Area supply object screening method, device, equipment and medium

Abstract

The present disclosure provides a regional supply object screening method, device, equipment and medium, the method comprises: obtaining a trained target regional supply object screening model of a target supply region; obtaining a candidate supply object of the target supply region, and based on the target regional supply object screening model, obtaining a predicted allocation rate and a predicted supply amount of the candidate supply object in the target supply region within a predicted time range; and screening a target supply object of the target supply region from the candidate supply object according to the predicted allocation rate and the predicted supply amount. The adaptation degree between the target supply object and the target supply region is improved, the waste of the product selection quota of the target supply region is avoided, the screening efficiency of the target supply object of the target supply region is improved, the possibility of the target supply region achieving the category operation target of the headquarters is improved, and the product selection method and product selection effect of the offline store as the target supply region are optimized.

Description

Technical Field

[0001] This disclosure relates to the field of data processing, and in particular to a method, apparatus, equipment, and medium for screening regional supply targets. Background Technology

[0002] With the development of technology, in the field of new retail, the types of products that offline stores can sell in-store are limited to a certain extent due to factors such as store size and shelf capacity.

[0003] In related technologies, headquarters staff can select products for offline stores to sell based on the store's level. However, among the products selected for offline stores using this method, some products may have performed poorly in the historical sales of offline stores. In this scenario, the staff of the offline stores may not allocate these products to the stores for sale, resulting in a waste of the offline stores' product selection quota.

[0004] Correspondingly, in related technologies, offline store staff can select products suitable for offline stores from all products based on the historical sales data of the products and sell them in the store. In this scenario, offline store staff cannot go through all products, resulting in fewer product categories sold in offline stores, and there is a possibility that the category operation goals of the headquarters cannot be achieved. Summary of the Invention

[0005] This disclosure aims to at least partially address one of the technical problems in the related art.

[0006] Therefore, the first aspect of this disclosure proposes a method for screening regional supply targets.

[0007] The second aspect of this disclosure proposes a regional supply target screening device.

[0008] The third aspect of this disclosure proposes an electronic device.

[0009] The fourth aspect of this disclosure provides for a computer-readable storage medium.

[0010] The first aspect of this disclosure proposes a method for screening regional supply targets. The method includes: acquiring a trained target regional supply target screening model for a target supply region; acquiring candidate supply targets for the target supply region, and based on the target regional supply target screening model, obtaining a predicted allocation rate and a predicted supply quantity of the candidate supply targets within the target supply region within a predicted time range; and selecting target supply targets for the target supply region from the candidate supply targets according to the predicted allocation rate and the predicted supply quantity.

[0011] In addition, the regional supply target selection method proposed in the first aspect of this disclosure may also have the following additional technical features:

[0012] According to one embodiment of this disclosure, obtaining a trained target region supply object screening model for the target supply region includes: obtaining a candidate region supply object screening model to be trained; obtaining a sample region feature set of a sample region and a sample supply object feature set of the sample supply objects in the sample region, wherein the sample region feature set includes at least one of the sample region features and sample region allocation features of the sample region, and the sample supply object feature set includes at least one of the sample supply object features and sample region supply features of the sample supply objects; obtaining training samples for the candidate region supply object screening model based on the sample region feature set and the sample supply object feature set; inputting the training samples into the candidate region supply object screening model for model training until training is completed, thereby obtaining the trained target region supply object screening model for the target supply region.

[0013] According to one embodiment of this disclosure, obtaining training samples for the candidate region supply object screening model based on the sample region feature set and the sample supply object feature set includes: concatenating the sample region features and sample region allocation features in the sample region feature set, and the sample supply object feature set including the sample supply object features and sample region supply features, to obtain concatenated sample features; obtaining the historical region allocation rate and historical region supply amount of the sample supply object in the sample region within a historical time range; and using the historical region allocation rate and historical region supply amount as sample labels for the concatenated sample features to obtain the training samples for the candidate region supply object screening model.

[0014] According to one embodiment of this disclosure, the step of inputting the training samples into the candidate region supply object screening model for model training until training is completed, to obtain a trained target region supply object screening model, includes: obtaining a training task set of the candidate region supply object screening model based on the training samples; obtaining candidate task model combinations of each training task in the training task set through the candidate gating unit in the candidate region supply object screening model; inputting the training samples into each candidate task model combination to obtain the task results of each training task output by each candidate task model combination; obtaining the training loss of the candidate region supply object screening model based on the task results of each training task and the sample label of the training samples; adjusting the model parameters of the candidate region supply object screening model based on the training loss, and returning to obtain the next training sample to continue model training of the parameter-adjusted candidate region supply object screening model until training is completed, to obtain the trained target region supply object screening model.

[0015] According to one embodiment of this disclosure, obtaining the candidate task model combination of each training task in the training task set through the candidate gating unit in the candidate region supply object screening model includes: obtaining the unit task model set of the candidate region supply object screening model; for any training task, obtaining at least one candidate unit task model for executing the training task from the unit task model set through the candidate gating unit, and determining the combination of the at least one candidate unit task model as the candidate task model combination of the training task.

[0016] According to one embodiment of this disclosure, the step of adjusting the model parameters of the candidate region supply object screening model based on the training loss, and returning to obtain the next training sample to continue training the parameter-adjusted candidate region supply object screening model until training ends, thereby obtaining the trained target region supply object screening model, includes: adjusting the model iteration parameters of the candidate region supply object screening model based on the training loss to obtain a parameter-adjusted candidate region supply object screening model, wherein the model iteration parameters include at least the parameters of each candidate task model in the candidate task model combination and the parameters of the candidate gating unit; returning to obtain the next training sample to continue training the parameter-adjusted candidate region supply object screening model until training ends, thereby obtaining the trained target region supply object screening model.

[0017] According to one embodiment of this disclosure, the step of obtaining candidate supply objects for the target supply region and obtaining the predicted allocation rate and predicted supply quantity of the candidate supply objects in the target supply region within a predicted time range based on the target region supply object screening model includes: obtaining a target gating unit in the target region supply object screening model; obtaining, through the target gating unit, a combination of allocation rate prediction task models for the allocation rate prediction task corresponding to the predicted allocation rate, and a combination of supply quantity prediction task models for the supply quantity corresponding to the predicted supply quantity from the target unit task model set; obtaining a target region feature set for the target supply region and a combination of candidate supply object features for the candidate supply objects, and obtaining the model inputs of the allocation rate prediction task model combination and the supply quantity prediction task model combination according to the target region feature set and the candidate supply object feature set; and obtaining, based on the model inputs, the predicted allocation rate of the candidate supply objects in the target supply region within the predicted time range output by the allocation rate prediction task model combination, and the predicted supply quantity of the candidate supply objects in the target supply region within the predicted time range output by the supply quantity prediction task model combination.

