Product recommendation method and device, electronic equipment, storage medium and program product

Abstract

The embodiment of the application discloses a product recommendation method and device, electronic equipment, storage medium and program product; in the embodiment of the application, the correlation weight between each object is determined according to the product correlation weight of multiple objects for multiple different products and the evaluation correlation weight of multiple objects for the same product in multiple different products, and the adjacent objects are determined according to the correlation weight to obtain an adjacent object group; the correlation weight between each object is updated according to the correlation weight between the objects in the adjacent object group to obtain the updated correlation weight; each object is divided according to the updated correlation weight to obtain an object group; and the target object is recommended for a product based on the product of other objects in the object group. The embodiment of the application can improve the accuracy of product recommendation. The embodiment of the application can be applied to cloud technology, artificial intelligence, intelligent transportation, auxiliary driving and various scenes.

Description

Technical Field

[0001] This application relates to the field of recommendation technology, specifically to a product recommendation method, apparatus, electronic device, storage medium, and program product, wherein the storage medium is a computer storage medium and the program product is a computer program product. Background Technology

[0002] With the development of the internet, online shopping is becoming increasingly popular. However, when shopping online, users face a wide variety of products, requiring them to spend considerable time selecting the items they need, resulting in a longer purchase timeframe.

[0003] Currently, in order to enable users to quickly purchase the products they need, products are recommended based on the user's browsing history. However, the accuracy of the recommended products is relatively low. Summary of the Invention

[0004] This application provides a product recommendation method, apparatus, electronic device, storage medium, and program product, which can solve the technical problem of low accuracy in product recommendation results.

[0005] A product recommendation method includes: Obtain the product association weights of multiple objects for various different products; Obtain the evaluation association weights of the same product across multiple objects and various different products; Based on the above product association weights and the above evaluation association weights, the association weights between each of the above objects are determined, and adjacent objects are determined based on the above association weights to obtain adjacent object groups. Based on the association weights between objects in the above adjacent object groups, the association weights between each of the above objects are updated to obtain the updated association weights; Based on the updated association weights, the objects are divided into groups. Based on the products of other objects within the aforementioned object group, product recommendations are made to the target object. The aforementioned other objects are objects within the aforementioned object group other than the target object.

[0006] Accordingly, embodiments of this application provide a product recommendation device, including: The first acquisition module is used to acquire the product association weights of multiple objects for various different products; The second acquisition module is used to acquire the evaluation association weights of the above multiple objects for the same product in multiple different types of products; The weight determination module is used to determine the association weight between each of the above objects based on the above product association weight and the above evaluation association weight, and to determine the adjacent objects based on the above association weight to obtain the adjacent object group. The weight update module is used to update the association weights between each of the above objects based on the association weights between objects in the above adjacent object groups, so as to obtain the updated association weights. The object partitioning module is used to partition each of the above objects according to the updated association weights to obtain object groups; The product recommendation module is used to recommend products to the target object based on the products of other objects within the aforementioned object group. The aforementioned other objects are objects within the aforementioned object group other than the target object.

[0007] Optionally, the weight update module is specifically used to perform: Candidate objects are selected from each of the above objects, and a first adjacent object group containing the above candidate objects and a second adjacent object group containing the adjacent objects of the above candidate objects are selected from each of the above adjacent object groups. Based on the aforementioned first adjacent object group, determine the first object weight of the aforementioned candidate object; Based on the aforementioned second adjacent object group, determine the second object weights of the adjacent objects of the aforementioned candidate objects; Based on the first object weight and the second object weight, the association weights between objects in the first adjacent object group are updated to obtain the updated association weights.

[0008] Optionally, the weight update module is specifically used to perform: Determine the number of the first adjacent object groups mentioned above; Based on the aforementioned number and the association weights between objects in the aforementioned first adjacent object group, the first object weight of the aforementioned candidate object is determined.

[0009] Optionally, the weight update module is specifically used to perform: The first object weight and the second object weight are fused together to obtain the target object weight. The weights of the target object and the association weights between objects in the first adjacent object group are updated and merged to obtain the updated association weights between objects in the first adjacent object group.

[0010] Optionally, the second acquisition module is specifically used to perform: Obtain evaluation information from multiple objects regarding the same product across various different products; Determine the type of the above evaluation information; Based on the types of evaluation information mentioned above, determine the evaluation association weights of the above multiple objects for the same product among multiple different products.

[0011] Optionally, the object partitioning module is also used to perform: From the above objects, filter out the objects to be divided that are not in the above object groups; Based on the updated association weights between the objects to be segmented and each object in the object group, determine the group association weights between the objects to be segmented and the object group. The objects to be divided are assigned to the object group corresponding to the largest group association weight.

[0012] Optionally, the object partitioning module is also used to perform: Filter out the target object group from the above object group, where the number of objects in the target object group is less than the preset number; Determine each of the neighboring object groups of the target object group, wherein the neighboring object groups are the groups of related objects that have the updated association weights with objects in the target object group; Based on the updated association weights between objects in the target object group and the associated objects, determine the group association weights between the target object group and the adjacent object groups. The target object group is updated based on the aforementioned group association weights.

[0013] Furthermore, this application also provides an electronic device, including a processor and a memory, wherein the memory stores a computer program, and the processor is used to run the computer program in the memory to implement the product recommendation method provided in this application.

[0014] Furthermore, embodiments of this application also provide a computer-readable storage medium storing a computer program adapted for loading by a processor to execute any of the product recommendation methods provided in embodiments of this application.

[0015] Furthermore, this application also provides a computer program product, including a computer program that, when executed by a processor, implements any of the product recommendation methods provided in this application.

[0016] In this embodiment, firstly, product association weights for multiple objects across various products and evaluation association weights for multiple objects across the same product across various product types are obtained. Then, based on the product association weights and evaluation association weights, the association weights between each object are determined, and adjacent objects are identified based on these association weights, resulting in adjacent object groups. Next, the association weights between each object are updated based on the association weights between objects in the adjacent object groups, resulting in updated association weights. Then, each object is divided into object groups based on the updated association weights. Finally, based on the products of other objects within the object groups, product recommendations are made for the target object; these other objects are all objects within the object groups other than the target object.

