Methods, equipment, storage media, and software products for determining resources for item exchange.

Abstract

This application provides a method, device, storage medium, and program product for determining the exchange resources for goods. It acquires attribute information of the item to be exchanged, and based on this, retrieves target similar items from an item pool that meet similarity requirements, accurately obtaining target similar items that are meaningful for determining the exchange resource value of the item to be exchanged. Based on the exchange resource value of the target similar items, a preliminary range of exchange resource values ​​for the item to be exchanged is defined to avoid deviations caused by fluctuations in a single reference value. Then, considering the actual exchange situation of the target similar items within a preset time period, the target exchange resource value for the item to be exchanged is selected from the exchange resource range. By introducing similar items and historical exchange data, the determined exchange resource value is more in line with market conditions, significantly improving the market adaptability of the exchange resource value and promoting the circulation efficiency of goods. The automated process reduces the subjectivity and cost of manual intervention, improving the efficiency of item exchange.

Description

Technical Field

[0001] This application relates to the field of goods exchange technology, and in particular to a method, device, storage medium and program product for determining goods exchange resources. Background Technology

[0002] In the context of goods exchange, the value of goods is usually expressed in the form of exchangeable resources. Users can publish goods through goods exchange systems (such as online exchange systems that offer second-hand goods in a consumer-to-consumer (C2C) model) to make them visible to more people and facilitate exchanges.

[0003] When listing items in an item exchange system, the user providing the item usually needs to set the exchange resource value. After the item is listed, the exchange resource value often needs to be adjusted to expedite the exchange process.

[0004] However, the users who provide the items are usually individuals, making it inefficient to determine the exchange value of resources. Summary of the Invention

[0005] This application provides a method, device, storage medium, and program product for determining the value of exchangeable resources for goods, thereby improving the efficiency of determining the value of exchangeable resources for goods.

[0006] In a first aspect, embodiments of this application provide a method for determining resources for exchanging goods. The method includes: Obtain the attribute information of the item to be redeemed; Based on the attribute information of the item to be redeemed, a target item of the same type that meets the similarity requirement to the item to be redeemed is obtained from the item pool; Based on the exchange resource value of the target item, determine the exchange resource range value of the item to be exchanged; Based on the exchange resource value of the target item and the exchange status within a preset time period, the target exchange resource value of the item to be exchanged is determined within the exchange resource range.

[0007] Secondly, embodiments of this application provide a device for determining resource exchange items, the device comprising: The acquisition module is used to obtain the attribute information of the items to be redeemed; The same item identification module is used to obtain target same items that meet the similarity requirements of the item to be redeemed from the item pool based on the attribute information of the item to be redeemed. The resource redemption determination module is used to determine the range of redemption resources for the item to be redeemed based on the redemption resource value of the target item; and to determine the target redemption resource value for the item to be redeemed within the range of redemption resources based on the redemption resource value of the target item and the redemption status within a preset time period.

[0008] Thirdly, embodiments of this application provide an electronic device, including: a memory, a processor, and a communication interface; wherein, the memory stores executable code, and when the executable code is executed by the processor, the processor performs the method described in the first aspect.

[0009] Fourthly, embodiments of this application provide a non-transitory machine-readable storage medium storing executable code, which, when executed by a processor of an electronic device, enables the processor to at least implement the method described in the first aspect.

[0010] Fifthly, embodiments of this application provide a computer program product, the computer program product including a computer program, which, when executed by a processor, can implement the method described in the first aspect.

[0011] The scheme for determining the exchange resources provided in this application embodiment obtains the attribute information of the item to be exchanged, and accordingly retrieves target similar items from the item pool that meet the similarity requirements of the item to be exchanged, accurately obtaining target similar items that are of reference significance for determining the exchange resource value of the item to be exchanged. Then, based on the exchange resource value of the target similar items, a preliminary range of exchange resource values ​​for the item to be exchanged is defined to avoid deviations caused by fluctuations in a single reference value. Furthermore, considering the actual exchange situation of the target similar items within a preset time period, the target exchange resource value for the item to be exchanged is selected from the exchange resource range. For items to be exchanged with single inventory and unstructured attribute information, by introducing similar items and historical exchange data, exchange resource values ​​that are more in line with market conditions can be obtained, significantly improving the market adaptability of exchange resource values ​​and promoting the circulation efficiency of goods. Automated determination of exchange resource values ​​reduces the subjectivity and cost of manual intervention, effectively improving the efficiency of item exchange. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1A schematic diagram of the hardware execution environment for a model optimization method provided in an embodiment of this application; Figure 2 This is an application diagram of a cloud computing environment provided in an embodiment of this application; Figure 3 A flowchart illustrating a method for determining resource exchange for goods, provided in an embodiment of this application; Figure 4 This application provides a schematic diagram illustrating the principle of determining the target exchange resource value in an embodiment of the present application. Figure 5 A flowchart illustrating another method for determining resource exchange for goods, provided in an embodiment of this application; Figure 6 A flowchart illustrating a training method for an item selection model provided in an embodiment of this application; Figure 7 A schematic diagram of the structure of a device for determining the exchange of goods resources provided in an embodiment of this application; Figure 8 This is a schematic diagram of the structure of an electronic device provided in this embodiment. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. In addition, the timing of the steps in the following method embodiments is only an example and not a strict limitation.

[0015] It should be noted that, in the cases involving user information in the embodiments of this application, the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse. In addition, the various models involved in this application (including but not limited to large language models or other models) comply with relevant laws and standards.

[0016] First, the terms or concepts involved in the embodiments of this application will be explained: Exchangeable resources are used to represent the value of items and can be general equivalents, such as currency or virtual currency. The exchangeable resource value represents the quantity of resources exchanged.

[0017] A Standard Product Unit (SPU) refers to a collection of items of the same type, model, and specifications, without distinction of quality, exchange value, or individual items.

[0018] Standardized products refer to items with unified market standards, clearly defined specifications and models, and a high degree of homogeneity among products, for which consumers have a fixed understanding of their function and appearance. Examples include electronic products, home appliances, and daily necessities (such as a brand's 550ml bottled water or a brand's size of facial tissue). Conversely, non-standardized products refer to items without unified standards and specifications. Examples include handicrafts, clothing, footwear, and agricultural products.

[0019] In item exchange scenarios, users need to set the exchange value when posting items. If the exchange value is set higher than the market price, the item may take a long time to be exchanged. If the exchange value is set lower than the market price, although the exchange can usually be completed quickly, it does not meet the user's expectations. However, after an item is posted, the market price is constantly changing, and the users who provide the item are usually individuals who are unlikely to perceive these changes, causing the set exchange value to be out of touch with the current market conditions. Therefore, there is an urgent need for a method to determine an item exchange value that reflects the current market conditions.

[0020] Currently, for various online redemption systems operating under the Business to Consumer (B2C) model, the descriptions of items are typically structured data, and items are often in multiple inventory formats. Based on structured historical data, the redemption resource value of items can be determined through methods such as data modeling.