[0018] According to one embodiment of this disclosure, obtaining the target region feature set of the target supply region and the candidate supply object feature set of the candidate supply objects, and obtaining the model inputs of the allocation rate prediction task model combination and the supply quantity prediction task model combination based on the target region feature set and the candidate supply object feature set, includes: obtaining the target region features and target region allocation features of the target supply region in the target region feature set, and the candidate supply object features and historical supply quantity features of the candidate supply objects in the target supply region in the candidate supply object feature set; concatenating the target region features, the target region allocation features, the candidate supply object features, and the historical supply quantity features to obtain the model inputs of the allocation rate prediction task model combination and the supply quantity prediction task model combination.

[0019] According to one embodiment of this disclosure, the step of selecting target supply objects for the target supply region from the candidate supply objects based on the predicted allocation rate and the predicted supply volume includes: obtaining the supply value of the candidate supply objects in the target supply region, and obtaining the total predicted supply value of the candidate supply objects in the target supply region within the predicted time range based on the supply value and the predicted supply volume; obtaining the candidate supply score of the candidate supply objects within the predicted time range based on the predicted allocation rate, the predicted supply volume, and the total predicted supply value, and selecting the target supply objects for the target supply region within the predicted time range from the candidate supply objects based on the candidate supply score.

[0020] A second aspect of this disclosure provides a regional supply target screening device, comprising: an acquisition module for acquiring a trained target regional supply target screening model for a target supply region; a prediction module for acquiring candidate supply targets for the target supply region and, based on the target regional supply target screening model, obtaining a predicted allocation rate and a predicted supply quantity of the candidate supply targets within the target supply region within a predicted time range; and a screening module for screening target supply targets for the target supply region from the candidate supply targets according to the predicted allocation rate and the predicted supply quantity.

[0021] In addition, the regional supply target screening device proposed in the second aspect of this disclosure may also have the following additional technical features:

[0022] According to one embodiment of this disclosure, the acquisition module is further configured to: acquire a candidate region supply object screening model to be trained; acquire a sample region feature set of the sample region and a sample supply object feature set of the sample supply objects of the sample region, wherein the sample region feature set includes at least one of the sample region features and sample region allocation features of the sample region, and the sample supply object feature set includes at least one of the sample supply object features and sample region supply features of the sample supply objects; obtain training samples for the candidate region supply object screening model based on the sample region feature set and the sample supply object feature set; input the training samples into the candidate region supply object screening model for model training until the training is completed, thereby obtaining the trained target region supply object screening model for the target supply region.

[0023] According to one embodiment of this disclosure, the acquisition module is further configured to: concatenate the sample region features and the sample region allocation features in the sample region feature set, and the sample supply object feature set including the sample supply object features and the sample region supply features, to obtain concatenated sample concatenation features; acquire the historical region allocation rate and historical region supply amount of the sample supply object in the sample region within a historical time range; and use the historical region allocation rate and historical region supply amount as sample labels of the concatenated sample features to obtain the training samples of the candidate region supply object screening model.

[0024] According to one embodiment of this disclosure, the acquisition module is further configured to: acquire a set of training tasks for the candidate region supply object screening model based on the training samples; acquire candidate task model combinations for each training task in the training task set through a candidate gating unit in the candidate region supply object screening model; input the training samples into each candidate task model combination to obtain the task results of each training task output by each candidate task model combination; acquire the training loss of the candidate region supply object screening model based on the task results of each training task and the sample labels of the training samples; adjust the model parameters of the candidate region supply object screening model based on the training loss, and return to acquire the next training sample to continue training the parameter-adjusted candidate region supply object screening model until training ends, thereby obtaining the trained target region supply object screening model.

[0025] According to one embodiment of this disclosure, the acquisition module is further configured to: acquire a set of unit task models for the candidate region supply object screening model; for any training task, acquire at least one candidate unit task model for executing the training task from the set of unit task models through the candidate gating unit, and determine the combination of the at least one candidate unit task model as the candidate task model combination for the training task.

[0026] According to one embodiment of this disclosure, the acquisition module is further configured to: adjust the model iteration parameters of the candidate region supply object screening model according to the training loss to obtain a parameter-adjusted candidate region supply object screening model, wherein the model iteration parameters include at least the parameters of each candidate task model in the candidate task model combination and the parameters of the candidate gating unit; return to acquire the next training sample to continue model training of the parameter-adjusted candidate region supply object screening model until the training ends, and obtain the trained target region supply object screening model.

[0027] According to one embodiment of this disclosure, the prediction module is further configured to: obtain target gating units in the target area supply object screening model; obtain, through the target gating units, a combination of allocation rate prediction task models for the allocation rate prediction task corresponding to the predicted allocation rate, and a combination of supply quantity prediction task models for the supply quantity prediction task corresponding to the predicted supply quantity from the target unit task model set; obtain a target area feature set for the target supply area, and a combination of candidate supply object features for the candidate supply objects, and obtain the model inputs for the allocation rate prediction task model combination and the supply quantity prediction task model combination based on the target area feature set and the candidate supply object feature set; and obtain, based on the model inputs, the predicted allocation rate of the candidate supply objects in the target supply area within the prediction time range output by the allocation rate prediction task model combination, and the predicted supply quantity of the candidate supply objects in the target supply area within the prediction time range output by the supply quantity prediction task model combination.

[0028] According to one embodiment of this disclosure, the prediction module is further configured to: obtain the target region features and target region allocation features of the target supply region in the target region feature set, and the candidate supply object features and historical supply volume features of the candidate supply object in the candidate supply object feature set; and concatenate the target region features, the target region allocation features, the candidate supply object features, and the historical supply volume features to obtain the model inputs of the allocation rate prediction task model combination and the supply volume prediction task model combination, respectively.

[0029] According to one embodiment of this disclosure, the screening module is further configured to: obtain the supply value of the candidate supply object in the target supply area, and obtain the total predicted supply value of the candidate supply object in the target supply area within the prediction time range based on the supply value and the predicted supply quantity; obtain the candidate supply score of the candidate supply object within the prediction time range based on the predicted allocation rate, the predicted supply quantity and the total predicted supply value, and screen the target supply object in the target supply area within the prediction time range from the candidate supply objects based on the candidate supply score.

[0030] A third aspect of this disclosure provides an electronic device, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute instructions to implement the regional supply object screening method as described in the first aspect and / or the regional supply object screening method as described in the second aspect.

[0031] This fourth aspect of the disclosure provides a computer-readable storage medium that, when executed by a processor of an electronic device, enables the electronic device to perform the regional supply object screening method as described in the first aspect and / or the regional supply object screening method as described in the second aspect.

[0032] The method and apparatus for selecting regional supply targets proposed in this disclosure acquire a trained target region supply target selection model for the target supply region, and obtain the predicted allocation rate and predicted supply volume of candidate supply targets through the target region supply target selection model. Then, based on the predicted allocation rate and predicted supply volume, the target supply targets for the target supply region are selected from the candidate supply targets. In this disclosure, selecting target supply targets for the target supply region from candidate supply targets based on the predicted allocation rate and predicted supply volume improves the fit between target supply targets and target supply regions, avoids waste of product selection quotas in the target supply region, and improves the selection efficiency of target supply targets by obtaining the predicted allocation rate and predicted supply volume of candidate supply targets based on the target region supply target selection model. This avoids the situation where the target supply region has a limited variety of target supply targets due to the inability of staff to traverse all candidate supply targets, increases the likelihood of the target supply region achieving the headquarters' category operation goals, optimizes the product selection method and effect for offline stores in the target supply region, and thus optimizes the product sales of offline stores in the target supply region.