[0017] In this embodiment, the association weights are first determined based on the evaluation association weights and product association weights between each object. Then, adjacent objects are determined based on the association weights to obtain adjacent object groups. Next, the association weights between each object are updated based on the association weights between objects in the adjacent object groups to obtain updated association weights. This updated association weights are used to divide each object into more accurate object groups. As a result, when recommending products to the target object based on the products of other objects within the object group, the recommended products are more in line with the needs of the target object, thus improving the accuracy of product recommendations. Attached Figure Description

[0018] 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 only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of a product recommendation process provided in an embodiment of this application; Figure 2 This is a flowchart illustrating the product recommendation method provided in the embodiments of this application; Figure 3 This is a schematic diagram of the products and objects provided in the embodiments of this application; Figure 4 This is a schematic diagram of the first relationship network diagram provided in the embodiments of this application; Figure 5 This is a flowchart illustrating another product recommendation method provided in an embodiment of this application; Figure 6 This is a schematic diagram of the influence ranking and friend ranking provided in the embodiments of this application; Figure 7 This is a schematic diagram of another product recommendation method provided in an embodiment of this application; Figure 8 This is a schematic diagram illustrating the generation process of the first and second relational network diagrams after partitioning, as provided in the embodiments of this application. Figure 9 This is a schematic diagram of the product recommendation device provided in the embodiments of this application; Figure 10 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0020] 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, and 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.

[0021] This application provides a product recommendation method, apparatus, electronic device, storage medium, and program product. The product recommendation apparatus can be integrated into an electronic device, which may be a server or a terminal, etc.

[0022] The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, network acceleration services (Content Delivery Network, CDN), as well as big data and artificial intelligence platforms.

[0023] Furthermore, multiple servers can form a blockchain, with the servers being nodes on the blockchain.

[0024] The terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, vehicle terminal, aircraft, etc., but is not limited to these. The embodiments of this application can be applied to various scenarios, including but not limited to cloud technology, artificial intelligence, smart transportation, and assisted driving.

[0025] Terminals and servers can be connected directly or indirectly via wired or wireless communication, and this application does not impose any restrictions on this.

[0026] For example, such as Figure 1As shown, the server obtains the product association weights of multiple objects for various types of products and the evaluation association weights of multiple objects for the same product among various different products. Then, based on the product association weights and evaluation association weights, it determines the association weights between each object and identifies adjacent objects based on these weights, thus obtaining adjacent object groups. Next, the server updates the association weights between each object according to the association weights between objects in the adjacent object groups, obtaining updated association weights. Based on the updated association weights, it then divides each object into object groups. Finally, when the server receives a recommendation request from the terminal, which includes the target object, the server recommends products to the target object based on the products of other objects within the object group and returns the recommended products to the terminal, which then displays the recommended products.

[0027] In the specific implementation of this application, data such as user information are involved. When the above embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0028] Furthermore, in the embodiments of this application, "multiple" refers to two or more. The terms "first" and "second," etc., in the embodiments of this application are used for distinguishing descriptions and should not be construed as implying relative importance.

[0029] The following sections provide detailed descriptions of each example. It should be noted that the order in which the embodiments are described is not intended to limit the preferred order of the embodiments.

[0030] In this embodiment, the description will be from the perspective of the product recommendation device. To facilitate the explanation of the product recommendation method of this application, the following will describe the product recommendation device integrated into the terminal in detail, that is, the terminal will be used as the execution subject in the detailed explanation.

[0031] Please see Figure 2 , Figure 2 This is a flowchart illustrating a product recommendation method provided in an embodiment of this application. The product recommendation method may include: S201. Obtain the product association weights of multiple objects for various different products.

[0032] A product refers to something that satisfies a certain need, including tangible goods and intangible services. For example, a product can be at least one of the following: merchandise, content (such as posts, links, videos, and push notifications), and services (where a service is a program that fulfills a need, such as sports services, ride-hailing services, and payment services). An object refers to a user associated with the product. For example, when the product is a merchandise, the object can be the user who purchased the product. When the product is content, the object can be the user who viewed the product; when the product is a service, the object can be the user who used the product.

[0033] To obtain the product association weights of an object for multiple different products, one approach is to retrieve the products of each object from its historical records and then determine the product association weights for each object based on those products. Alternatively, an input interface can be displayed, and in response to the object's input on that interface, the products of each object can be retrieved, and then the product association weights for each object can be determined based on those products.

[0034] The method for obtaining the product information can be chosen according to the actual situation; this example does not impose any restrictions.

[0035] After the terminal obtains multiple different products from various objects, it can filter out multiple different products of the same type from these different products, and then determine the product association weight between each object based on these multiple different products of the same type. The same type can refer to at least one of the following: same manufacturer, same brand, and same product category.

[0036] For example, if the product is a commodity, object a buys commodity 1, object b buys commodity 2, and object c buys commodity 3, commodity 1 and commodity 3 are both commodities of merchant T, and commodity 2 is a commodity of merchant Q. Then, the various different products of each object include commodity 1, commodity 2, and commodity 3. The various different products under the same type can be commodity 1 and commodity 3. Therefore, based on commodity 1 and commodity 3, the product association weight between object a and object c can be obtained.

[0037] The product association weight between objects of different products of the same type can be a pre-set first target value. For example, the first target value can be 2, that is, if the product association weight is set to 2, the product association weight between the two objects is 2 when both objects have purchased products from merchant T.

[0038] If the products of two objects are not of the same type, the product association weight between the two objects can be a second target value, which can be, for example, 0.

[0039] The terminal can obtain the product association weights of multiple objects for various different products when it receives a recommendation request. Alternatively, the terminal can obtain the product association weights of multiple objects for various different products when it detects that the accounts of each object have successfully logged in.

[0040] The method for a terminal to obtain the product association weights of multiple objects for various different products can be selected according to the actual situation, and this embodiment does not limit it here.

[0041] S202. Obtain the evaluation association weights of multiple objects for the same product across multiple different products.

[0042] The terminal can simultaneously obtain the product association weights of multiple objects for various different products, and the evaluation association weights of multiple objects for the same product among these various products. Alternatively, the terminal can first obtain the product association weights of multiple objects for various different products, and then obtain the evaluation association weights of multiple objects for the same product among these various products.

[0043] Alternatively, the terminal can first obtain the evaluation association weights of multiple objects for the same product across multiple different products, and then obtain the product association weights of multiple objects for multiple different products.

[0044] The terminal can first filter out products that are associated with each object from the products of each object, and then determine the evaluation association weight between each object based on the products that are associated with each object.