[0021] However, in goods exchange scenarios, such as C2C online exchange systems that offer secondhand goods, the items available for exchange are mostly single-inventory items. The descriptions provided by individuals offering the items are unstructured, sometimes even lacking crucial information like condition and version, resulting in a lack of directly referable historical data. Therefore, it's impossible to determine the exchange value of items using the aforementioned data modeling methods.

[0022] Based on the above, this application proposes a method for determining the exchange resources for goods. The method involves obtaining the attribute information of the item to be exchanged, and then matching it with target similar items from an item pool that meet the similarity requirements. This accurately yields target similar items that are meaningful for determining the exchange resource value of the item to be exchanged. Next, based on the exchange resource value of the target similar items, a preliminary range of exchange resource values ​​for the item to be exchanged is defined to avoid deviations caused by fluctuations in a single reference value. Then, considering the actual exchange situation of the target similar items within a preset time period, the target exchange resource value for the item to be exchanged is selected from the range of exchange resource values. For items with single inventory and unstructured attribute information, by introducing similar items and historical exchange data, exchange resource values ​​that better reflect market conditions can be obtained, significantly improving the market adaptability of exchange resource values ​​and promoting the circulation efficiency of goods. Automated determination of exchange resource values ​​reduces the subjectivity and cost of manual intervention, effectively improving the efficiency of item exchange.

[0023] The following describes the scheme for determining the exchange resources for goods provided in the embodiments of this application.

[0024] The method for determining the resources for item exchange provided in this application embodiment can be applied to an item exchange system. Figure 1 This is a schematic diagram of an item exchange system provided in an embodiment of this application, such as... Figure 1 As shown, the item redemption system may include client devices and server devices 101, with the client devices communicating with the server devices 101. The server devices 101 may be cloud servers from cloud service providers. The client devices 102 may be laptops, tablets, PCs, robots, etc. Figure 1 The example shows three client devices: a first client device 102, a second client device 103, and a third client device 104. Figure 1 This is just one example and does not constitute a limit on the number of client devices.

[0025] The following example illustrates the item posting process: The user providing the item logs into the item exchange platform via the first client device 102 and enters the attribute information of the item to be exchanged. The attribute information describes the multi-dimensional attributes of the item. Optionally, the attribute information may include an image and / or a description of the item. For example, the description information might include the brand, model, specifications, and condition of the item. The first client device 102 then sends the attribute information of the item to the server device 101. The server device 101 then posts the item on the item exchange platform based on the attribute information.

[0026] In one possible implementation, the first client device 102 sends the attribute information of the item to be redeemed to the server device 101, along with the set redemption resource value for the item. The server device 101 then publishes the item based on its attribute information and the redemption resource value.

[0027] In another possible implementation, the attribute information of the item to be redeemed includes its image and description. Server device 101 retrieves a target similar item from the item pool that meets the similarity requirement based on the attribute information of the item to be redeemed. Based on the redemption resource value of the target similar item, it determines the redemption resource range for the item to be redeemed. Based on the redemption resource value of the target similar item and the redemption status within a preset time period, it determines the target redemption resource value for the item to be redeemed within the redemption resource range. Then, server device 101 can feed back the target redemption resource value to the first client device 102. The first client device 102 can display the target redemption resource value to provide a reference for the user to set the redemption resource value for the item to be redeemed. The first client device 102 receives the redemption resource value input by the user and sends it to server device 101. Server device 101 publishes the item based on its attribute information and the redemption resource value.

[0028] After an item is published, other users can view it through a client device. This client device can be the same as or different from the first client device 102, as long as the user logging into the client device is different from the user logging into the first client device 102.

[0029] Furthermore, after an item is released, due to changes in market conditions, the server device 101 retrieves a target similar item from the item pool that meets the similarity requirement to the item to be redeemed, based on the attribute information of the item to be redeemed. Based on the redemption resource value of the target similar item, the server device 101 determines the redemption resource range for the item to be redeemed. Based on the redemption resource value of the target similar item and the redemption situation within a preset time period, the server device 101 determines the target redemption resource value for the item to be redeemed within this range. Subsequently, in one possible implementation, the server device 101 can push this target redemption resource value to the first client device 102 to provide a reference for the user to adjust the redemption resource value of the item to be redeemed. In another possible implementation, the server device 101 can also adjust the redemption resource value of the item to be redeemed based on the target redemption resource value, with the authorization of the user who provided the item. Here, user authorization refers to the server device 101 receiving authorization information instructing it to adjust the redemption resource value. For example, the user signs a document agreeing to the server device 101 adjusting the redemption resource value of the item to be redeemed. For example, if a user sets a lower limit for the exchange of resources, the server device 101 can directly adjust the exchange resource value of the item to be exchanged if it is above this lower limit.

[0030] In some embodiments, during the process of determining the target exchange resource value, the server device 101 may invoke at least one model or intelligent agent to obtain the target item and / or the range of exchange resources. This at least one model or intelligent agent may be deployed in the server device 101 or in other devices communicatively connected to the server device 101 for invocation by the server device 101. Optionally, the models involved in this application are models trained based on deep learning and artificial intelligence technologies, such as AI-based language models (LM), image recognition models, or multimodal models (MM), possessing significant generalization and intelligence capabilities, thereby enabling various task processing capabilities, such as language understanding, generation, and image recognition. Furthermore, the models involved in the embodiments of this application may be separately deployed models or models within intelligent agents.

[0031] In practical applications, the aforementioned server-side device 101 can be a cloud server maintained by a cloud service provider—referred to as a computing node. In cases such as... Figure 2 The cloud computing environment shown may include several distributed deployments. Figure 2 The diagram illustrates compute nodes (201-1, 201-2, ...), each possessing processing resources such as computing and storage. In a cloud computing environment, multiple compute nodes can be organized to provide a specific service; conversely, a single compute node can provide one or more services. Figure 2The diagram illustrates services A, B, C, and D. In a cloud computing environment, these services can be provided through external service interfaces, which client devices then call to use the corresponding services. Service interfaces can take the form of Software Development Kits (SDKs) or Application Programming Interfaces (APIs).

[0032] The services described above are deployed using various virtualization technologies supported by cloud computing environments, such as virtual machine-based and container-based virtualization technologies. Taking container-based virtualization technology as an example, several containers corresponding to a service can be assembled into a container group (pod). For example... Figure 2 The illustrated service B can be configured with one or more pods, and each pod can include a proxy and one or more containers. The one or more containers in the pod are used to handle requests related to one or more corresponding functions of the service, and the proxy in the pod is used to control network functions related to the service, such as routing and load balancing.

[0033] During operation, executing requests from client devices may require invoking one or more services in the cloud computing environment, and executing one or more functions of one service may require invoking one or more functions of another service. For example... Figure 2 As shown, after receiving a request from a client device, service A can call service B. Service B can request service D to perform one or more functions. In this embodiment of the application, a cloud service for determining the resources for exchanging items can be deployed in one or more computing nodes.

[0034] The execution process of the method for determining the exchangeable resources provided in this application embodiment is described in detail below with reference to the accompanying drawings. This method for determining the exchangeable resources can be executed by the server-side device 101 in the aforementioned exchangeable resources system.