[0033] It should be understood that the description herein is not intended to identify key or essential features of the embodiments thereof, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0034] The above and / or additional aspects and advantages of this disclosure will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, in which:

[0035] Figure 1 This is a flowchart illustrating a regional supply target screening method according to an embodiment of the present disclosure;

[0036] Figure 2 This is a schematic diagram illustrating product selection in an offline store according to an embodiment of the present disclosure.

[0037] Figure 3 This is a flowchart illustrating a regional supply target screening method according to another embodiment of the present disclosure;

[0038] Figure 4 This is a schematic diagram of a regional supply target screening model according to an embodiment of the present disclosure;

[0039] Figure 5 This is a flowchart illustrating a regional supply target screening method according to another embodiment of the present disclosure;

[0040] Figure 6 This is a schematic diagram of the structure of a regional supply object screening device according to an embodiment of the present disclosure;

[0041] Figure 7 This is a block diagram of an electronic device according to an embodiment of the present disclosure. Detailed Implementation

[0042] Embodiments of this disclosure are described in detail below. Examples of these embodiments are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting this disclosure.

[0043] The following description, with reference to the accompanying drawings, outlines a method, apparatus, device, and medium for screening regional supply targets according to embodiments of this disclosure.

[0044] Figure 1 This is a flowchart illustrating a regional supply target selection method according to an embodiment of the present disclosure, as shown below. Figure 1 As shown, the method includes:

[0045] S101, Obtain the trained target area supply object screening model for the target supply area.

[0046] In some implementations, offline stores face certain limitations in the types of products they can sell due to factors such as store size, merchandise capacity, and display space.

[0047] In this embodiment of the disclosure, a trained sales product selection model can be used to select sales products for offline stores. The offline stores that need to select sales products can be marked as target supply areas, and the products sold in the offline stores can be marked as target supply objects in the target supply areas.

[0048] As an example, it can be Figure 2 The offline stores 1, 2, 3, ..., n shown are labeled as target supply area 1, target supply area 2, target supply area 3, ..., target supply area n, respectively.

[0049] like Figure 2 As shown, the products sold in offline store 1 are product 2, product 3 and product m. Therefore, product 2, product 3 and product m can be marked as target supply object 2, target supply object 3 and target supply object m of target supply area 1.

[0050] Optionally, a model can be trained to screen supply objects for the target supply area, and the trained model can be labeled as the trained target area supply object screening model for the target supply area.

[0051] In this scenario, a pre-trained target area supply object screening model can be used to screen offline stores that are in the target supply area.

[0052] It should be noted that the target area supply object screening model can be built based on a neural network (Multi-gate Mixture-of-Experts, MMoE) or other neural networks; no specific limitation is made here.

[0053] S102, obtain candidate supply objects in the target supply area, and based on the target area supply object screening model, obtain the predicted allocation rate and predicted supply volume of candidate supply objects in the target supply area within the predicted time range.

[0054] In this embodiment of the disclosure, there may be multiple candidate supply objects in the target supply area, and the target supply object in the target supply area can be obtained by screening from the multiple candidate supply objects.

[0055] This can be understood as follows: an offline store may have multiple products to be screened for in-store delivery. The store can then select from these multiple products to obtain the target products that will be actually sold in that offline store. The products to be screened are the candidate supply objects for offline stores marked as the target supply area, and the target products that are actually sold in that offline store are the target supply objects for offline stores marked as the target supply area.

[0056] Optionally, staff at offline stores may obtain products that sell well in the store based on the store's historical sales data and transfer them from the warehouse to the store as products actually sold in the offline store. Therefore, for product 1 and product 2, when the transfer rate of product 1 is greater than that of product 2, it can be determined that product 1 sells better in offline stores than product 2.

[0057] In this scenario, the allocation rate and sales volume of candidate suppliers in the target supply area can be predicted within the predicted time range, and the target suppliers in the target supply area can be selected from the candidate suppliers based on the prediction results.

[0058] For any candidate supply object, a pre-trained target area supply object screening model can be obtained, and the allocation rate and supply volume of the candidate supply object in the target supply area can be predicted through the target area supply object screening model, thereby obtaining the predicted allocation rate and predicted supply volume of the candidate supply object in the target supply area within the prediction time range.

[0059] S103. Based on the predicted allocation rate and predicted supply volume, select the target supply objects for the target supply area from the candidate supply objects.

[0060] In this embodiment of the disclosure, the predicted allocation rate and predicted supply quantity of each of the candidate supply objects can be obtained, and the predicted allocation rate and predicted supply quantity of each of the candidate supply objects can be compared to obtain the target supply object of the target supply area among the candidate supply objects.

[0061] Optionally, a set number of candidate supply objects can be selected from the predicted allocation rates of all candidate supply objects from high to low, and a set number of candidate supply objects can be selected from the predicted supply quantities of all candidate supply objects from high to low. The candidate supply objects included in the intersection of the two selected candidate supply object groups are determined as the target supply objects of the target supply area.

[0062] Optionally, the candidate supply objects and target supply areas can be matched based on the predicted allocation rate and predicted supply volume, and a portion of the candidate supply objects whose matching degree meets the preset conditions can be selected from all the candidate supply objects as the target supply objects of the target supply area.

[0063] The regional supplier selection method proposed in this disclosure obtains a trained target region supplier selection model for the target supply region, and uses this model to obtain the predicted allocation rate and predicted supply volume of candidate suppliers. Then, based on the predicted allocation rate and predicted supply volume, the target suppliers for the target supply region are selected from the candidate suppliers. This disclosure improves the fit between target suppliers and target supply regions by selecting target suppliers from candidate suppliers based on the predicted allocation rate and predicted supply volume, avoiding waste of product selection quotas in the target supply region. Obtaining the predicted allocation rate and predicted supply volume of candidate suppliers based on the target region supplier selection model improves the selection efficiency of target suppliers in the target supply region, avoiding the situation where the target supply region staff cannot traverse all candidate suppliers, resulting in a limited variety of target suppliers in the target supply region. This increases the likelihood that the target supply region will achieve the headquarters' category operation goals, optimizes the product selection method and effect for offline stores in the target supply region, and ultimately optimizes the product sales of offline stores in the target supply region.

[0064] In the above embodiments, the acquisition of the target area supply object screening model can be combined with... Figure 3 To understand further, Figure 3 This is a flowchart illustrating a regional supply target screening method according to another embodiment of the present disclosure, as shown below. Figure 3 As shown, the method includes:

[0065] S301, Obtain the candidate region supply object screening model to be trained.