[0045] For example, if there are multiple different products including product 1, product 2, and product 3, and multiple objects including object a, object b, and object c, and objects a, b, and c all purchased product 1, then obtain the evaluation association weights of objects a, b, and c for product 1.

[0046] In some embodiments, evaluation information of multiple objects for the same product among multiple different products can be obtained, and the evaluation association weight of multiple objects for the same product among multiple different products can be determined based on the content of the evaluation information.

[0047] For example, the content of the evaluation information can be identified to obtain the identification results. Then, based on the identification results, the degree of liking of the object corresponding to the evaluation information for the product can be determined. Next, based on the degree of liking of each object for the product, the evaluation association weight between each object can be determined.

[0048] Different levels of liking can correspond to different evaluation association weights. For example, if object a likes product 1 to level s1 and object b likes product 1 to level s3, since the evaluation association weight between s1 and s3 is 1, the evaluation association weight between object a and object b is also 1.

[0049] In other embodiments, the evaluation association weights of multiple objects for the same product across multiple different products are obtained, including: Retrieve evaluation information of the same product from multiple objects across various different product categories; Determine the type of evaluation information; Based on the type of evaluation information, determine the evaluation association weights of multiple objects for the same product across multiple different products.

[0050] In this embodiment, the type of evaluation information can refer to the level of evaluation information. Then, the evaluation association weight of each object for the same product can be determined directly based on the type of evaluation information, so that the evaluation association weight can be obtained quickly.

[0051] The process of determining the evaluation association weights of multiple objects for the same product across multiple different products, based on the type of evaluation information, can be as follows: When the evaluation information of various objects for the same product is of the same type, the evaluation association weight between objects with the same evaluation information type is determined as the third target value; When different objects have different types of evaluation information for the same product, the evaluation association weight between objects with the same type of evaluation information is determined as the fourth target value, and the third target value is greater than the fourth target value.

[0052] For example, if we set the third target value to 3 and the fourth target value to 1, and both object a and object b purchased product 1, and their reviews of product 1 are positive, and both object b and object c purchased product 2, with object b giving a positive review and object c giving a negative review, then the review association weight between object a and object b is 3, and the review association weight between object b and object c is 1.

[0053] It should be noted that there are instances where the type of review information does not match the content of the review information. For example, if the review information is classified as positive, but the content describes the product's shortcomings, then the type of review information does not match the content of the review information.

[0054] To more accurately determine the evaluation association weights between various objects based on the type of evaluation information, before determining the evaluation association weights of multiple objects for the same product across multiple different products, sentiment recognition can be performed on the evaluation information content to obtain the sentiment recognition result. If the sentiment recognition result does not match the type of evaluation information, the type of evaluation information is adjusted based on the sentiment recognition result to obtain the adjusted type of evaluation information. Finally, the terminal determines the evaluation association weights of multiple objects for the same product across multiple different products based on the adjusted type of evaluation information.

[0055] S203. Based on the product association weight and the evaluation association weight, determine the association weight between each object, and determine the adjacent objects based on the association weight to obtain the adjacent object group.

[0056] After obtaining the product association weight and the evaluation association weight, the terminal can use these weights as the association weights between various objects. For example, if object a and object b both purchased product 1, object b also purchased product 2, and object c purchased product 3, where both product 2 and product 3 are products of merchant T, then the association weight between object a and object b is the evaluation association weight, and the association weight between object b and object c is the product association weight.

[0057] After obtaining the product association weight and the evaluation association weight, the terminal can take the objects with association weight as adjacent objects to obtain the adjacent object group.

[0058] It should be noted that a group of adjacent objects can include two adjacent objects, or three or more adjacent objects.

[0059] For example, if the association weight between object a and object b is 3, the association weight between object b and object c is 2, and the association weight between object a and object c is 1, then object a and object b are adjacent objects, forming an adjacent object group ab; object b and object c are adjacent objects, forming an adjacent object group bc; and object a and object c are adjacent objects, forming an adjacent object group ac. Alternatively, object a, object b, and object c form an adjacent object group abc.

[0060] Alternatively, after obtaining the product association weight and the evaluation association weight, the terminal can use the product association weight and the evaluation association weight as connecting edges and the various objects as nodes to construct the first relationship network graph.

[0061] For example, such as Figure 3As shown, objects a and b purchased product 1, objects b and c purchased product 2, and object d purchased product 3. Both product 1 and product 3 are products of merchant T. The evaluation information of objects a and b for product 1 is of the same type, so the third objective value is 3, meaning the evaluation association weight between objects a and b is 3. The evaluation information of objects b and c for product 2 is of different types, so the fourth objective value is 1, meaning the evaluation association weight between objects b and c is 1. The first objective value is 2, meaning the product association weight between objects d and a is 2, and the product association weight between objects d and b is 2.

[0062] Using objects a, b, c, and d as nodes, and product association weights and evaluation association weights as connecting edges, the constructed association network graph can be shown as follows: Figure 4 As shown.

[0063] After obtaining the first relational network graph, the terminal determines the adjacent objects based on the first relational network graph to obtain the adjacent object group. At this time, the adjacent objects can refer to two objects that have a connecting edge.

[0064] For example, such as Figure 4 As shown, objects a and b are connected by an edge, so objects a and b are adjacent objects, forming an adjacent object group ab. Objects a and d are connected by an edge, so objects a and d are adjacent objects, forming an adjacent object group ad. Objects b and c are connected by an edge, so objects b and c are adjacent objects, forming an adjacent object group bc. Objects b and d are connected by an edge, so objects b and d are adjacent objects, forming an adjacent object group bd.

[0065] In some embodiments, after obtaining the product association weight and the evaluation association weight, the friend relationship between each object can also be obtained. If there is a friend relationship between objects with product association weight and evaluation association weight, a sixth target value can be added to the product association weight or the evaluation association weight to obtain the association weight, thereby increasing the association weight between objects with product association weight and evaluation association weight.

[0066] For example, such as Figure 4 As shown, the evaluation association weight between object a and object b is 3, and the sixth objective value is 1. If object a and object b are friends, the evaluation association weight 3 can be increased by 1, resulting in an association weight of 4 between object a and object b.

[0067] S204. Update the association weights between objects based on the association weights between objects in adjacent object groups to obtain the updated association weights.

[0068] After obtaining each adjacent object group, the terminal can filter out candidate objects from each object, and filter out the first adjacent object group containing the candidate objects and the second adjacent object group containing the adjacent objects of the candidate objects from each adjacent object group.

