; input the activity configuration information and the historical performance data of the registered driver into the order completion type model, and determine the possible order completion types of the registered driver according to the output of the order completion type model.

[0166] In one embodiment, the first determination module 1102 is specifically configured to obtain the type of the registered driver; and determine the target prediction model from a preset model list according to the type of the registered driver.

[0167] In one embodiment, the first determination module 1102 is specifically configured to determine the performance type of a registered driver according to the identification information of the registered driver; determine the service platform type of the registered driver according to the service platform information of the registered driver, where the service platform type includes a single service platform and a multi-service platform; and determine the type of the registered driver according to the performance type of the registered driver and the service platform type of the registered driver.

[0168] In one embodiment, the preset model list includes a single-platform model list and a multi-platform model list. The first determination module 1102 is specifically configured to, if the service platform type of the registered driver is a single service platform type, determine a target prediction model corresponding to the online car-hailing platform served by the registered driver and the performance type of the registered driver from the single-platform model list, where the single-platform model list is used to store the correspondence between different online car-hailing platforms and performance types and prediction models; if the service platform type of the registered driver is a multi-service platform type, obtain the target platform combination corresponding to the registered driver, where the target platform combination includes multiple online car-hailing platforms served by the registered driver, determine a sub-list of prediction models corresponding to the target platform combination and the performance type of the registered driver from the multi-platform model list, and determine the target prediction model corresponding to the target online car-hailing platform from the sub-list of prediction models, where the multi-platform model list is used to store the correspondence between different platform combinations and performance types and sub-lists of prediction models, and the sub-list of prediction models is used to store the correspondence between different online car-hailing platforms and prediction models.

[0169] In one embodiment, the second determination module 1103 is specifically configured to, for each registered driver, determine the probability estimate of the possible order completion type of the registered driver, and determine the expected value of the driver quantity contribution of the registered driver to the order completion type according to the probability estimate and the online probability; for each order completion type, determine the driver quantity corresponding to the order completion type according to the sum of the expected values of the driver quantity contributions of each registered driver to the order completion type.

[0170] In one embodiment, the third determination module 1104 is specifically configured to, for each order completion type, determine the reward resources of the order completion type according to the driver quantity and the reward reference value; and determine the activity resources of the target online car-hailing platform according to the sum of the reward resources of each order completion type.

[0171] Each module in the above resource determination device can be implemented in whole or in part by software, hardware, and their combination. Each of the above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form so that the processor can call and execute the operations corresponding to each of the above modules.

[0172] In an exemplary embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be asFigure 12 As shown in the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a resource determination method.

[0173] Those skilled in the art can understand that Figure 12 the structure shown in the figure is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0174] In an exemplary embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, the steps described in any of the above method embodiments are implemented.

[0175] In an embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, the steps described in any of the above method embodiments are implemented.

[0176] In an embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, the steps described in any of the above method embodiments are implemented.

[0177] It should be noted that the user information (including but not limited to driver device information, driver personal information, etc.) and data (including but not limited to data for analysis (such as historical performance data and activity configurations of each registered driver), stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0178] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, artificial intelligence (AI) processors, etc., without limitation.

[0179] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as within the scope recorded in the present application.

[0180] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A resource determination method, characterized in that, The method includes: Obtaining the activity configuration information of the target online car-hailing platform and the historical performance data of each registered driver; For each of the registered drivers, determining the target prediction model corresponding to the registered driver, and inputting the activity configuration information and the historical performance data of the registered driver into the target prediction model, and determining the online probability of the registered driver and the possible order completion types according to the output of the target prediction model; Determining the number of drivers corresponding to different order completion types according to the online probabilities and the possible order completion types of each of the registered drivers; Determining the activity resources of the target online car-hailing platform according to the number of drivers corresponding to different order completion types and the reward reference values corresponding to different order completion types.

2. The method according to claim 1, characterized in that, The target prediction model includes an online probability model and an order completion type model. The inputting the activity configuration information and the historical performance data of the registered driver into the target prediction model, and determining the online probability of the registered driver and the possible order completion types according to the output of the target prediction model includes: Inputting the activity configuration information and the historical performance data of the registered driver into the online probability model, and determining the online probability of the registered driver according to the output of the online probability model; Inputting the activity configuration information and the historical performance data of the registered driver into the order completion type model, and determining the possible order completion types of the registered driver according to the output of the order completion type model.

3. The method according to claim 1, wherein The determining the target prediction model corresponding to the registered driver includes: Obtaining the type of the registered driver; Determining the target prediction model from a preset model list according to the type of the registered driver.

4. The method according to claim 3, wherein The obtaining the type of the registered driver includes: Determining the performance type of the registered driver according to the identification information of the registered driver; Determining the service platform type of the registered driver according to the service platform information of the registered driver, where the service platform type includes a single service platform and a multi-service platform; Determining the type of the registered driver according to the performance type of the registered driver and the service platform type of the registered driver.

5. The method according to claim 4, wherein The preset model list includes a single-platform model list and a multi-platform model list. The determining the target prediction model from the preset model list according to the type of the registered driver includes: If the service platform type of the registered driver is a single service platform type, determining the target prediction model corresponding to the online car-hailing platform served by the registered driver and the performance type of the registered driver from the single-platform model list, where the single-platform model list is used to store the corresponding relationships between different online car-hailing platforms and performance types and prediction models; If the service platform type of the registered driver is a multi-service platform type, obtain the target platform combination corresponding to the registered driver, where the target platform combination includes multiple online car-hailing platforms served by the registered driver, determine a sub-list of prediction models corresponding to the target platform combination and the performance type of the registered driver from the multi-platform model list, and determine the target prediction model corresponding to the target online car-hailing platform from the sub-list of prediction models. The multi-platform model list is used to store the correspondence between different platform combinations and performance types and the sub-list of prediction models, and the sub-list of prediction models is used to store the correspondence between different online car-hailing platforms and prediction models.

6. The method according to claim 1, characterized in that The determining of the number of drivers corresponding to different order completion types according to the online probabilities and order completion types of the registered drivers includes: For each registered driver, determine the probability estimate of the possible order completion types of the registered driver, and determine the expected value of the contribution of the registered driver to the number of drivers of the order completion type according to the probability estimate and the online probability; For each order completion type, determine the number of drivers corresponding to the order completion type according to the sum of the expected values of the contribution of each registered driver to the number of drivers of the order completion type.

7. The method according to claim 1, wherein The determining of the activity resources of the target online car-hailing platform according to the number of drivers corresponding to different order completion types and the reward reference values corresponding to different order completion types includes: For each order completion type, determine the reward resources of the order completion type according to the number of drivers and the reward reference value; Determine the activity resources of the target online car-hailing platform according to the sum of the reward resources of each order completion type.

8. A method and apparatus for resource determination, characterized in that, The device includes: An acquisition module, configured to acquire the activity configuration information of the target online car-hailing platform and the historical performance data of each registered driver; A first determination module, configured to, for each registered driver, determine the target prediction model corresponding to the registered driver, input the activity configuration information and the historical performance data of the registered driver into the target prediction model, and determine the online probability and the possible order completion types of the registered driver according to the target prediction model; A second determination module, configured to determine the number of drivers corresponding to different order completion types according to the online probabilities and the possible order completion types of the registered drivers; A third determination module, configured to determine the activity resources of the target online car-hailing platform according to the number of drivers corresponding to different order completion types and the reward reference values corresponding to different order completion types.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images