[0066] S302, obtain the sample region feature set of the sample region and the sample supply object feature set of the sample supply object of the sample region, wherein the sample region feature set includes at least one of the sample region features and the sample region allocation features, and the sample supply object feature set includes at least one of the sample supply object features and the sample region supply features.

[0067] In this embodiment of the disclosure, the sample region has a corresponding sample region feature set, wherein the sample region feature set includes at least one of the sample region features and the sample region allocation features.

[0068] Accordingly, objects supplied within the sample area can be marked as sample supply objects within the sample area. Each sample supply object has a corresponding set of sample supply object features, which includes at least one of the sample supply object features of the sample supply object and the sample area supply features of the sample object within the sample area.

[0069] This can be understood as follows: in a scenario where the sample area is the sample offline stores and the sample supply object is the sample products sold in the sample offline stores, the sample area feature set includes at least one of the sample store features and sample store allocation features of the sample offline stores, and the sample supply object feature set includes at least one of the sample product features and sample supply features of the sample products sold in the sample offline stores.

[0070] As an example, let's define the sample area as offline stores. The store features labeled as sample area characteristics and the store transfer features labeled as sample area transfer characteristics, as well as the sample supply object characteristics and the sample area supply characteristics of the sample objects within the sample area, can be extracted from the store data and transfer data in the table below:

[0071]

[0072] Among them, the products marked as sample supply objects can be understood as the smallest stock keeping unit (SKU) of offline stores, and the standardized product unit (SPU) in the table above can be understood as the product category of products marked as sold in offline stores.

[0073] In this scenario, store data of sample stores marked as sample areas, sales data and product data of products marked as sample supply objects, and transfer data of products in different stores and different regions can be obtained from the above table. This results in a sample area feature set including at least one of the sample area features and sample area transfer features, and a sample supply object feature set including at least one of the sample supply object features and sample area supply features.

[0074] S303, Based on the sample region feature set and the sample supply object feature set, obtain the training samples for the candidate region supply object screening model.

[0075] Optionally, the sample region features and sample region allocation features in the sample region feature set, as well as the sample supply object feature set including sample supply object features and sample region supply features, can be spliced ​​together to obtain the spliced ​​sample spliced ​​features.

[0076] In this embodiment of the disclosure, sample region features and sample region allocation features, as well as the splicing order of sample supply object features and sample region supply features, can be obtained. Based on the splicing order, the sample region features, sample region allocation features, sample supply object features, and sample region supply features are spliced ​​together using the feature splicing method in related technologies to obtain the spliced ​​features. The spliced ​​features are then marked as sample splicing features.

[0077] Optionally, the historical allocation rate and supply volume of the sample supply objects within the sample area can be obtained within a historical time range, and the historical allocation rate and supply volume can be used as sample labels for sample splicing features to obtain training samples for the candidate area supply object screening model.

[0078] In this embodiment of the disclosure, the allocation rate of the sample supply object within a historical time range can be obtained and marked as the historical area allocation rate, and the supply amount of the sample supply object within the sample area within a historical time range can be marked as the historical area supply amount.

[0079] Furthermore, the historical regional allocation rate and historical regional supply volume are used as sample labels for sample splicing features, thereby obtaining training samples for the candidate regional supply object screening model.

[0080] It should be noted that the training samples may include samples used to train the allocation rate of the sample supply objects output by the candidate region supply object screening model, and samples used to train the supply quantity of the sample supply objects output by the candidate region supply object screening model.

[0081] The samples used to train the allocation rate of the sample supply objects output by the candidate region supply object screening model can include positive samples and negative samples. Positive samples can be samples composed of sample supply regions and sample supply objects that have been allocated after distribution, while negative samples can be samples composed of sample supply regions and sample supply objects that have not been allocated after distribution.

[0082] S304. Input the training samples into the candidate region supply object screening model for model training until the training is completed, and obtain the trained target region supply object screening model for the target supply region.

[0083] Optionally, obtain the training task set of the candidate region supply object screening model based on the training samples.

[0084] In this embodiment of the disclosure, the candidate region supply object screening model is a multi-task model, and the training samples input when training the candidate region supply object screening model are multi-task training samples.

[0085] In a scenario where the candidate region supply target selection model outputs the supply quantity and allocation rate, the training tasks of the candidate region supply target selection model include training tasks corresponding to the supply quantity and training tasks corresponding to the allocation rate.

[0086] In this scenario, the training samples for the candidate region supply object screening model include training tasks corresponding to the supply quantity and training tasks corresponding to the allocation rate. The set of training tasks corresponding to the supply quantity and training tasks corresponding to the allocation rate can be marked as the training task set of the training samples.

[0087] Optionally, candidate gating units in the candidate region supply object screening model are used to obtain candidate task model combinations for each training task in the training task set.

[0088] As an example, the candidate region supply object screening model can be as follows: Figure 4 As shown, by Figure 4 It can be seen that the candidate region supply object screening model includes gating unit A and gating unit B. Through gating unit A and gating unit B, the corresponding task model combination can be selected for each training task in the training task set, thereby realizing multi-task training of the candidate region supply object model.

[0089] Optionally, a set of unit task models for the candidate region supply object screening model can be obtained. For any training task, at least one candidate unit task model for executing the training task can be obtained from the set of unit task models through the candidate gating unit, and the combination of at least one candidate unit task model can be determined as the candidate task model combination for the training task.

[0090] In this embodiment of the disclosure, the training task for the training samples may include a training task corresponding to the allocation rate and a training task corresponding to the supply quantity. As an example, it can be achieved through... Figure 4 The gate unit A shown obtains at least one unit task model from the set of unit task models to execute the training task corresponding to the allocation rate.

[0091] Specifically, gating unit A can select unit task models from the set of unit task models that process training tasks corresponding to the allocation rate and mark them as candidate unit task models. The combination of these candidate unit task models can be marked as candidate task model combination A for executing training tasks corresponding to the allocation rate.

[0092] Accordingly, it can be done through Figure 4 The gate unit B shown obtains at least one unit task model from the set of unit task models to execute the training task corresponding to the supply quantity.

[0093] Specifically, the gating unit B can select unit task models from the set of unit task models that process the training tasks corresponding to the supply quantity and mark them as candidate unit task models. The combination of these candidate unit task models is then marked as candidate task model combination B for executing the training tasks corresponding to the supply quantity.

[0094] Optionally, the training samples are input into the candidate task model combination to obtain the task results of each training task output by the candidate task model combination.

[0095] As an example, such as Figure 4 As shown, training samples can be input into candidate task model combination A, which is composed of at least one candidate unit task model selected from the set of unit task models by gating unit A, and candidate task model combination B, which is composed of at least one candidate unit task model selected from the set of unit task models by gating unit B.

[0096] like Figure 4 As shown, candidate task model combination A can be obtained by performing a training task based on training samples to determine the allocation rate. Figure 4 The output A shown, and the candidate task model combination B, can be obtained by performing a training task based on the training samples. Figure 4 The output B shown is then implemented. Figure 4 The candidate region supply screening model shown is trained using a multi-task approach.