[0069] For example, adjacent object groups include adjacent object groups ab, ad, bc, bd, and ce. Adjacent object group ab includes object a and object b, adjacent object group ad includes object a and object d, adjacent object group bc includes object b and object c, adjacent object group bd includes object b and object d, and adjacent object group ce includes object c and object e.

[0070] When the candidate object is object a, the adjacent objects of the candidate object are object b and object d. The first adjacent object group is adjacent object group ab and adjacent object group ad, and the second adjacent object group is adjacent object group bc and adjacent object group bd.

[0071] Then, based on the association weights between objects in the second adjacent object group, the association weights between objects in the first adjacent object group are updated to obtain the updated association weights. That is, based on the adjacent objects of the candidate object and the indirect adjacent objects of the candidate object (e.g., ... Figure 4 As shown, object c is an indirect neighbor of object a) and the association weight between them is updated.

[0072] Alternatively, after obtaining the first adjacent object group and the second adjacent object group, the terminal can first determine the first object weight of the candidate object based on the first adjacent object group, determine the second object weight of the adjacent objects of the candidate object based on the second adjacent object group, and then update the association weight between objects in the first adjacent object group based on the first object weight and the second object weight to obtain the updated association weight.

[0073] The first object weight is used to represent the importance of the candidate object in the partitioning process, and the second object weight is used to represent the importance of the candidate object's neighboring objects in the partitioning process. The greater the association weight between the candidate object and its neighboring objects, the greater the importance of the candidate object. The more neighboring objects a candidate object has, the less important the candidate object becomes.

[0074] The terminal can determine the first object weight of a candidate object based on the association weight between each object in the first adjacent object group and the number of adjacent objects of the candidate object in the first adjacent object group.

[0075] It should be understood that the relationships between candidate objects and their adjacent objects, as well as the relationships between the weights of the first and second objects, are relative. For example, ... Figure 4 As shown, when object a is a candidate object, its neighboring objects are objects b and d. The object weight of object a is the first object weight, and the object weight of object b is the second object weight. When object b is a candidate object, its neighboring objects are objects a, c, and d. In this case, the object weight of object b is the first object weight, and the object weight of object a is the second object weight.

[0076] When an adjacent object group includes two objects, in some embodiments, determining a first object weight for a candidate object based on the first adjacent object group includes: Determine the number of the first adjacent object group; The first object weight of the candidate object is determined based on the number of objects and the association weight between objects in the first adjacent object group.

[0077] The number of the first adjacent object group is also the number of adjacent objects of the candidate object. The terminal can substitute the number of adjacent objects of the candidate object and the association weight between each adjacent object of the candidate object into the following formula to calculate the average weight of the candidate object:

[0078] W(user i ) Indicates the first i The first object weight of each candidate object friend_w ij Indicates the first i The candidate and the first i The first candidate j The association weight between adjacent objects n i Indicates the first i The number of neighboring objects of each candidate object.

[0079] For example, such as Figure 4 As shown, the first object weight of object a is 2.5, the first object weight of object b is 2, the first object weight of object d is 2, and the first object weight of object c is 1.

[0080] It should be understood that the method for determining the second object weight of the neighboring objects of the candidate object based on the second neighboring object group is the same as the method for determining the first object weight of the candidate object based on the first neighboring object group, and will not be described again in this embodiment.

[0081] Optionally, since the calculation methods for the first object weight and the second object weight are the same, in order to reduce the amount of computation, after obtaining the first object weight of the candidate object and the second object weight of the candidate object's neighboring objects, the neighboring objects of the candidate object can be marked so that the candidate object can be selected from the unmarked objects in the future, thereby reducing the amount of computation.

[0082] For example, let's say we have objects a, b, and c, where b is an adjacent object of a. When object a is a candidate object, we can get the weight of the first object of a as 2 and the weight of the second object of b as 2.5. If we continue to filter candidates from each object, object b might be selected as a candidate object. In this case, since object a is an adjacent object of object b, the weight of the first object of b would also be 2.5, and the weight of the second object of a would also be 2, resulting in double calculation.

[0083] If objects a and b are marked, then only object c will be considered as a candidate object, avoiding duplicate calculations and thus reducing the amount of computation.

[0084] After obtaining the weights of the first and second objects, the terminal updates the association weights between objects in the first adjacent object group based on the weights of the first and second objects, thus obtaining the updated association weights. This allows the updated association weights to more accurately represent the relationship between candidate objects and their adjacent objects, improving the accuracy of subsequent segmentation.

[0085] In other embodiments, the association weights between objects in the first adjacent object group are updated based on the first object weight and the second object weight to obtain the updated association weights, including: The weights of the first object and the weights of the second object are fused to obtain the weight of the target object. The target object weight and the association weights between objects in the first adjacent object group are updated and fused to obtain the updated association weights between objects in the first adjacent object group.

[0086] The first fusion process can be an additive fusion process, or it can include both additive and multiplicative fusion processes. That is, the first object weight and the first coefficient can be multiplicatively fused to obtain the first fusion result, the second object weight and the second coefficient can be multiplicatively fused to obtain the second fusion result, and then the first fusion result and the second fusion result can be additively fused to obtain the target object weight.

[0087] The terminal can input the weight of the first object, the weight of the second object, and the association weights between objects in the first adjacent object group into the following formula for fusion processing to obtain the updated association weights:

[0088] W(user i ) Indicates the first i The candidate and the first i The first candidate j Updated association weights between adjacent objects W(user j ) Indicates the first i The first candidate j The second object weight of the nth neighboring object (when the nth) i The first candidate j When two adjacent objects are used as candidate objects W(user j ) It can also represent the first i The first candidate j The first object weight of each of the adjacent objects.

[0089] For example, such as Figure 4 As shown, the first object weight of object a is 2.5, the first object weight of object b is 2 (when object a is a candidate object, 2 is also the second object weight of object b), the first object weight of object d is 2, and the first object weight of object c is 1. Therefore, the updated association weight between object a and object b is 12.5.

[0090] In this embodiment, a first fusion process is performed on the first object weight and the second object weight to obtain the target object weight. Then, an update fusion process is performed on the target object weight and the association weight between objects in the first adjacent object group to obtain the updated association weight. This ensures that the larger the first object weight and the second object weight are, the larger the updated association weight is, and the smaller the first object weight and the second object weight are, the smaller the updated association weight is. By using the first object weight and the second object weight, the relationship between the candidate object and the candidate object's adjacent objects becomes more accurate, thereby improving the accuracy of the segmentation.