[0035] Figure 3 A flowchart illustrating a method for determining resource exchange for goods, as provided in this application embodiment, is shown below. Figure 3 As shown, the method includes the following steps: 301. Obtain the attribute information of the item to be redeemed.

[0036] 302. Based on the attribute information of the item to be redeemed, obtain a target item of the same type from the item pool that meets the similarity requirement of the item to be redeemed.

[0037] 303. Determine the range of exchange resources for the item to be exchanged based on the exchange resource value of the same target item.

[0038] 304. Based on the exchange resource value of the target item and the exchange status within the preset time period, determine the target exchange resource value of the item to be exchanged within the exchange resource range.

[0039] In practical applications, the server-side device obtains the attribute information of the item to be redeemed when or after the item is published. The process is as described in the above embodiments and will not be repeated here.

[0040] Based on the attribute information of the item to be redeemed, obtain a target item of the same type from the item pool that meets the similarity requirement to the item to be redeemed.

[0041] The item pool contains items that have already been redeemed. Furthermore, the item pool can also contain posted items. Posted items refer to items that have been posted but not yet redeemed.

[0042] In this context, "target identical item" refers to an item that is identical to the item to be redeemed. Furthermore, "target identical item" refers to an item with the same key attributes as the item to be redeemed. Key attributes are those that significantly influence the redemption resource value of an item. It's evident that the key attributes differ across item categories. Alternatively, "target identical item" can be an item belonging to the same SPU (Single Product Unit) as the item to be redeemed. For example, the key attributes of electronic products are typically: brand, model, and storage capacity. The key attributes of clothing and footwear are typically: brand, style, and material.

[0043] For example, if the item to be redeemed is a mobile phone, then the target item must be the same as the item in key attributes such as brand, model, and memory. If the item to be redeemed is a book, then the target item must be the same as the item in key attributes such as International Standard Book Number (ISBN), edition, and publisher.

[0044] In one possible embodiment, the method for obtaining the target item can be to first search for similar items to the item to be redeemed in the item pool based on the feature vector of the attribute information, and then filter out the target item to be redeemed from the similar items. This will be explained in detail below. Step 302 can be implemented through steps 3021-3023. The execution order of steps 3021 and 3022 is not sequential.

[0045] Step 3021: In the item pool, retrieve the first similar item whose distance between the feature vectors of the attribute information of the item to be redeemed meets the condition.

[0046] Calculate the distance between the feature vector of the attribute information of the item to be redeemed and the feature vectors of the attribute information of each item in the item pool. The smaller the distance between the two feature vectors, the more similar the corresponding items are. Therefore, items with a distance less than or equal to a distance threshold can be identified as the first similar items. This initially identifies items similar to the item to be redeemed.

[0047] The feature vector of the attribute information of the item to be redeemed refers to the vector representation obtained by extracting features from the attribute information of the item to be redeemed. The feature vector of the attribute information of the item to be redeemed is used to represent the attribute information of the item to be redeemed. In one possible embodiment, the attribute information includes descriptive information and an image. When obtaining the first similar item, the following steps are taken: First, in the item pool, a third similar item is retrieved whose distance from the feature vector of the descriptive information of the item to be redeemed meets a certain condition. Second, in the item pool, a fourth similar item is retrieved whose distance from the feature vector of the image of the item to be redeemed meets a certain condition. The aforementioned distance conditions all refer to distances being less than or equal to a corresponding threshold. Therefore, the first similar item includes both the third and fourth similar items.

[0048] Furthermore, an image vector pool containing feature vectors of images of items in the item pool can be pre-constructed. The feature vector of the image of the item to be redeemed is determined in the same way as obtaining the feature vectors of the images of items in the item pool. Then, a third similar item is retrieved from the image vector pool whose distance from the feature vector of the image of the item to be redeemed meets a certain condition.

[0049] Furthermore, a description information vector pool containing feature vectors of description information for items in the item pool can be pre-constructed. The feature vector of the description information for the item to be redeemed is determined in the same way as obtaining the feature vectors of the description information for items in the item pool. Then, a fourth similar item is retrieved from the description information vector pool whose distance from the feature vector of the description information of the item to be redeemed meets a certain condition.

[0050] Step 3022: In the item pool, retrieve the second similar item that appears simultaneously with the item to be redeemed in the historical search results.

[0051] Collaborative filtering can be used to retrieve second-similar items to the item to be redeemed. These second-similar items can be those that appeared simultaneously with the item in historical search results.

[0052] Historical search results refer to the search results returned to the client device when a user searches for items on the item redemption platform in the past. In each search result, items are typically sorted according to their relevance to the search information, or displayed across multiple search pages. Items in the historical search results that are less than a threshold away from the item to be redeemed are considered the closest to the item in the historical search results. Alternatively, items appearing simultaneously with the item to be redeemed in the historical search results can be considered items appearing on the same search page. All of these can be considered items that are relatively similar to the item to be redeemed and therefore can be considered as the second most similar items.

[0053] Among them, the first and second similar items are both similar items to the item to be exchanged.

[0054] Step 3023: Based on the attribute information of the item to be redeemed, the attribute information of the first similar item, and the attribute information of the second similar item, filter out the target item from the first similar item and the second similar item.

[0055] In practical applications, the first and second similar items obtained are only similar to the item to be redeemed, but there are differences in the exchange resource values ​​between some similar items. Therefore, it is necessary to filter out the target items that are the same type as the item to be redeemed from the first and second similar items.

[0056] In one possible embodiment, since a relatively small number of similar items have already been obtained from an item library containing a large number of items, the target identical item can be determined from the similar items using an item filtering model. Specifically, step 3023 can be implemented as follows: The attribute information of the item to be redeemed, the attribute information of the first similar item, the attribute information of the second similar item, and a prompt message are input into the item filtering model, so that the item filtering model, based on the prompt message, performs at least one of the following processes to filter the target identical item from the first and second similar items: The system identifies the item types of the item to be redeemed, the first similar item, and the second similar item. Based on the item type, it extracts key attributes from their respective attribute information to filter for target items of the same type. Then, it filters out combined redemption items and items with unreasonable redemption resource values ​​from the first and second similar items.

[0057] The prompt information is used to indicate the filtering process of the above-mentioned item filtering model.

[0058] The item type can be similar to the electronic products, clothing, and footwear mentioned above. The key attributes here are similar to those mentioned earlier. Furthermore, the correspondence between item type and key attributes can be provided in the prompt message.

[0059] In this context, "items for combination redemption" refers to redeeming in bulk or as a bundled set. Bulk redemption, for example, is indicated in the description as a package of 10 items, requiring the determination of the redemption resource value for one item. Bundle redemption, for example, is redeeming a clothes hanger and a storage box together. Furthermore, information about the items for combination redemption can be provided in the prompt message.