[0097] It should be noted that, Figure 4 Each unit task model in the illustrated set of unit task models can be a multi-layered, small-scale fully connected layer. Gating unit A and gating unit B share the set of unit task models, where each unit task model has its corresponding prediction direction.

[0098] In this scenario, the gating unit can obtain the prediction direction of each unit task model in the unit task model set, and select at least one unit task model that matches the training task corresponding to the training sample from the prediction direction of each unit task model, thereby obtaining a combination of candidate unit task models to perform the training task.

[0099] Optionally, based on the task results of each training task and the sample labels of the training samples, the training loss of the candidate region supply object screening model is obtained, and the model parameters of the candidate region supply object screening model are adjusted according to the training loss. Then, the model is returned to obtain the next training sample and the candidate region supply object screening model with adjusted parameters is trained again until the training ends, and the trained target region supply object screening model is obtained.

[0100] In this embodiment of the disclosure, the task results of each training task and the sample labels of the training samples can be processed by the loss acquisition algorithm in the related technology to obtain the training loss of the candidate region supply object screening model.

[0101] In this scenario, the model parameters of the candidate region supply object screening model can be adjusted based on the training loss. The parameters that are adjusted during model iteration can be marked as the model iteration parameters of the candidate region supply object screening model.

[0102] Optionally, the model iteration parameters of the candidate region supply object selection model can be adjusted according to the training loss to obtain the parameter-adjusted candidate region supply object selection model. The model iteration parameters include at least the parameters of each candidate task model in the candidate task model combination and the parameters of the candidate gating units.

[0103] In this embodiment of the disclosure, the model iteration parameters participating in the iteration of the candidate region supply object screening model may include the parameters of each candidate task model in the candidate task model combination selected by the gating unit for the training task, as well as the parameters of the gating unit, and may also include other model iteration parameters of the candidate region supply object screening model, which are not specifically limited here.

[0104] In this scenario, the model iteration parameters can be adjusted based on the training loss to obtain a candidate region supply object selection model with adjusted parameters. Then, the next training sample is obtained to continue training the candidate region supply object selection model with adjusted parameters until the training ends, resulting in a trained target region supply object selection model.

[0105] In this embodiment of the disclosure, after completing the model training of the current round and obtaining the candidate region supply object screening model with adjusted parameters, it can return to obtain the next training sample, continue to train the candidate region supply object screening model with adjusted parameters, until the training is completed, and obtain the trained target region supply object screening model.

[0106] Optionally, the training termination condition of the candidate region supply object selection model can be set according to the training round. For the model training of the current round, if the training round meets the preset model training termination condition, the model training of the candidate region supply object selection model can be terminated, and the candidate region supply object selection model obtained in the current round of training can be determined as the trained target region supply object selection model.

[0107] Optionally, the training termination condition of the candidate region supply object screening model can be set according to the training output results. For the current round of model training, if the output result of the current round of model training meets the preset model training termination condition, the model training of the candidate region supply object screening model can be terminated, and the candidate region supply object screening model obtained in the current round of training can be determined as the trained target region supply object screening model.

[0108] The regional supply target selection method proposed in this disclosure obtains a candidate regional supply target selection model to be trained, as well as training samples for the candidate regional supply target selection model. The model is then trained using these training samples until training is complete, resulting in a trained target regional supply target selection model. This disclosure trains the candidate regional supply target selection model through multiple training tasks, enabling the trained target regional supply target selection model to perform multi-task predictions. This improves the prediction allocation rate of candidate supply targets and the efficiency and accuracy of predicting supply volume, thereby enhancing the selection efficiency of target supply targets in the target supply region and optimizing the product selection methods and effects for offline stores in the target supply region.

[0109] In the above embodiments, regarding obtaining target supply objects for the target supply region through a trained target region supply object screening model, it can be combined with... Figure 5 To understand further, Figure 5 This is a flowchart illustrating a regional supply target screening method according to another embodiment of the present disclosure, as shown below. Figure 5 As shown, the method includes:

[0110] S501, obtain candidate supply objects in the target supply area, and based on the target area supply object screening model, obtain the predicted allocation rate and predicted supply volume of candidate supply objects in the target supply area within the prediction time range.

[0111] Optionally, the target gating unit in the target area supply object screening model can be obtained.

[0112] Among them, the gating units in the trained target region supply object screening model can be marked as target gating units in the target region supply object screening model.

[0113] As an example, such as Figure 4 As shown, settings Figure 4 The example shown is a target region provided with an object selection model. Figure 4 The gated units A and B shown are the target gated units in the target area supply object screening model.

[0114] Optionally, the target gating unit can obtain the allocation rate prediction task model combination of the allocation rate prediction task corresponding to the predicted allocation rate and the supply quantity prediction task combination of the predicted supply quantity from the target unit task model set.

[0115] Among them, the set of unit task models in the trained target region supply object selection model can be marked as the target unit task model set of the target region supply object selection model.

[0116] As an example, such as Figure 4 As shown, settings Figure 4 The example shown is a target region provided with an object selection model. Figure 4 The set of unit task models shown is the set of target unit task models in the target area supply object screening model.

[0117] In this embodiment of the disclosure, the predicted allocation rate and the predicted supply quantity can be obtained through a trained target area supply object screening model. The model task of obtaining the predicted allocation rate can be marked as the allocation rate prediction task of the target area supply object screening model, and the model task of obtaining the predicted supply quantity can be marked as the supply quantity prediction task of the target area supply object screening model.

[0118] As an example, such as Figure 4 As shown, in Figure 4 In the scenario shown, where a pre-trained target region is provided with an object filtering model, it can be achieved through... Figure 4 The target gating unit A shown retrieves at least one target unit task model from the target unit task model set of the target area supply object screening model to perform the allocation rate prediction task, and marks the combination of the at least one target unit task model as the allocation rate prediction task model combination for performing the allocation rate prediction task. In this example, it can be... Figure 4 The task model combination A shown is considered as the allocation rate prediction task model combination.

[0119] As another example, still as Figure 4 As shown, in Figure 4 In the scenario shown, where a pre-trained target region is provided with an object filtering model, it can be achieved through... Figure 4 The target gating unit B shown retrieves at least one target unit task model from the target unit task model set of the target area supply object screening model to perform the supply quantity forecasting task, and marks the combination of the at least one target unit task model as the supply quantity forecasting task model combination for performing the supply quantity forecasting task. In this example, it can be... Figure 4 The task model combination B shown is considered as the supply forecast task model combination.

[0120] Optionally, the target area feature set and the candidate supply object feature set of the target supply area are obtained, and the model inputs of the allocation rate prediction task model combination and the supply quantity prediction task model combination are obtained according to the target area feature set and the candidate supply object feature set.