[0091] S205. Divide each object into groups based on the updated association weights.

[0092] After receiving the updated association weights, the terminal can group objects with updated association weights greater than the first preset weight into the same object group.

[0093] For example, such as Figure 4As shown, the updated association weight between object a and object b is 12.5, and the updated association weight between object a and object d is 9. If the first preset weight is 10, then object a and object b will be assigned to object group C1.

[0094] At this point, object d has not been assigned to object group C1, meaning a situation arises where it is necessary to separate objects from object groups. Therefore, in some embodiments, after dividing each object according to the updated association weight to obtain object groups, the process further includes: From the various objects, filter out the objects to be divided that are not in the object group; The group association weight between the object to be divided and the object group is determined based on the updated association weight between the object to be divided and each object in the object group. The objects to be divided are assigned to the object group corresponding to the largest group association weight.

[0095] The terminal can directly add the updated association weights between the object to be divided and each object in the object group to obtain the group association weight between the object to be divided and the object group.

[0096] For example, such as Figure 4 As shown, objects a and b are in object group C1, while object d is not in object group C1. Object d is the object to be partitioned. Since the updated association weight between object d and object a in object group C1 is 9, and the updated association weight between object d and object b in object group C1 is 8, the group association weight between object d and object group C1 is 17.

[0097] If at this time object d still has an updated association weight with objects in object group C2 and the group association weight between object d and object group C2 is 10, and object d has an updated association weight with objects in object group C3 and the group association weight between object d and object group C3 is 9, since the maximum group association weight is 17, object d is assigned to object group C1.

[0098] After obtaining the various object groups, if the number of objects in an object group is small, the object groups may become isolated. Therefore, in some embodiments, after dividing the objects according to the updated association weights to obtain object groups, the process further includes: Filter the target object group from the object group, where the number of objects in the target object group is less than the preset number; Determine each adjacent object group of the target object group. The adjacent object group is the group containing the associated objects that have updated association weights with objects in the target object group. Based on the updated association weights between objects in the target object group and associated objects, determine the group association weights between the target object group and adjacent object groups; Update the target object group based on the group association weight.

[0099] The terminal can directly add up the updated association weights between objects in the target object group and related objects to obtain the group association weight between the target object group and adjacent object groups.

[0100] For example, there are object groups C1, C2, and C3. Object group C1 contains 99 objects, object group C2 contains 150 objects, and the total number of objects in object group C3 is 200. The default number is 100. There is an updated association weight between object a in object group C1 and objects c and d in object group C2, and there is an updated association weight between object b in object group C1 and object e in object group C3. Then, the target object group is object group C1, object groups C2 and C3 are both adjacent object groups of object group C1, and objects c, d, and e are associated objects.

[0101] If the updated association weight between object a and object c is 10, the updated association weight between object a and object d is 12, and the updated association weight between object b and object e is 8, then the group association weight between object group C1 and object group C2 is 22, and the group association weight between object group C1 and object group C3 is 8.

[0102] After obtaining the group association weight, the terminal can directly assign the target object group to the adjacent object group corresponding to the largest group association weight.

[0103] Alternatively, the terminal can update the object groups by referring to the object partitioning method. That is, the terminal determines the first group object weight of the target object group based on the group association weight between the target object group and the adjacent object groups, and determines the second group object weight of the adjacent object groups based on the group association weight between the adjacent object groups and the indirectly adjacent object groups. Then, the terminal updates the group association weight between the target object group and the adjacent object groups based on the first group object weight and the second group object weight to obtain the updated group association weight. Finally, the terminal assigns the target object group to the adjacent object group corresponding to the updated group association weight that is greater than the second preset weight.

[0104] In this embodiment, after obtaining each object group, the object groups are further subdivided to avoid the phenomenon of isolated object groups. Here, an object group can also be referred to as a community.

[0105] In this embodiment, object groups are constructed using product association weights and evaluation association weights, eliminating the need to construct object groups based on friend relationships. Furthermore, the objects can be anonymized, resulting in anonymous object groups that enhance the security of protected objects.

[0106] S206. Based on the products of other objects within the object group, recommend products to the target object. Other objects are objects within the object group other than the target object.

[0107] Other objects' products refer to products that are related to other objects. For example, if other objects purchase product 1, then product 1 is related to other objects. Or, if other objects view video 1, then video 1 is related to other objects.

[0108] As described above, in this embodiment, the product association weights of multiple objects for various different products and the evaluation association weights of multiple objects for the same product among various different types of products are first obtained. Then, based on the product association weights and evaluation association weights, the association weights between each object are determined, and adjacent objects are identified based on these association weights, resulting in adjacent object groups. Next, the association weights between each object are updated based on the association weights between objects in the adjacent object groups, resulting in updated association weights. Then, each object is divided based on the updated association weights, resulting in object groups. Finally, based on the products of other objects within the object groups, product recommendations are made for the target object; these other objects are all objects within the object groups other than the target object.

[0109] In this embodiment, the association weights are first determined based on the evaluation association weights and product association weights between each object. Then, adjacent objects are determined based on the association weights to obtain adjacent object groups. Next, the association weights between each object are updated based on the association weights between objects in the adjacent object groups to obtain updated association weights. This updated association weights are used to divide each object into more accurate object groups. As a result, when recommending products to the target object based on the products of other objects within the object group, the recommended products are more in line with the needs of the target object, thus improving the accuracy of product recommendations.

[0110] Based on the methods described in the above embodiments, the following examples will provide further detailed explanations.

[0111] Please see Figure 5 , Figure 5 This is a flowchart illustrating the product recommendation method provided in an embodiment of this application. The product recommendation method may include: S501. The terminal obtains the product association weights of multiple objects for various different purchased products of the same type in the platform, and determines the type of evaluation information based on the evaluation information of multiple objects for the same purchased product in the platform.

[0112] The platform can be a shopping platform. The platform can take the form of at least one of a mini-program, an application, or a webpage. The terminal can filter out multiple different purchased items of the same type from a variety of purchased items on the platform, and then determine the product association weight between the objects corresponding to these multiple different purchased items of the same type as the first target value.

[0113] The same group of people who have purchased multiple different products of the same type can be referred to as the same type of people.