[0060] Items with unreasonable exchange resource values ​​refer to items with significantly different exchange resource values ​​compared to other similar items, for example, a difference greater than 50%. The target similar items include items with outlier exchange resource values; that is, when calculating the exchange resource values ​​of the target similar items, some individual items have exchange resource values ​​that significantly deviate from the overall distribution of exchange resource values ​​for the target similar items, exhibiting abnormally high or low values. Further information on items with unreasonable exchange resource values ​​that need to be excluded can be provided in the prompt message.

[0061] In one possible implementation, since the reference value of the exchange resource value of the same item that has already been redeemed is usually more significant than that of the exchange resource value of the published same item, it is possible to control the proportion of the target same item that has already been redeemed to be greater than a proportion threshold, or to control the number of the target same item that has already been redeemed to be greater than a preset threshold.

[0062] Next, referencing the exchange resource values ​​of the same target item, determine the range of exchange resource values ​​for the item to be exchanged. This allows for obtaining a reasonable fluctuation range in the exchange resource values ​​of the item to be exchanged.

[0063] If the target item has already been redeemed, the redemption resource value is the same as the redemption value of the target item that has already been redeemed. If the target item has been released but not yet redeemed, the redemption resource value is the same as the redemption resource value of the target item when it was released.

[0064] In one possible embodiment, the range of exchange resources is determined based on the exchange resource value of the target item and the attribute factors that affect the exchange resource value of the target item.

[0065] Optionally, the attribute factors affecting the exchange value of resources include: the key attributes of the item to be exchanged. Since key attributes are usually the most important factors affecting the exchange value of resources, the most important factor affecting the exchange value of the same target item is the key attribute. The exchange value of the item to be exchanged can be analyzed and determined based on the exchange value determined by the key attributes.

[0066] Furthermore, the attribute factors that affect the value of exchanged resources may include, but are not limited to, at least one of the following attribute factors: brand premium information, quality information, accessory completeness information, and scarcity information.

[0067] Brand premium refers to the additional exchange value that users are willing to pay for a brand's intangible assets (such as reputation, history, design philosophy, and social recognition), exceeding the functional cost of the product itself. When exchanging resources, even if two items are similar in material and function, well-known brands (especially luxury brands) will receive a higher exchange value due to their associated social status and quality assurance.

[0068] Condition refers to the newness and / or wear and tear of an item. Items of the same SPU in different conditions usually have different exchange resource values. For example, for the same mobile phone model, a phone with its original screen protector still in its original packaging may have an exchange resource value very close to the original value of that phone; while a phone of the same model with a worn casing and scratched screen may have an exchange resource value that is half or even lower.

[0069] Accessories refer to packaging boxes, instruction manuals, warranty cards, chargers, invoices, and free gifts, excluding the main product. Accessory completeness is used to assess whether the original factory-fitted accessories are retained when the item is redeemed.

[0070] Scarcity information refers to the relationship between the supply and demand of a good in the market. Scarcity occurs when the supply of a good is significantly less than the demand. For example, limited-production items, discontinued items, or items from a long history may all be scarce. Scarcity information indicates the degree of scarcity of a good. The higher the scarcity level, the higher the corresponding exchange value of resources.

[0071] Furthermore, a resource exchange value generation model can be used to obtain the range of exchangeable resources. The exchangeable resource value of the target item and the attribute factors affecting its exchangeable resource value are input into the resource exchange value generation model, causing it to output the range of exchangeable resources.

[0072] Optionally, the resource value generation model and the aforementioned item filtering model can be the same model or different models. If the resource value generation model and the aforementioned item filtering model are the same model, then the target item and the range of exchangeable resources can be obtained with a single model call.

[0073] Then, based on the exchange resource value of the target item and the exchange status within the preset time period, the target exchange resource value of the item to be exchanged is determined from the exchange resource range.

[0074] In one possible embodiment, the target exchange resource value of the item to be exchanged can be determined as follows: based on the exchange resource value of the same target item and the exchange situation within a preset time period, a relationship expression between the exchange resource value of the item to be exchanged and the exchange probability is obtained. The target exchange resource value is then determined based on this relationship expression.

[0075] A relational expression is an expression that reflects the relationship between the probability of exchange and the value of the exchanged resource. For example, it can be an expression in the form of a supply and demand function. Based on the exchange resource value of the target item and the exchange situation within a preset time period, the relational expression between the probability of exchange and the exchange resource value of the target item can be calculated, and this can be used as the relational expression for the item to be exchanged. Once the relational expression is determined, the target exchange resource value can be obtained based on it.

[0076] The preset duration refers to the expected time from when the item is posted to when it is redeemed. For example, the preset duration could be 7 days, 15 days, or 30 days.

[0077] Furthermore, the relationship expression can be obtained as follows: Based on the average exchange resource value, the midpoint of the exchange resource range, or the median of the exchange resource range for the same target item, a representative exchange resource value is derived. Here, the average value refers to the average exchange resource value of all target items of the same type. The representative exchange resource value refers to the value within the exchange resource range that best represents that range. Then, based on the exchange performance of the target item within a preset time period, the exchange probability corresponding to the representative exchange resource value is determined. Based on the representative exchange resource value and the exchange probability, the relationship expression between the exchange resource value and the exchange probability of the item to be exchanged is obtained.

[0078] For example, taking the supply and demand function in the following formula (1) as an example, we can construct a relationship expression between the exchange resource value and the exchange probability of the item to be exchanged.

[0079] Formula (1) in, It represents the probability of redeeming the item. The exchange resource value of the items to be exchanged. For parameters.

[0080] Please see Figure 4 , Figure 4 This is a schematic diagram illustrating the principle of determining the target exchange resource value, provided as an embodiment of this application. Figure 4 As shown on the left, the figure is a curve fitted according to formula (1), where the horizontal axis represents the exchange resource value of the item to be exchanged, the vertical axis represents the exchange probability of the item to be exchanged, and p1-p2 is the exchange resource range of the item to be exchanged. It can be seen that the parameters in formula (1) need to be determined. This allows us to determine the relational expression and thus fit a result such as... Figure 4 The curve shown on the left. The situation of the item to be redeemed can be compared by the redemption resource value of the target item and the redemption situation within the preset time. Assuming the preset time is 30 days, the target items include: the first item with a redemption resource value of 30 after 18 days, the second item with a redemption resource value of 35 after 30 days, the third item with a redemption resource value of 28 after 7 days, and the fourth item with a redemption resource value of 32 after 28 days. Then the redemption resource representative value can be the midpoint of 28, 30, 32 and 35, which is 31.5. Since the first, third and fourth items are the ones that are redeemed within 30 days of release, the redemption probability corresponding to the redemption resource representative value is 3 / 4*100%=75%, so a point (31.5, 75%) in formula (1) can be determined. Thus, the parameters can be calculated based on this point. In other words, the relational expression is determined. A corresponding curve can be fitted using this determined relational expression, such as... Figure 4 As shown on the right. Afterwards, it can be achieved by obtaining (… The largest To determine the target exchange resource value of the item to be exchanged from the exchange resource range p1-p2. .