[0121] Specifically, the target region features and target region allocation features of the target supply region in the target region feature set, as well as the candidate supply object features and historical supply volume features of the candidate supply object in the candidate supply object feature set, can be obtained. The target region features, target region allocation features, candidate supply object features, and historical supply volume features are then concatenated to obtain the model inputs for the allocation rate prediction task model combination and the supply volume prediction task model combination, respectively.

[0122] In this embodiment of the disclosure, the acquisition of the target area feature set including target area features and target area allocation features, as well as the candidate supply object feature set including candidate supply object features and historical supply volume features, can be understood in conjunction with the acquisition process of the sample area feature set including sample area features and sample area allocation features, and the sample supply object feature set including sample supply object features and sample area supply features proposed in the above embodiments, and will not be repeated here.

[0123] Furthermore, the splicing of target area characteristics, target area allocation characteristics, candidate supply object characteristics, and historical supply volume characteristics can be understood in conjunction with the splicing process of sample area characteristics, sample area allocation characteristics, sample supply object characteristics, and sample area supply characteristics proposed in the above embodiments, and will not be elaborated here.

[0124] Furthermore, the features obtained by splicing together the target area features, target area allocation features, candidate supply object features, and historical supply volume features are determined as the model inputs for the allocation rate prediction task model combination and the supply volume prediction task model combination, respectively.

[0125] Optionally, based on the model input, the predicted allocation rate of the candidate supply object in the target supply area within the prediction time range output by the allocation rate prediction task model combination, and the predicted supply amount of the candidate supply object in the target supply area within the prediction time range output by the supply amount prediction task model combination, are obtained.

[0126] As an example, such as Figure 4 As shown, task model combination A, which is a combination of task models for predicting allocation rates, can output its predicted allocation rate based on the model input and output. Here, it can be... Figure 4 The output A shown is considered as the predicted allocation rate of the combined output of the allocation rate prediction task model.

[0127] Accordingly, such as Figure 4 As shown, task model combination B, which is a combination of supply forecasting task models, can output its predicted supply based on the model input and output. Here, it can be... Figure 4 The output B shown is considered as the predicted supply output of the combined output of the supply forecasting task model.

[0128] S502, based on the predicted allocation rate and predicted supply volume, select the target supply objects for the target supply area from the candidate supply objects.

[0129] Optionally, the supply value of the candidate supply object in the target supply area is obtained, and based on the supply value and the predicted supply quantity, the total predicted supply value of the candidate supply object in the target supply area within the prediction time range is obtained.

[0130] As an example, the target supply area is set as offline stores, the candidate supply objects are the products to be screened in offline stores, the supply value of the candidate supply objects in the target supply area is the selling price of the products to be screened in offline stores, and the predicted supply volume is the predicted sales volume of the products to be screened in offline stores within the predicted time range.

[0131] In this example, the predicted Gross Merchandise Volume (GMV) of the products to be screened for in-store sales can be obtained by multiplying the unit selling price of the products to be screened for in-store sales in offline stores and the predicted sales volume of the products to be screened for in-store sales in offline stores within the predicted time range.

[0132] Specifically, the total predicted sales amount can be denoted as the sum of the predicted supply value of candidate suppliers within the target supply area during the prediction period.

[0133] Optionally, based on the predicted allocation rate, predicted supply quantity, and predicted supply value, candidate supply scores are obtained for candidate supply objects within the predicted time period, and target supply objects within the target supply area within the predicted time period are selected from the candidate supply objects based on the candidate supply scores.

[0134] In this embodiment of the disclosure, the predicted allocation rate, the predicted supply quantity, and the total predicted supply value each have their own weighting coefficients. In this scenario, the predicted allocation rate, the predicted supply quantity, and the total predicted supply value can be weighted and fused based on the weighting system, and the result of the weighted fusion can be determined as the candidate supply score of the candidate supply object in the target supply area within the prediction time range.

[0135] Furthermore, all candidate supply objects can be sorted according to the candidate supply score, and a preset number of candidate supply objects can be selected from high to low as the target supply objects for the target supply area.

[0136] The regional supply target screening method proposed in this disclosure obtains the predicted allocation rate and predicted supply quantity of candidate supply targets through a trained target regional supply target screening model. Based on the supply value and predicted supply quantity of candidate supply targets, the total predicted supply value of candidate supply targets in the target supply region within the predicted time range is obtained. Furthermore, candidate supply scores of candidate supply targets are obtained based on the predicted allocation rate, predicted supply quantity, and total predicted supply value. Finally, target supply targets for the target supply region are screened from all candidate supply targets based on the candidate supply scores. In this disclosure, target suppliers for the target supply region are selected from candidate suppliers based on predicted allocation rates and predicted supply volumes. This improves the fit between target suppliers and target supply regions, avoids wasting product selection quotas in the target supply region, and improves the selection efficiency of target suppliers in the target supply region by obtaining predicted allocation rates and predicted supply volumes of candidate suppliers based on the target region supplier selection model. This avoids the situation where the target supply region staff cannot visit all candidate suppliers, resulting in a small number of target supplier categories in the target supply region. This increases the possibility of the target supply region achieving the headquarters' category operation goals, optimizes the product selection method and effect of offline stores in the target supply region, and thus optimizes the product sales of offline stores in the target supply region.

[0137] Corresponding to the regional supply target screening methods proposed in the above embodiments, an embodiment of this disclosure also proposes a regional supply target screening device. Since the regional supply target screening device proposed in this disclosure corresponds to the regional supply target screening methods proposed in the above embodiments, the implementation methods of the above regional supply target screening methods are also applicable to the regional supply target screening device proposed in this disclosure, and will not be described in detail in the following embodiments.

[0138] Figure 6 This is a schematic diagram of the structure of a regional supply object screening device according to an embodiment of the present disclosure, as shown below. Figure 6 As shown, the regional supply target screening device 600 includes an acquisition module 61, a prediction module 62, and a screening module 63, wherein:

[0139] Module 61 is used to acquire a trained target area supply object screening model for the target supply area;

[0140] The prediction module 62 is used to obtain candidate supply objects in the target supply area and, based on the target area supply object screening model, obtain the predicted allocation rate and predicted supply volume of candidate supply objects in the target supply area within the prediction time range.

[0141] The screening module 63 is used to screen the target supply objects for the target supply area from the candidate supply objects based on the predicted allocation rate and the predicted supply volume.

[0142] In this embodiment of the disclosure, the acquisition module 61 is further configured to: acquire a candidate region supply object screening model to be trained; acquire a sample region feature set of the sample region and a sample supply object feature set of the sample region, wherein the sample region feature set includes at least one of the sample region features and the sample region allocation features, and the sample supply object feature set includes at least one of the sample supply object features and the sample region supply features; obtain training samples for the candidate region supply object screening model based on the sample region feature set and the sample supply object feature set; input the training samples into the candidate region supply object screening model for model training until the training is completed, thereby obtaining the trained target region supply object screening model for the target supply region.

