Object recommendation method, device, electronic device and storage medium

Abstract

Embodiments of the present invention relate to the field of artificial intelligence and disclose an object recommendation method, apparatus, electronic device, and storage medium. The method comprises: obtaining an object recommendation request for an object to be recommended, sent by a user; wherein the object recommendation request includes object-related data to be recommended and the user's multidimensional recommendation-related factors; generating a multidimensional co-occurrence matrix matching the object recommendation request based on the user's multidimensional recommendation-related factors and the object recommendation-related data; generating an object similarity matrix matching the object recommendation request based on the object-related data to be recommended and the object recommendation-related data; generating a multidimensional relationship matrix matching the object recommendation request based on the multidimensional co-occurrence matrix and the object similarity matrix; and determining the recommended object matching the object recommendation request based on the multidimensional relationship matrix. The technical solution of the embodiments of the present invention can improve the accuracy of object recommendations.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of intelligent recommendation technology, and in particular to an object recommendation method, device, electronic device, and storage medium. Background Art

[0002] Intelligent recommendation can provide personalized recommendation services based on big data and artificial intelligence technology, combined with the accumulation of multiple industry fields.

[0003] Intelligent recommendation is a key area within the field of artificial intelligence and is crucial for businesses, especially those in the retail sector. Intelligent recommendation systems are widely used in today's internet companies. They can quickly present recommendations to customers based on their preferences, habits, and browsing history. Recommendations can include products, information, audio and video data, effectively retaining users' browsing time and extending their stay.

[0004] In the process of realizing the present invention, the inventors discovered that the prior art has the following defects: In traditional intelligent recommendation systems, when recommending objects, they often only perform simple screening logic processing on the objects to be recommended, and display the objects to be recommended to users in the form of a waterfall flow according to the release time, object name or object classification. Generally speaking, if the number of objects that can be recommended is limited and there are not many objects to display, then this display method is not a big problem and will not cause a bad user experience for users. However, with the development of the times, the amount and types of Internet information are becoming more and more numerous. Traditional object recommendation methods will recommend more objects to users, but a large part of these recommended objects are not objects that users are interested in, which can easily cause a poor experience for users and reduce user stay time. Summary of the Invention

[0005] Embodiments of the present invention provide an object recommendation method, device, electronic device, and storage medium, which can improve the accuracy of object recommendation and thus enhance user experience.

[0006] In a first aspect, an embodiment of the present invention provides an object recommendation method, comprising:

[0007] Obtaining an object recommendation request sent by a user; wherein the object recommendation request includes object-related data of the object to be recommended and the multi-dimensional recommendation-related factors of the user;

[0008] generating a multidimensional co-occurrence matrix matching the object recommendation request according to the multidimensional recommendation association factors and the object recommendation association data;

[0009] generating an object similarity matrix matching the object recommendation request according to the to-be-recommended object association data and the object recommendation association data;

[0010] generating a multidimensional relationship matrix matching the object recommendation request according to the multidimensional co-occurrence matrix and the object similarity matrix;

[0011] Determine a recommended object that matches the object recommendation request according to the multi-dimensional relationship matrix.

[0012] In a second aspect, an embodiment of the present invention further provides an object recommendation device, comprising:

[0013] An object recommendation request acquisition module is configured to acquire an object recommendation request sent by a user; wherein the object recommendation request includes object-related data of the object to be recommended and the multi-dimensional recommendation-related factors of the user;

[0014] A multi-dimensional co-occurrence matrix generating module, configured to generate a multi-dimensional co-occurrence matrix matching the object recommendation request based on the user's multi-dimensional recommendation association factors and the object recommendation association data;

[0015] An object similarity matrix generating module, configured to generate an object similarity matrix matching the object recommendation request based on the to-be-recommended object association data and the object recommendation association data;

[0016] a multi-dimensional relationship matrix generating module, configured to generate a multi-dimensional relationship matrix matching the object recommendation request based on the multi-dimensional co-occurrence matrix and the object similarity matrix;

[0017] The recommended object determination module is used to determine the recommended object that matches the object recommendation request according to the multi-dimensional relationship matrix.

[0018] In a third aspect, an embodiment of the present invention further provides an electronic device, comprising:

[0019] one or more processors;

[0020] a storage device for storing one or more programs;

[0021] When the one or more programs are executed by the one or more processors, the one or more processors implement the object recommendation method provided by any embodiment of the present invention.