[0081] In summary, the solution provided in the above embodiments of this application obtains the attribute information of the item to be redeemed, and accordingly retrieves target similar items from the item pool that meet the similarity requirements, accurately obtaining target similar items that are meaningful for determining the redemption resource value of the item to be redeemed. Then, based on the redemption resource value of the target similar items, a preliminary range of redemption resource values ​​for the item to be redeemed is defined to avoid deviations caused by fluctuations in a single reference value. Furthermore, considering the actual redemption situation of the target similar items within a preset time period, the target redemption resource value for the item to be redeemed is selected from the redemption resource range. For items to be redeemed with single inventory and unstructured attribute information, by introducing similar items and historical redemption data, a redemption resource value that better reflects market conditions can be obtained, significantly improving the market adaptability of the redemption resource value and promoting the circulation efficiency of the item. Automated determination of redemption resource values ​​reduces the subjectivity and cost of manual intervention, effectively improving the efficiency of item redemption.

[0082] In addition, by determining the range of exchangeable resources through an item screening model and the target exchangeable resource value from the range through a demand function, the limitations of the item screening model (which is prone to exceeding the limit) and the pure demand function model (which heavily relies on historical data) are overcome. This approach is more suitable for determining the exchangeable resource value of items with single inventory and unstructured attribute information, thus effectively improving the efficiency of item exchange.

[0083] Please see Figure 5 , Figure 5 A flowchart illustrating another method for determining resource exchange for goods provided in this application embodiment is shown below. Figure 5 As shown, the method includes the following steps: 501. Obtain the attribute information of the item to be redeemed.

[0084] 502. In the item pool, retrieve the third similar item whose distance from the feature vector of the item to be redeemed meets the condition.

[0085] 503. In the item pool, retrieve the fourth similar item whose distance between the feature vector of the image of the item to be redeemed meets the condition. The first similar item includes the third similar item and the fourth similar item.

[0086] 504. In the item pool, retrieve the second similar item that appears simultaneously with the item to be redeemed in the historical search results.

[0087] 505. Determine whether the item to be redeemed belongs to the standard product type.

[0088] If yes, proceed to step 506; otherwise, proceed to step 507.

[0089] 506. Based on the priority order of third-similar items, fourth-similar items, and second-similar items, obtain a preset number of items that are highly similar to the item to be redeemed, as the target identical items.

[0090] 507. Based on the priority order of fourth similar items, third similar items, and second similar items, obtain a preset number of items that are highly similar to the item to be redeemed, as the target identical items.

[0091] 508. Determine the range of exchange resources for the item to be exchanged based on the exchange resource value of the same target item.

[0092] 509. Based on the exchange resource value of the target item and the exchange status within the preset time period, determine the target exchange resource value of the item to be exchanged within the exchange resource range.

[0093] Steps 501-504 and 508-509 in this embodiment are similar to the steps in the above embodiments, and will not be repeated here.

[0094] In practical applications, regardless of whether the item to be redeemed is a standardized or non-standardized product, the probability that the second similar item recalled by collaborative filtering is the same as the item to be redeemed is relatively low compared to the first similar item recalled by image feature vectors and descriptive information feature vectors.

[0095] For standardized items to be redeemed, it's typically impossible to distinguish their model, configuration, or other information affecting redemption value from their appearance alone; that is, it's impossible to identify the model and configuration from the image of the item. It's usually necessary to rely on the specifications and model information included in the description to determine if items of the same type are identical. Therefore, for standardized items to be redeemed, a predetermined number of items with a high degree of similarity to the item to be redeemed are retrieved according to the priority order of third-similar items, fourth-similar items, and second-similar items, as target identical items. That is, target identical items are prioritized from third-similar items, followed by fourth-similar items. If no target identical item can be found using the above method, second-similar items can be used as a reference to determine the target identical item.

[0096] For non-standard items to be redeemed, which lack a standardized model and whose redemption value depends heavily on appearance, material, design, and condition, image matching can be used to find corresponding target items. Therefore, for non-standard items, a predetermined number of items with a high degree of similarity to the item to be redeemed are retrieved according to the priority order of fourth, third, and second similar items to be considered as target items. That is, target items are prioritized from the fourth similar items, followed by the third. If no target item can be found using the above method, the second similar items can be used as a reference to determine the target item.

[0097] In this embodiment, when determining the target identical item for both standardized and non-standardized items to be redeemed, the priority order of similar items recalled through different methods is used based on the different characteristics of standardized and non-standardized items to determine a more accurate target identical item.

[0098] In some embodiments, before determining the range of exchangeable resources for the item to be exchanged, it is also necessary to obtain the lower limit of exchangeable resources for the item. The lower limit of exchangeable resources refers to the value of the item's exchangeable resources adjusted based on the target exchangeable resource value when the target exchangeable resource value is greater than or equal to the lower limit. Furthermore, the lower limit of exchangeable resources can be set by the user.

[0099] In one possible embodiment, after obtaining the target exchange resource value, it is necessary to determine whether the target exchange resource value is greater than or equal to the lower limit value of the exchange resource.

[0100] If the target exchange resource value is greater than or equal to the minimum exchange resource value, the exchange resource value of the item to be exchanged will be adjusted according to the target exchange resource value.

[0101] If the target exchange resource value is less than the minimum exchange resource value, the exchange resource value of the item to be exchanged will not be adjusted, or the exchange resource value of the item to be exchanged will be adjusted according to the minimum exchange resource value.

[0102] In this embodiment, by setting a lower limit for exchangeable resources, after obtaining the target exchangeable resource value, it is necessary to determine whether the target exchangeable resource value reaches or exceeds the lower limit. If the target exchangeable resource value is greater than or equal to the lower limit, the exchangeable resource value of the item to be exchanged is adjusted according to the target exchangeable resource value; if the target value is less than the lower limit, the original exchangeable resource value remains unchanged, or it is directly adjusted according to the lower limit. This ensures the dynamic adaptability of the exchangeable resource value and avoids situations where the exchangeable resource value is unreasonable or does not meet user expectations due to the target exchangeable resource value being too low, reducing unnecessary frequent adjustments and improving the operating efficiency of the item exchange system.

[0103] In some embodiments, the attribute information of the item to be redeemed can be input into the abnormal item identification model, so that the abnormal item identification model performs the following processing: Based on the attribute information of the item to be redeemed, determine the item type of the item to be redeemed. If the item type of the item to be redeemed is not an abnormal item type, then obtain a target item of the same type from the item pool that meets the similarity requirement of the item to be redeemed. If the item type of the item to be redeemed is an abnormal item type, then the next item to be redeemed can be processed.

[0104] Among them, abnormal item types refer to item types for which the target exchange resource value cannot be obtained.

[0105] Optionally, the anomalous item type includes at least one of the following: Key attributes are missing (e.g., the phone has no model number); The exchange value for resources is far lower than the reasonable exchange value (for example, a brand new mobile phone has an exchange value of 10,000 resources, while the same mobile phone has an exchange value of 1). The description of the item to be redeemed contains multiple different items; Items for which there is no intention to exchange (e.g., items intended for recycling, etc.); Customized items.