[0114] S502. When the evaluation information of each object for the same purchased product is of the same type, the terminal determines the evaluation association weight between the objects corresponding to the same evaluation information type as the third target value.

[0115] People who share the same type of evaluation information can be called a group of like-minded individuals.

[0116] S503. When the types of evaluation information for the same purchased product are different for each object, the terminal determines the evaluation association weight between objects with the same type of evaluation information as the fourth target value, and the third target value is greater than the fourth target value.

[0117] Different types of evaluation information correspond to different groups of people and can be called related groups.

[0118] S504. The terminal determines the association weight between each object based on the product association weight and the evaluation association weight, and constructs the first relationship network graph with the association weight as the connecting edge and the object as the node.

[0119] S505. The terminal determines adjacent objects based on the first relational network diagram, obtains adjacent object groups, and filters out candidate objects from each object. It then filters out a first adjacent object group containing candidate objects and a second adjacent object group containing adjacent objects from each adjacent object group.

[0120] S506. The terminal determines the number of the first adjacent object group and determines the first object weight of the candidate object based on the number and the association weight between objects in the first adjacent object group.

[0121] S507. The terminal determines the number of the second adjacent object groups, and determines the second object weight of the adjacent objects of the candidate object based on the number of the two adjacent object groups and the association weight between the objects in the second adjacent object groups.

[0122] S508. The terminal performs a first fusion process on the weight of the first object and the weight of the second object to obtain the weight of the target object, and performs an update fusion process on the weight of the target object and the association weight between objects in the first adjacent object group to obtain the updated association weight between objects in the first adjacent object group.

[0123] S509. The terminal will group objects with updated associated weights that are greater than the first preset weight into the same object group.

[0124] S5010. The terminal filters out the objects to be divided from the objects that are not in the object group, and determines the group association weight between the objects to be divided and the object group based on the updated association weight between the objects to be divided and each object in the object group.

[0125] S5011, The terminal assigns the objects to be divided to the object group corresponding to the largest group association weight.

[0126] S5012. The terminal filters out the target object group from the object group. The number of objects in the target object group is less than the preset number. The terminal also determines each adjacent object group and indirect adjacent object group of the target object group. The adjacent object group is the group where the associated objects with updated association weights with the objects in the target object group are located. The indirect adjacent object group is the group where the indirect associated objects with updated association weights with the associated objects are located.

[0127] S5013. The terminal determines the group association weight between the target object group and the adjacent object group based on the updated association weight between the object and the associated object in the target object group, and determines the group association weight between the adjacent object group and the indirectly adjacent object group based on the updated association weight between the associated object and the indirectly associated object.

[0128] S5014. The terminal determines the first group object weight of the target object group based on the group association weight between the target object group and the adjacent object group, and determines the second group object weight of the adjacent object group based on the group association weight between the adjacent object group and the indirectly adjacent object group.

[0129] S5015. The terminal updates the group association weight between the target object group and the adjacent object group according to the weight of the first group object and the weight of the second group object, obtains the updated group association weight, and divides the target object group into the adjacent object group corresponding to the updated group association weight that is greater than the second preset weight, thus obtaining the first relationship network graph after division.

[0130] S5016. The terminal obtains the recommendation request of the target object, and recommends the products purchased by other objects in the corresponding object group in the first relational network graph after division to the target object.

[0131] After obtaining the first relational network graph after partitioning, when the terminal receives a recommendation request from a target object, it can recommend products purchased by other objects in the corresponding object group within the partitioned first relational network graph to the target object. Furthermore, it can sort and display the products purchased by other objects based on the updated association weights between the target object and other objects. For example... Figure 6 As shown in section 601, the higher the association weight after the update, the earlier the order in which other objects purchase products.

[0132] Furthermore, the terminal can also obtain the friend relationships between users based on the mini-programs they share, and generate a friend list based on these relationships. The generated friend list can be, for example, like... Figure 6 As shown in Figure 602.

[0133] It should be noted that if an object is a new object on the platform, meaning that the object has not yet purchased any products on the platform, then when the terminal detects that an object is a new object, it can display an interest set page. The interest set page includes at least one of the following interest set options: category set, brand set, and product set. Figure 7 As shown. The category options can include apparel, personal care, sports, and jewelry. The brand options can include brand A, brand B, and brand C.

[0134] Then, in response to the new object's selection of the interest set, the terminal constructs the first feature vector of the new object. The terminal also constructs the second feature vector of the historical object based on the platform's historical object's access records of products, evaluation information of products, and like information of comments.

[0135] Next, the terminal calculates the first feature vector and the second feature vector, calculates the similarity between the new object and the historical objects, and forms a friend list of the new object with historical objects whose similarity is greater than the preset similarity. It can also recommend products purchased by historical objects whose similarity is greater than the preset similarity to the new object.

[0136] It should be noted that if the new object does not select an interest set, a popular friend set can be obtained, and a friend list of the new object can be selected from the popular friend set through a random interleaving algorithm.

[0137] In some embodiments, in addition to constructing a first relationship network graph based on product association weights and evaluation association weights, a second relationship network graph can also be constructed based on the object's mailing address. The process of constructing the second relationship network graph based on the object's mailing address can be as follows: The terminal first converts the recipient's address into latitude and longitude, then encodes the latitude and longitude of the object using the geohash algorithm to obtain the object's string. Then, based on the precision of the geohash algorithm, the object's string is divided into the same interval (for example, when the geohash algorithm's precision is 1 kilometer, it can generate friends within 1 kilometer). Objects corresponding to strings within the same interval are considered friends, thus obtaining the surrounding population. Finally, a second relationship network graph is generated based on the objects corresponding to the strings in each interval.

[0138] For example, such as Figure 8 As shown, the terminal can generate a first relationship network diagram based on evaluation information, or it can generate a second relationship network diagram based on the recipient address.

[0139] For details on the specific implementation method and corresponding beneficial effects in this example, please refer to the above-mentioned product recommended method implementation examples. This example will not repeat them here.

[0140] To facilitate better implementation of the product recommendation method provided in this application, this application also provides an apparatus based on the above-described product recommendation method. The meanings of the terms used are the same as in the product recommendation method described above, and specific implementation details can be found in the descriptions within the method embodiments.

[0141] For example, such as Figure 9 As shown, the product recommendation device may include: The first acquisition module 901 is used to acquire the product association weights of multiple objects for various different products.

[0142] The second acquisition module 902 is used to acquire the evaluation association weights of multiple objects for the same product in multiple different types of products.