[0143] In this embodiment of the disclosure, the acquisition module 61 is further configured to: concatenate the sample region features and sample region allocation features in the sample region feature set, and the sample supply object feature set including sample supply object features and sample region supply features, to obtain the concatenated sample concatenated features; acquire the historical region allocation rate and historical region supply amount of the sample supply object in the sample region within a historical time range; and use the historical region allocation rate and historical region supply amount as sample labels for the sample concatenated features to obtain training samples for the candidate region supply object screening model.

[0144] In this embodiment of the disclosure, the acquisition module 61 is further configured to: acquire a set of training tasks for the candidate region supply object screening model based on training samples; acquire candidate task model combinations for each training task in the training task set through the candidate gating unit in the candidate region supply object screening model; input training samples into each candidate task model combination to obtain the task results of each training task output by each candidate task model combination; acquire the training loss of the candidate region supply object screening model based on the task results of each training task and the sample labels of the training samples; adjust the model parameters of the candidate region supply object screening model based on the training loss, and return to acquire the next training sample to continue training the candidate region supply object screening model after parameter adjustment until the training ends, thereby obtaining the trained target region supply object screening model.

[0145] In this embodiment of the disclosure, the acquisition module 61 is further configured to: acquire a set of unit task models for the candidate region supply object screening model; for any training task, acquire at least one candidate unit task model for executing the training task from the set of unit task models through the candidate gating unit, and determine the combination of at least one candidate unit task model as the candidate task model combination for the training task.

[0146] In this embodiment of the disclosure, the acquisition module 61 is further configured to: adjust the model iteration parameters of the candidate region supply object screening model according to the training loss to obtain the parameter-adjusted candidate region supply object screening model, wherein the model iteration parameters include at least the parameters of each candidate task model in the candidate task model combination and the parameters of the candidate gating unit; return to acquire the next training sample to continue model training of the parameter-adjusted candidate region supply object screening model until the training ends, and obtain the trained target region supply object screening model.

[0147] In this embodiment of the disclosure, the prediction module 62 is further configured to: obtain target gating units in the target area supply object screening model; obtain, through the target gating units, a combination of allocation rate prediction task models corresponding to the allocation rate prediction task and a combination of supply quantity prediction task models corresponding to the supply quantity prediction task from the target unit task model set; obtain a target area feature set of the target supply area and a candidate supply object feature set of the candidate supply objects, and obtain the model inputs of the allocation rate prediction task model combination and the supply quantity prediction task model combination according to the target area feature set and the candidate supply object feature set; and obtain, according to the model inputs, the predicted allocation rate of the candidate supply objects in the target supply area within the prediction time range output by the allocation rate prediction task model combination and the predicted supply quantity of the candidate supply objects in the target supply area within the prediction time range output by the supply quantity prediction task model combination.

[0148] In this embodiment of the disclosure, the prediction module 62 is further configured to: obtain the target area features and target area allocation features of the target supply area in the target area feature set, and the candidate supply object features and historical supply volume features of the candidate supply object in the candidate supply object feature set; and concatenate the target area features, target area allocation features, candidate supply object features and historical supply volume features to obtain the model inputs of the allocation rate prediction task model combination and the supply volume prediction task model combination, respectively.

[0149] In this embodiment of the disclosure, the screening module 63 is further configured to: obtain the supply value of the candidate supply object in the target supply area, and obtain the total predicted supply value of the candidate supply object in the target supply area within the prediction time range based on the supply value and the predicted supply quantity; obtain the candidate supply score of the candidate supply object within the prediction time range based on the predicted allocation rate, the predicted supply quantity and the total predicted supply value, and screen the target supply object in the target supply area within the prediction time range from the candidate supply objects based on the candidate supply score.

[0150] The regional supplier selection device proposed in this disclosure acquires a trained target region supplier selection model for the target supply region, and obtains the predicted allocation rate and predicted supply volume of candidate suppliers through the target region supplier selection model. Then, based on the predicted allocation rate and predicted supply volume, the target suppliers for the target supply region are selected from the candidate suppliers. In this disclosure, selecting target suppliers for the target supply region from candidate suppliers based on the predicted allocation rate and predicted supply volume improves the fit between target suppliers and target supply regions, avoids wasting the product selection quota of the target supply region, and improves the selection efficiency of target suppliers based on the target region supplier selection model. This avoids the situation where the target supply region has a limited variety of target suppliers due to the inability of staff to traverse all candidate suppliers, thus increasing the likelihood that the target supply region will achieve the headquarters' category operation goals. It also optimizes the product selection method and effect for offline stores in the target supply region, thereby optimizing the product sales of offline stores in the target supply region.

[0151] To achieve the above embodiments, this disclosure also provides an electronic device, a computer-readable storage medium, and a computer program product.

[0152] Figure 7 This is a block diagram of an electronic device 700 according to an embodiment of the present disclosure, as follows: Figure 7 As shown, the electronic device 700 includes a memory 71, a processor 72, and a computer program stored on the memory 71 and executable on the processor 72. When the processor 72 executes program instructions, it implements the regional supply object filtering method for servers provided in the above embodiments.

[0153] The regional supplier selection method proposed in this disclosure obtains a trained target region supplier selection model for the target supply region, and uses this model to obtain the predicted allocation rate and predicted supply volume of candidate suppliers. Then, based on the predicted allocation rate and predicted supply volume, the target suppliers for the target supply region are selected from the candidate suppliers. This disclosure improves the fit between target suppliers and target supply regions by selecting target suppliers from candidate suppliers based on the predicted allocation rate and predicted supply volume, avoiding waste of product selection quotas in the target supply region. Obtaining the predicted allocation rate and predicted supply volume of candidate suppliers based on the target region supplier selection model improves the selection efficiency of target suppliers in the target supply region, avoiding the situation where the target supply region staff cannot traverse all candidate suppliers, resulting in a limited variety of target suppliers in the target supply region. This increases the likelihood that the target supply region will achieve the headquarters' category operation goals, optimizes the product selection method and effect for offline stores in the target supply region, and ultimately optimizes the product sales of offline stores in the target supply region.

[0154] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0155] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0156] The program code used to implement the method itself can be written in any combination of one or more programming languages. This program code can be provided to the processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0157] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0158] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0159] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or grid browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication grid). Examples of communication grids include local area networks (LANs), wide area networks (WANs), the Internet, and blockchain grids.

[0160] Computer systems can include clients and servers. Clients and servers are generally geographically separated and typically interact through a communication mesh. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, a hosting product within the cloud computing service system, addressing the shortcomings of traditional physical hosts and VPS (Virtual Private Server, or simply "VPS") services, such as high management difficulty and weak business scalability. Servers can also be servers in distributed systems or servers integrated with blockchain technology.

[0161] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0162] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this disclosure, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0163] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of this disclosure includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of this disclosure pertain.

[0164] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0165] It should be understood that various parts of this disclosure can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0166] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware, and the program can be stored in a computer-readable storage medium. When executed, the program includes one or a combination of the steps of the method embodiments.

[0167] Furthermore, the functional units in the various embodiments of this disclosure can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0168] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of the present disclosure have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present disclosure. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present disclosure.