[0106] Optionally, the abnormal item type can be indicated by prompt words. Prompt words containing the attribute information of the item to be redeemed and the abnormal item type can be input into the abnormal item recognition model; and / or the abnormal item recognition model can be trained to learn the abnormal item type, thereby obtaining a trained abnormal item recognition model.

[0107] Furthermore, the abnormal item identification model can be the same as or a different model from the item filtering model mentioned above. If the abnormal item identification model and the item filtering model are different models, the attribute information of the item to be redeemed can be input into the abnormal item identification model so that the abnormal item identification model outputs the item type of the item to be redeemed. If the item type of the item to be redeemed does not belong to the abnormal item type, then the above process continues. Figure 3 or Figure 4 The method steps of the illustrated embodiment are as follows: If the item to be redeemed is of an abnormal item type, then the next item to be redeemed can be processed.

[0108] Optionally, if the abnormal item identification model and the item filtering model are the same model, after obtaining the first and second similar items of the item to be redeemed, the item filtering model is used to identify the item type of the item to be redeemed. If the item type of the item to be redeemed does not belong to the abnormal item type, then based on the attribute information of the item to be redeemed, a target item that meets the similarity requirement of the item to be redeemed is obtained from the first and second similar items. If the item type of the item to be redeemed belongs to the abnormal item type, then the next item to be redeemed can be processed. In this way, multiple processing steps can be achieved by calling the model once.

[0109] In some embodiments, the aforementioned models (item selection model, resource value generation model, and abnormal item identification model) can be obtained through training. Furthermore, the aforementioned models can be trained using supervised fine-tuning (SFT) or reinforcement learning (RL).

[0110] If the resource value generation model and the item selection model mentioned above are the same model, they are referred to as the item selection model in the following embodiments. The training process of the resource value generation model and the item selection model if they are the same model is described below.

[0111] Please see Figure 6 , Figure 6 A flowchart illustrating a training method for an item selection model provided in this application embodiment is shown below. Figure 5 As shown, the method includes the following steps: 601. Input the attribute information of the sample items that have been redeemed into the initial item screening model so that the initial item screening model outputs the predicted redemption resource range value of the sample items.

[0112] 602. Based on the exchange resource value of the sample items and the predicted exchange resource range value, determine the accuracy of the predicted exchange resource range value.

[0113] 603. Based on the accuracy of the predicted range of exchangeable resources and the fact that the range of exchangeable resources meets the width constraint, a reward signal is obtained.

[0114] 604. Adjust the parameters of the initial item selection model according to the reward signal to obtain the item selection model.

[0115] In this embodiment, the item selection model can be obtained through reinforcement learning training. The attribute information and exchange resource value of sample items that have already been redeemed are used as sample data. The attribute information of the sample items serves as the input data for the initial item selection model. The exchange resource value of the sample item refers to the actual exchange resource value at the time of successful redemption, and is used as the label (Ground Truth) for that sample item.

[0116] The attribute information of the sample items is input into the initial item selection model to obtain the predicted exchange resource range value of the sample items output by the initial item selection model. For example, the initial item selection model can be an LLM (Local Level Management Model).

[0117] The predicted exchange resource range for sample items needs to be compared with their corresponding exchange resource values. It's necessary to determine whether the exchange resource value falls within the predicted range to assess the accuracy of the prediction. If it falls within the range, the prediction is valid and a reward can be given; if it falls outside the range, the prediction is inaccurate and a penalty signal is given. Furthermore, certain constraints need to be set on the width of the predicted exchange resource range. For example, a predicted range of 1 to 10000, while encompassing the exchange resource value, provides no guidance. Therefore, constraints need to be set to minimize the width of the predicted exchange resource range while including the exchange resource value. For instance, the range width could be set to 20% of the maximum exchange resource value of the target exchange item.

[0118] Furthermore, the reward function can be constructed as follows to obtain the reward signal: If the exchanged resource value falls within the predicted exchanged resource range, and the width of the predicted exchanged resource range meets the constraints, then the prediction is valid, and a positive reward signal can be given. If the exchanged resource value falls outside the predicted exchanged resource range, then the prediction is inaccurate, and a negative reward signal (penalty signal) is given. If the exchanged resource value falls within the predicted exchanged resource range, but the width of the predicted exchanged resource range is too large, i.e., it does not meet the constraints (exceeding the preset threshold or penalty coefficient), a lighter negative reward signal is given.

[0119] Next, the reward signal guides the direction of model parameter updates. Optionally, a loss function can be calculated using a reinforcement learning algorithm, such as Group Relative Policy Optimization (GRPO). The loss is then backpropagated to the internal neural network layers of the initial item selection model. The model's weights and bias parameters are then adjusted.

[0120] Furthermore, the above process can be repeated until the initial item selection model converges, thus obtaining the item selection model.

[0121] In one possible embodiment, the output data of the initial item selection model can also be in a set format, such as JavaScript Object Notation (JSON) format.

[0122] Furthermore, during the training process described above, reward signals can be provided based on the format of the output data of the initial item selection model to facilitate training. For example, a positive reward signal is provided when the output data of the initial item selection model is in the set format; a negative reward signal is provided when the output data of the initial item selection model is not in the set format.

[0123] In this embodiment, the "width constraint" and reward mechanism prevent the model from intentionally outputting an excessively large range in pursuit of high accuracy, making the model's output more reasonable and in line with expectations. Furthermore, by utilizing the reward mechanism of reinforcement learning, the constraints on accuracy and interval width can be flexibly adjusted to adapt to different application scenarios.

[0124] Furthermore, if the resource value generation model and item selection model mentioned above are two different models, then the two models can be obtained through separate training processes or joint training processes. The joint training process is similar to the one described above. Figure 6 The training process shown is similar. The training process for each individual will be described separately below.

[0125] Furthermore, the item selection model can be trained using reinforcement learning. The training process is as follows: The attribute information of the sample item and the attribute information of its similar items are input into the initial item selection model, enabling it to select predicted identical items from the similar items. Then, based on the predicted identical items and the identical items of the sample item, a reward signal reflecting the accuracy of the predicted identical items is determined. The parameters of the initial item selection model are then adjusted according to the reward signal to obtain the final item selection model.

[0126] Furthermore, a resource generation model can be trained using reinforcement learning. The training process is as follows: The resource value of a previously redeemed sample item (of the same type) is input into the initial resource generation model. This allows the model to output a predicted resource range for the sample item based on its redemption value. Then, based on the sample item's redemption value and the predicted resource range, the accuracy of the predicted range is determined. A reward signal is obtained based on the accuracy of the predicted range and whether the range's width satisfies the width constraint. Finally, the parameters of the initial resource generation model are adjusted according to the reward signal to obtain the final resource generation model.

[0127] Figure 7 This is a schematic diagram of the structure of a device for determining resource exchange for goods, provided in an embodiment of this application. Figure 7 As shown, the device includes: Module 701 is used to obtain the attribute information of the item to be redeemed; The same item recognition module 702 is used to obtain target same items that meet the similarity requirements of the item to be redeemed from the item pool based on the attribute information of the item to be redeemed. The resource exchange determination module 703 is used to determine the range of exchange resources for the item to be exchanged based on the exchange resource value of the target item; and to determine the target exchange resource value for the item to be exchanged within the range of exchange resources based on the exchange resource value of the target item and the exchange status within a preset time period.