[0143] The weight determination module 903 is used to determine the association weight between each object based on the product association weight and the evaluation association weight, and to determine the adjacent objects based on the association weight to obtain the adjacent object group.

[0144] The weight update module 904 is used to update the association weights between objects based on the association weights between objects in adjacent object groups, and obtain the updated association weights.

[0145] The object partitioning module 905 is used to partition each object according to the updated association weight to obtain object groups.

[0146] Product recommendation module 906 is used to recommend products to a target object based on products from other objects within the object group. Other objects are objects in the object group other than the target object.

[0147] Optionally, the weight update module 904 is specifically used to perform: Filter out candidate objects from each object, and filter out the first adjacent object group containing the candidate objects and the second adjacent object group containing the adjacent objects of the candidate objects from each adjacent object group; Based on the first adjacent object group, determine the first object weight of the candidate object; Based on the second adjacent object group, determine the second object weight of the adjacent objects of the candidate object; Based on the weights of the first and second objects, the association weights between objects in the first adjacent object group are updated to obtain the updated association weights.

[0148] Optionally, the weight update module 904 is specifically used to perform: Determine the number of the first adjacent object group; The first object weight of the candidate object is determined based on the number of objects and the association weight between objects in the first adjacent object group.

[0149] Optionally, the weight update module 904 is specifically used to perform: The weights of the first object and the weights of the second object are fused to obtain the weight of the target object. The target object weight and the association weights between objects in the first adjacent object group are updated and fused to obtain the updated association weights between objects in the first adjacent object group.

[0150] Optionally, the second acquisition module 902 is specifically used to perform: Retrieve evaluation information of the same product from multiple objects across various different product categories; Determine the type of evaluation information; Based on the type of evaluation information, determine the evaluation association weights of multiple objects for the same product across multiple different products.

[0151] Optionally, the object partitioning module 905 is also used to perform: From the various objects, filter out the objects to be divided that are not in the object group; The group association weight between the object to be divided and the object group is determined based on the updated association weight between the object to be divided and each object in the object group. The objects to be divided are assigned to the object group corresponding to the largest group association weight.

[0152] Optionally, the object partitioning module 905 is also used to perform: Filter the target object group from the object group, where the number of objects in the target object group is less than the preset number; Determine each adjacent object group of the target object group. The adjacent object group is the group containing the associated objects that have updated association weights with objects in the target object group. Based on the updated association weights between objects in the target object group and associated objects, determine the group association weights between the target object group and adjacent object groups; Update the target object group based on the group association weight.

[0153] In practice, each of the above modules can be implemented as an independent entity or can be combined arbitrarily to be implemented as the same or several entities. For the specific implementation methods and corresponding beneficial effects of each of the above modules, please refer to the previous method embodiments, which will not be repeated here.

[0154] This application also provides an electronic device, which may be a server or a terminal, etc. Figure 10 As shown, it illustrates a structural schematic diagram of the electronic device involved in the embodiments of this application, specifically: The electronic device may include components such as a processor 1001 with one or more processing cores, a memory 1002 with one or more computer-readable storage media, a power supply 1003, and an input unit 1004. Those skilled in the art will understand that... Figure 10 The electronic device structure shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, or combine certain components, or have different component arrangements. Wherein: The processor 1001 is the control center of the electronic device, connecting various parts of the device via various interfaces and lines. It executes computer programs and / or modules stored in the memory 1002, and calls data stored in the memory 1002, to perform various functions and process data. Optionally, the processor 1001 may include one or more processing cores; preferably, the processor 1001 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and applications, and the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into the processor 1001.

[0155] The memory 1002 can be used to store computer programs and modules. The processor 1001 executes various functional applications and data processing by running the computer programs and modules stored in the memory 1002. The memory 1002 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, computer programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 1002 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 1002 may also include a memory controller to provide the processor 1001 with access to the memory 1002.

[0156] The electronic device also includes a power supply 1003 that supplies power to various components. Preferably, the power supply 1003 can be logically connected to the processor 1001 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. The power supply 1003 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0157] The electronic device may also include an input unit 1004, which can be used to receive input digital or character information, and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control.

[0158] Although not shown, the electronic device may also include a display unit, etc., which will not be described in detail here. Specifically, in this embodiment, the processor 1001 in the electronic device loads the executable files corresponding to the processes of one or more computer programs into the memory 1002 according to the following instructions, and the processor 1001 runs the computer programs stored in the memory 1002 to realize various functions, such as: Obtain the product association weights of multiple objects for various different products; Obtain the evaluation association weights of multiple objects for the same product across multiple different products; Based on the product association weight and the evaluation association weight, the association weight between each object is determined, and the adjacent objects are determined based on the association weight to obtain the adjacent object group; Based on the association weights between objects in adjacent object groups, the association weights between each object are updated to obtain the updated association weights. Based on the updated association weights, each object is divided into groups to obtain object groups; Based on the products of other objects within the object group, product recommendations are made to the target object. Other objects are objects within the object group other than the target object.

[0159] For details on the specific implementation methods and corresponding beneficial effects of each of the above operations, please refer to the detailed description of the product recommendation method above, which will not be repeated here.

[0160] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by a computer program, or by a computer program controlling related hardware. The computer program can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0161] Therefore, embodiments of this application provide a computer-readable storage medium storing a computer program that can be loaded by a processor to execute the steps in any of the product recommendation methods provided in embodiments of this application. For example, the computer program can execute the following steps: Obtain the product association weights of multiple objects for various different products; Obtain the evaluation association weights of multiple objects for the same product across multiple different products; Based on the product association weight and the evaluation association weight, the association weight between each object is determined, and the adjacent objects are determined based on the association weight to obtain the adjacent object group; Based on the association weights between objects in adjacent object groups, the association weights between each object are updated to obtain the updated association weights. Based on the updated association weights, each object is divided into groups to obtain object groups; Based on the products of other objects within the object group, product recommendations are made to the target object. Other objects are objects within the object group other than the target object.

[0162] For details on the specific implementation methods and corresponding beneficial effects of the above operations, please refer to the previous embodiments, which will not be repeated here.

[0163] The computer-readable storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0164] Since the computer program stored in the computer-readable storage medium can execute the steps in any of the product recommendation methods provided in the embodiments of this application, the beneficial effects that any of the product recommendation methods provided in the embodiments of this application can achieve can be realized, as detailed in the preceding embodiments, and will not be repeated here.