[0169] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0170] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for selecting regional supply targets, characterized in that, The method includes: Obtain a trained target supply object screening model for the target supply area; The candidate supply objects of the target supply area are obtained, and the combination of allocation rate prediction task model and supply volume prediction task model selected by the target area supply object screening model are used to obtain the predicted allocation rate and predicted supply volume of the candidate supply objects in the target supply area within the prediction time range by means of the target area characteristics, target area allocation characteristics, candidate supply object characteristics of the candidate supply objects and historical supply volume characteristics of the target supply area. Based on the predicted allocation rate and the predicted supply volume, the target supply objects for the target supply area are selected from the candidate supply objects.

2. The method according to claim 1, characterized in that, The trained target region supply object screening model for obtaining the target supply region includes: Obtain the candidate region supply object screening model to be trained; Obtain the sample region feature set of the sample region and the sample supply object feature set of the sample supply object of the sample region, wherein the sample region feature set includes the sample region features and sample region allocation features of the sample region, and the sample supply object feature set includes the sample supply object features and sample region supply features of the sample supply object; The training samples for the candidate region supply object screening model are obtained based on the sample region feature set and the sample supply object feature set. The training samples are input into the candidate region supply object screening model for model training until the training is completed, resulting in the trained target region supply object screening model for the target supply region.

3. The method according to claim 2, characterized in that, The step of obtaining training samples for the candidate region supply object screening model based on the sample region feature set and the sample supply object feature set includes: The sample region features and the sample region allocation features in the sample region feature set, and the sample supply object feature set including the sample supply object features and the sample region supply features are spliced ​​together to obtain the spliced ​​sample spliced ​​features. Within a historical time frame, the historical regional allocation rate and historical regional supply volume of the sample supply object within the sample area are obtained. The historical regional allocation rate and the historical regional supply volume are used as sample labels for the sample splicing features to obtain the training samples for the candidate regional supply object screening model.

4. The method according to claim 2, characterized in that, The step of inputting the training samples into the candidate region supply target selection model for model training, until the training is completed, to obtain the trained target region supply target selection model, includes: Obtain the set of training tasks for the candidate region supply object screening model based on the training samples; By using the candidate gating unit in the candidate region supply object screening model, the candidate task model combination of each training task in the training task set is obtained; The training samples are input into each candidate task model combination to obtain the task results of each training task output by each candidate task model combination. Based on the task results of each training task and the sample labels of the training samples, the training loss of the candidate region supply object screening model is obtained. The model parameters of the candidate region supply object screening model are adjusted according to the training loss, and the next training sample is obtained to continue training the candidate region supply object screening model with adjusted parameters until the training ends, so as to obtain the trained target region supply object screening model.

5. The method according to claim 4, characterized in that, The step of obtaining candidate task model combinations for each training task in the training task set by using the candidate gating unit in the candidate region supply object screening model includes: Obtain the set of unit task models for the candidate region supply object screening model; For any training task, at least one candidate unit task model for executing the training task is obtained from the set of unit task models through the candidate gating unit, and the combination of the at least one candidate unit task model is determined as the candidate task model combination for the training task.

6. The method according to claim 4, characterized in that, The step of adjusting the model parameters of the candidate region supply object selection model based on the training loss, and then returning to obtain the next training sample to continue training the parameter-adjusted candidate region supply object selection model until training is completed, thereby obtaining the trained target region supply object selection model, includes: Based on the training loss, the model iteration parameters of the candidate region supply object screening model are adjusted to obtain the parameter-adjusted candidate region supply object screening model. The model iteration parameters include at least the parameters of each candidate task model in the candidate task model combination and the parameters of the candidate gating unit. Return to obtain the next training sample and continue training the candidate region supply object screening model with adjusted parameters until the training is completed, and obtain the trained target region supply object screening model.

7. The method according to claim 1, characterized in that, The process of acquiring candidate supply objects for the target supply region, and based on the combination of allocation rate prediction task model and supply quantity prediction task model selected by the target region supply object screening model, obtains the predicted allocation rate and predicted supply quantity of the candidate supply objects within the target supply region within the prediction time range through the target region characteristics, target region allocation characteristics, candidate supply object characteristics, and historical supply quantity characteristics of the candidate supply objects. Obtain the target gating unit in the target area supply object screening model; The target gating unit obtains the allocation rate prediction task model combination of the allocation rate prediction task corresponding to the predicted allocation rate and the supply quantity prediction task model combination of the supply quantity prediction task corresponding to the predicted supply quantity from the target unit task model set. The target region features, the target region allocation features, the candidate supply object features, and the historical supply volume features of the candidate supply object are obtained to obtain the model inputs of the allocation rate prediction task model combination and the supply volume prediction task model combination, respectively. Based on the model input, the predicted allocation rate of the candidate supply object within the target supply area within the predicted time range, output by the combined model of the allocation rate prediction task, and the predicted supply amount of the candidate supply object within the target supply area within the predicted time range, output by the combined model of the supply amount prediction task, are obtained.

8. The method according to claim 7, characterized in that, The step of obtaining the target region characteristics, the target region allocation characteristics, and the candidate supply object characteristics and historical supply volume characteristics of the candidate supply object to obtain the model inputs for the allocation rate prediction task model combination and the supply volume prediction task model combination includes: The target area features, target area allocation features, candidate supply object features, and historical supply volume features are concatenated to obtain the model inputs for the allocation rate prediction task model combination and the supply volume prediction task model combination, respectively.

9. The method according to claim 1, characterized in that, The step of selecting target supply objects for the target supply region from the candidate supply objects based on the predicted allocation rate and the predicted supply volume includes: Obtain the supply value of the candidate supply object in the target supply area, and based on the supply value and the predicted supply quantity, obtain the total predicted supply value of the candidate supply object in the target supply area within the predicted time range; Based on the predicted allocation rate, the predicted supply quantity, and the sum of the predicted supply value, a candidate supply score is obtained for the candidate supply objects within the predicted time range. Based on the candidate supply score, the target supply objects within the target supply area within the predicted time range are selected from the candidate supply objects.

10. A regional supply target screening device, characterized in that, The device includes: The acquisition module is used to acquire the trained target area supply object screening model for the target supply area; The prediction module is used to obtain candidate supply objects in the target supply area, and based on the combination of allocation rate prediction task model and supply quantity prediction task model selected by the target area supply object screening model, through the target area characteristics, target area allocation characteristics, candidate supply object characteristics and historical supply quantity characteristics of the candidate supply objects in the target supply area within the prediction time range, the predicted allocation rate and predicted supply quantity of the candidate supply objects in the target supply area within the prediction time range. The filtering module is used to filter the target supply objects for the target supply area from the candidate supply objects based on the predicted allocation rate and the predicted supply volume.

11. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute instructions to implement the method as claimed in any one of claims 1-9.

12. A computer-readable storage medium, wherein when instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform the method of any one of claims 1-9.

Images