[0128] In one possible embodiment, the same item identification module 702 is specifically used to: retrieve, in the item pool, a first similar item whose distance between the feature vectors of the attribute information of the item to be redeemed meets the condition; in the item pool, obtain a second similar item that appears simultaneously with the item to be redeemed in the historical search results; and filter out the target same item from the first similar item and the second similar item based on the attribute information of the item to be redeemed, the attribute information of the first similar item, and the attribute information of the second similar item.

[0129] In one possible embodiment, the same item identification module 702 is specifically used to: input the attribute information of the item to be redeemed, the attribute information of the first similar item, the attribute information of the second similar item, and the prompt information into the item filtering model, so that the item filtering model can filter out the target same item from the first similar item and the second similar item by performing at least one of the following processes according to the prompt information: identifying the item types of the item to be redeemed, the first similar item, and the second similar item, and extracting key attributes from the corresponding attribute information according to the item type, so as to filter the target same item according to the item type and key attributes; filtering out items that are combined for redemption and items with unreasonable redemption resource values ​​from the first similar item and the second similar item.

[0130] In one possible embodiment, the attribute information includes description information and an image; the same item recognition module 702 is specifically used to: retrieve, in the item pool, a third similar item whose distance between the feature vector of the description information of the item to be redeemed meets the condition; and retrieve, in the item pool, a fourth similar item whose distance between the feature vector of the image of the item to be redeemed meets the condition, wherein the first similar item includes the third similar item and the fourth similar item.

[0131] In one possible embodiment, the same item identification module 702 is specifically used to: if the item to be redeemed is a standard item, then according to the priority order of the third similar item, the fourth similar item, and the second similar item, obtain a preset number of items with a greater degree of similarity to the item to be redeemed, as the target same item; if the item to be redeemed is a non-standard item, then according to the priority order of the fourth similar item, the third similar item, and the second similar item, obtain a preset number of items with a greater degree of similarity to the item to be redeemed, as the target same item.

[0132] In one possible embodiment, the exchange resource determination module 703 is specifically used to: determine the exchange resource range value based on the exchange resource value of the target item and the attribute factors that affect the exchange resource value of the target item.

[0133] In one possible embodiment, the exchange resource determination module 703 is specifically used to: input the exchange resource value of the target same item and the attribute factors of the target same item that affect the exchange resource value into the item filtering model, so that the item filtering model outputs the exchange resource range value; The device also includes: a training model for inputting attribute information of sample items that have been redeemed into an initial item selection model, so that the initial item selection model outputs the predicted redemption resource range value of the sample items; determining the accuracy of the predicted redemption resource range value based on the redemption resource value and the predicted redemption resource range value; obtaining a reward signal based on the accuracy of the predicted redemption resource range value and whether the interval width of the redemption resource range value meets the width constraint condition; and adjusting the parameters of the initial item selection model based on the reward signal to obtain the item selection model.

[0134] In one possible embodiment, the acquisition module is further configured to: acquire the minimum exchange resource value of the item to be redeemed; The resource exchange determination module 703 is also used to: determine the range of exchange resources for the item to be exchanged if the target exchange resource value is greater than or equal to the lower limit of the exchange resource value, based on the exchange resource value of the target item; and adjust the exchange resource value of the item to be exchanged based on the target exchange resource value.

[0135] In one possible embodiment, the resource redemption determination module 703 is specifically used to: obtain a relational expression between the redemption resource value and the redemption probability of the item to be redeemed based on the redemption resource value of the target item and the redemption status within a preset time period; and determine the target redemption resource value based on the relational expression.

[0136] In one possible embodiment, the resource redemption determination module 703 is specifically used to: obtain a representative value of the redemption resource based on the average value of the redemption resource for the same target item, or the midpoint or median of the range of redemption resource values; determine the redemption probability corresponding to the representative value of the redemption resource based on the redemption status of the same target item within a preset time period; and obtain a relational expression between the redemption resource value and the redemption probability of the item to be redeemed based on the representative value of the redemption resource and the redemption probability.

[0137] In one possible embodiment, the same item identification module 702: inputs the attribute information of the item to be redeemed into the abnormal item identification model, so that the abnormal item identification model performs the following processing: determines the item type of the item to be redeemed based on the attribute information of the item to be redeemed; if the item type of the item to be redeemed does not belong to the abnormal item type, then, based on the attribute information of the item to be redeemed, obtains a target same item from the item pool that meets the similarity requirement of the item to be redeemed.

[0138] Figure 7 The apparatus shown can perform the steps in the methods in the foregoing embodiments. For detailed execution process and technical effects, please refer to the description in the foregoing embodiments, which will not be repeated here.

[0139] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application, such as... Figure 8 As shown, in practice, this electronic device includes a memory 21 and a processor 22.

[0140] Memory 21 is used to store computer programs and can be configured to store various other data to support operation on the electronic device. Examples of this data include instructions for any application or method used to operate on the electronic device, data structures, contact data, phone book data, messages, pictures, videos, etc.

[0141] The processor 22, coupled to the memory 21, is used to execute the computer program in the memory 21 to implement the method provided in the foregoing embodiments.

[0142] Furthermore, such as Figure 8 As shown, the electronic device also includes other components such as a communication component 23, a display 24, a power supply component 25, and an audio component 26. Figure 8 The diagram only shows some components and does not mean that the electronic device includes only these components. Figure 8 The components shown are as follows. The electronic device in this embodiment can be a terminal device such as a desktop computer, laptop computer, smartphone, or IoT device, or a server device such as a conventional server, cloud server, or server array.

[0143] The aforementioned memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random-Access Memory (SRAM), Electrically Erasable Programmable Read Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0144] The aforementioned communication component is configured to facilitate wired or wireless communication between the device containing the communication component and other devices. The device containing the communication component can access wireless networks based on communication standards, such as 2G, 3G, 4G / LTE, 5G, or combinations thereof. In one exemplary embodiment, the communication component receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel.

[0145] The aforementioned display includes a screen, which may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a Touch Panel, the screen can be implemented as a touchscreen to receive input signals from the user. The Touch Panel includes one or more touch sensors to sense touches, swipes, and gestures on the Touch Panel. The touch sensors can sense not only the boundaries of touch or swipe actions but also the duration and pressure associated with the touch or swipe operation.

[0146] The aforementioned power supply components provide power to various components within the device in which they reside. These power supply components may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power to the device in which they reside.

[0147] The aforementioned audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC) configured to receive external audio signals when the device containing the audio component is in an operating mode, such as call mode, recording mode, or voice recognition mode. The received audio signals can be further stored in memory or transmitted via a communication component. In some embodiments, the audio component also includes a speaker for outputting audio signals.