[0165] According to one aspect of this application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method described above.

[0166] The foregoing has provided a detailed description of a product recommendation method, apparatus, electronic device, storage medium, and program product provided in the embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A product recommendation method characterized by, include: Obtain the product association weights of multiple objects for various different products; Obtain the evaluation association weights of the multiple objects for the same product across multiple different products; Based on the product association weight and the evaluation association weight, the association weight between each of the objects is determined, and adjacent objects are determined based on the association weight to obtain an adjacent object group; The association weights between objects in the adjacent object groups are updated based on their association weights to obtain updated association weights. This process includes: selecting candidate objects from each object; selecting a first adjacent object group containing the candidate objects and a second adjacent object group containing the adjacent objects of the candidate objects from each adjacent object group; determining a first object weight for the candidate objects based on the first adjacent object group; determining a second object weight for the adjacent objects of the candidate objects based on the second adjacent object group; and updating the association weights between objects in the first adjacent object group based on the first object weights and the second object weights to obtain updated association weights. The objects are divided according to the updated association weights to obtain object groups; Based on the products of other objects within the object group, product recommendations are made to the target object, where the other objects are objects within the object group other than the target object.

2. The product recommendation method according to claim 1, characterized in that, The step of determining the first object weight of the candidate object based on the first adjacent object group includes: Determine the number of the first adjacent object group; The first object weight of the candidate object is determined based on the number of objects and the association weight between objects in the first adjacent object group.

3. The product recommendation method according to claim 1, characterized in that, The step of updating the association weights between objects in the first adjacent object group based on the first object weight and the second object weight to obtain the updated association weights includes: The first object weight and the second object weight are fused together to obtain the target object weight. The target object weight and the association weight between objects in the first adjacent object group are updated and fused to obtain the updated association weight between objects in the first adjacent object group.

4. The product recommendation method according to claim 1, characterized in that, The step of obtaining the evaluation association weights of the multiple objects for the same product among multiple different products includes: Obtain evaluation information from the multiple objects for the same product across multiple different products; Determine the type of the evaluation information; Based on the type of evaluation information, determine the evaluation association weights of the multiple objects for the same product among multiple different products.

5. The product recommendation method according to any one of claims 1-4, characterized in that, After dividing the objects according to the updated association weights to obtain object groups, the method further includes: From each of the objects, filter out the objects to be divided that are not in the object group; The group association weight between the object to be divided and the object group is determined based on the updated association weight between the object to be divided and each object in the object group. The objects to be divided are assigned to the object group corresponding to the largest group association weight.

6. The product recommendation method according to any one of claims 1-4, characterized in that, After dividing the objects according to the updated association weights to obtain object groups, the method further includes: Filter out a target object group from the object group, wherein the number of objects in the target object group is less than a preset number; Determine each neighboring object group of the target object group, wherein the neighboring object group is the group containing the associated objects that have the updated association weight with the objects in the target object group; Based on the updated association weights between objects in the target object group and the associated objects, determine the group association weights between the target object group and the adjacent object groups; The target object group is updated based on the group association weight.

7. A product recommendation device, characterized in that, include: The first acquisition module is used to acquire the product association weights of multiple objects for various different products; The second acquisition module is used to acquire the evaluation association weights of the multiple objects for the same product among multiple different types of products; The weight determination module is used to determine the association weight between each of the objects based on the product association weight and the evaluation association weight, and to determine adjacent objects based on the association weight to obtain an adjacent object group; The weight update module is used to update the association weights between each object according to the association weights between objects in the adjacent object group to obtain updated association weights. The module includes: filtering candidate objects from each object; filtering a first adjacent object group containing the candidate objects and a second adjacent object group containing the adjacent objects of the candidate objects from each adjacent object group; determining a first object weight for the candidate objects based on the first adjacent object group; determining a second object weight for the adjacent objects of the candidate objects based on the second adjacent object group; and updating the association weights between objects in the first adjacent object group based on the first object weights and the second object weights to obtain updated association weights. An object partitioning module is used to partition each of the objects according to the updated association weights to obtain object groups; The product recommendation module is used to recommend products to the target object based on the products of other objects within the object group, wherein the other objects are objects within the object group other than the target object.

8. The product recommendation device according to claim 7, characterized in that, The weight update module is specifically used to perform: Determine the number of the first adjacent object groups mentioned above; Based on the aforementioned number and the association weights between objects in the aforementioned first adjacent object group, the first object weight of the aforementioned candidate object is determined.

9. The product recommendation device according to claim 7, characterized in that, The weight update module is specifically used to perform: The first object weight and the second object weight are fused together to obtain the target object weight. The weights of the target object and the association weights between objects in the first adjacent object group are updated and merged to obtain the updated association weights between objects in the first adjacent object group.

10. The product recommendation device according to claim 7, characterized in that, The second acquisition module is specifically used for execution: Obtain evaluation information from multiple objects regarding the same product across various different products; Determine the type of the above evaluation information; Based on the types of evaluation information mentioned above, determine the evaluation association weights of the above multiple objects for the same product among multiple different products.

11. The product recommendation device according to any one of claims 7 to 10, characterized in that, The object partitioning module is also used to perform: From the above objects, filter out the objects to be divided that are not in the above object groups; Based on the updated association weights between the objects to be segmented and each object in the object group, determine the group association weights between the objects to be segmented and the object group. The objects to be divided are assigned to the object group corresponding to the largest group association weight.

12. The product recommendation device according to any one of claims 7 to 10, characterized in that, The object partitioning module is also used to perform: Filter out the target object group from the above object group, where the number of objects in the target object group is less than the preset number; Determine each of the neighboring object groups of the target object group, wherein the neighboring object groups are the groups of related objects that have the updated association weights with objects in the target object group; Based on the updated association weights between objects in the target object group and the associated objects, determine the group association weights between the target object group and the adjacent object groups. The target object group is updated based on the aforementioned group association weights.

13. An electronic device, characterized in that, It includes a processor and a memory, the memory storing a computer program, and the processor running the computer program in the memory to perform the product recommendation method according to any one of claims 1 to 6.

14. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program adapted for loading by a processor to perform the product recommendation method according to any one of claims 1 to 6.

15. A computer program product, characterized in that, The computer program product stores a computer program adapted for loading by a processor to execute the product recommendation method according to any one of claims 1 to 6.

Images