[0148] Accordingly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to implement the steps in the above-described method embodiments. The computer-readable storage medium includes volatile or non-volatile components, or a combination thereof, and can be removable or non-removable. Examples of computer-readable storage media include, but are not limited to, phase-change random access memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random-access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), flash memory or other memory technologies, CD-ROM, digital video disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium.

[0149] Accordingly, this application also provides a computer program product, which includes a computer program or instructions. When the computer program or instructions are executed by a processor, the processor is able to implement the steps in the above method embodiments. It should be understood that each step or combination of steps in the above method flow can be implemented by the computer program or instructions. In addition, these computer programs or instructions can be applied to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device, so that the processor of the general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device can be implemented as a means to implement the corresponding functions in the above method embodiments.

[0150] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for determining resources for exchanging goods, characterized in that, The method includes: Obtain the attribute information of the item to be redeemed; Based on the attribute information of the item to be redeemed, a target item of the same type that meets the similarity requirement to the item to be redeemed is obtained from the item pool; Based on the exchange resource value of the target item, determine the exchange resource range value of the item to be exchanged; Based on the exchange resource value of the target item and the exchange status within a preset time period, the target exchange resource value of the item to be exchanged is determined within the exchange resource range.

2. The method according to claim 1, characterized in that, The step of obtaining a target item of the same type that meets the similarity requirement to the item to be redeemed from the item pool based on the attribute information of the item to be redeemed includes: In the item pool, retrieve the first similar item whose distance between the feature vectors of the attribute information of the item to be redeemed meets the condition; From the item pool, retrieve the second similar item that appears simultaneously with the item to be redeemed in the historical search results; Based on the attribute information of the item to be redeemed, the attribute information of the first similar item, and the attribute information of the second similar item, the target identical item is selected from the first similar item and the second similar item.

3. The method according to claim 2, characterized in that, The step of filtering the target identical item from the first similar item and the second similar item based on the attribute information of the item to be redeemed, the attribute information of the first similar item, and the attribute information of the second similar item includes: The attribute information of the item to be redeemed, the attribute information of the first similar item, the attribute information of the second similar item, and the prompt information are input into the item filtering model, so that the item filtering model can filter out the target identical item from the first similar item and the second similar item by performing at least one of the following processes based on the prompt information: Identify the item types of the item to be redeemed, the first similar item, and the second similar item, and extract key attributes from their respective attribute information based on the item type, so as to filter the target similar item based on the item type and the key attributes; Filter out items that can be combined for exchange and items with unreasonable exchange resource values ​​from the first and second similar items.

4. The method according to claim 2, characterized in that, The attribute information includes descriptive information and an image; retrieving, from the item pool, a first similar item whose distance between its feature vector and the attribute information of the item to be redeemed meets a certain condition includes: In the item pool, a third similar item whose distance between the feature vectors of the description information of the item to be redeemed meets the condition is retrieved; In the item pool, a fourth similar item is retrieved that meets the distance condition between the feature vector of the image of the item to be redeemed. The first similar item includes the third similar item and the fourth similar item.

5. The method according to claim 4, characterized in that, The step of filtering the target identical item from the first similar item and the second similar item includes: If the item to be redeemed is a standard item, then according to the priority order of the third similar item, the fourth similar item, and the second similar item, a preset number of items with a high degree of similarity to the item to be redeemed are obtained as the target identical item; If the item to be redeemed is a non-standard product, then according to the priority order of the fourth similar item, the third similar item, and the second similar item, the preset number of items with a greater degree of similarity to the item to be redeemed are obtained as the target identical item.

6. The method according to any one of claims 1-5, characterized in that, The step of determining the exchange resource range of the item to be exchanged based on the exchange resource value of the target item includes: The range of exchange resources is determined based on the exchange resource value of the target item and the attribute factors that affect the exchange resource value of the target item.

7. The method according to claim 6, characterized in that, The step of determining the exchange resource range based on the exchange resource value of the target item and the attribute factors affecting the exchange resource value of the target item includes: The exchange resource value of the target item and the attribute factors that affect the exchange resource value of the target item are input into the item filtering model so that the item filtering model outputs the exchange resource range value. The method further includes: The attribute information of the sample items that have been redeemed is input into the initial item screening model so that the initial item screening model outputs the predicted redemption resource range value of the sample items; Based on the exchange resource value of the sample items and the predicted exchange resource range value, determine the accuracy of the predicted exchange resource range value; A reward signal is obtained based on the accuracy of the predicted range of exchangeable resources and the fact that the range of exchangeable resources satisfies the width constraint. The parameters of the initial item selection model are adjusted based on the reward signal to obtain the item selection model.

8. The method according to any one of claims 1-5 and 7, characterized in that, The method further includes: Obtain the minimum exchange resource value for the item to be exchanged; If the target exchange resource value is greater than or equal to the lower limit of the exchange resource value, then the exchange resource value of the item to be exchanged is adjusted according to the target exchange resource value.

9. The method according to any one of claims 1-5 and 7, characterized in that, The step of determining the target exchange resource value of the item to be exchanged within the exchange resource range based on the exchange resource value of the target item and the exchange status within a preset time period includes: Based on the exchange resource value of the target item and the exchange status within a preset time period, the relationship expression between the exchange resource value and the exchange probability of the item to be exchanged is obtained. The target exchange resource value is determined based on the relational expression.

10. The method according to claim 9, characterized in that, The step of obtaining the relationship expression between the exchange resource value and the exchange probability of the item to be exchanged based on the exchange resource value of the target item and the exchange situation within a preset time period includes: The representative value of the exchange resource is obtained based on the average value of the exchange resource for the same target item, the midpoint of the exchange resource range, or the median of the exchange resource range. Based on the redemption status of the target item within a preset time period, determine the redemption probability corresponding to the redemption resource representative value; Based on the exchange resource representative value and the exchange probability, the relationship expression between the exchange resource value and the exchange probability of the item to be exchanged is obtained.

11. The method according to any one of claims 2-5, 7, and 10, characterized in that, The step of obtaining a target item of the same type that meets the similarity requirement to the item to be redeemed from the item pool based on the attribute information of the item to be redeemed includes: The attribute information of the item to be redeemed is input into the abnormal item identification model, so that the abnormal item identification model performs the following processing: determining the item type of the item to be redeemed based on the attribute information of the item to be redeemed; if the item type of the item to be redeemed does not belong to the abnormal item type, then obtaining a target identical item that meets the similarity requirement of the item to be redeemed from the item pool based on the attribute information of the item to be redeemed.

12. An electronic device, characterized in that, include: The device includes a memory, a processor, and a communication interface; wherein the memory stores executable code, which, when executed by the processor, causes the processor to perform the method as described in any one of claims 1 to 11.

13. A non-transitory machine-readable storage medium, characterized in that, The non-transitory machine-readable storage medium stores executable code that, when executed by a processor of an electronic device, causes the processor to perform the method as described in any one of claims 1 to 11.

14. A computer program product, characterized in that, include: A computer program, when executed by a processor of an electronic device, causes the processor to perform the method as described in any one of claims 1 to 11.

Images