[0022] In a fourth aspect, an embodiment of the present invention further provides a computer storage medium having a computer program stored thereon, which, when executed by a processor, implements the object recommendation method provided by any embodiment of the present invention.

[0023] The embodiment of the present invention generates a multidimensional co-occurrence matrix for matching object recommendation requests based on the user's multidimensional recommendation association factors and object recommendation association data included in the object recommendation request for the object to be recommended sent by the user, and generates an object similarity matrix for matching object recommendation requests based on the object association data to be recommended and the object recommendation association data included in the object recommendation request, so as to generate a multidimensional relationship matrix for matching object recommendation requests based on the multidimensional co-occurrence matrix and the object similarity matrix, and finally determines the recommended object matching the object recommendation request based on the generated multidimensional relationship matrix, thereby solving the problem of low recommendation accuracy existing in existing object recommendation methods, and can improve the accuracy of object recommendation, thereby improving user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a flowchart of an object recommendation method provided by Example 1 of the present invention;

[0025] Figure 2 This is a flowchart of an object recommendation method provided by Embodiment 2 of the present invention;

[0026] Figure 3 is a schematic diagram of an object recommendation device provided by Embodiment 3 of the present invention;

[0027] Figure 4 This is a structural diagram of an electronic device provided in Example 4 of the present invention. DETAILED DESCRIPTION

[0028] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention.

[0029] It should also be noted that, for ease of description, only the part relevant to the present invention, rather than all of the content, is shown in the accompanying drawings. Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processing or methods depicted as flow charts. Although flow charts describe various operations (or steps) as sequential processing, many operations therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of various operations can be rearranged. When its operation is completed, the processing can be terminated, but can also have additional steps not included in the accompanying drawings. The processing can correspond to methods, functions, procedures, subroutines, subprograms, etc.

[0030] The terms "first," "second," and so on, in the description, claims, and drawings of the embodiments of the present invention are used to distinguish between different items, not to describe a specific order. Furthermore, the terms "including," "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or elements is not limited to the listed steps or elements and may include steps or elements that are not listed.

[0031] Example 1

[0032] Figure 1 This is a flowchart of an object recommendation method provided in the first embodiment of the present invention. This embodiment is applicable to the case where object recommendation is performed based on a multidimensional relationship matrix generated by a multidimensional co-occurrence matrix and an object similarity matrix. The method can be executed by an object recommendation device, which can be implemented by software and / or hardware and can generally be integrated into an electronic device. The electronic device can be a terminal device or a server device. The electronic device can run an object recommendation system to execute the object recommendation method. The embodiment of the present invention does not limit the specific device type of the electronic device. Optionally, the server can be an independent server 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 communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0033] Embodiments of the present invention can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results.

[0034] Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0035] Correspondingly, such as Figure 1 As shown, the method includes the following operations:

[0036] S110. Obtain an object recommendation request sent by a user; wherein the object recommendation request includes object-related data of the object to be recommended and multi-dimensional recommendation-related factors of the user.

[0037] The recommended object may be an object requested by the user. Optionally, the recommended objects that can be recommended to the user may include multiple types, such as, but not limited to, commodities, information, or audio and video data. Any object that can be recommended to the user may be a recommended object, and the embodiments of the present invention do not limit the object type of the recommended object. The object recommendation request may be a request initiated by the user and may represent the user's object recommendation needs. The recommended object associated data may be attribute data related to the recommended object. For example, when the recommended object is a device, the recommended object associated data may be descriptive data of the device, such as the device name, model, or performance data. When the recommended object is audio and video data, the recommended object associated data may be descriptive data of the audio and video data, such as the audio and video title, release time, or data size. Multi-dimensional recommendation association factors may be reference factors used to determine the user's object recommendation needs. For example, when the recommended object is a device, the multi-dimensional recommendation association factors may include, but are not limited to, the user's industry identification, the user's region, and the utilization rate of the user's device. When the object to be recommended is audio or video data to be recommended, the multi-dimensional recommendation related factors may include but are not limited to the user's historical browsing habits and the user's interest preferences.

[0038] In an embodiment of the present invention, an object recommendation system can respond to a user's object recommendation request and accurately recommend objects to the user using locally stored information about candidate recommended objects (i.e., objects that can be recommended). For example, the object recommendation system can be a product recommendation system or an information recommendation system. The embodiment of the present invention does not limit the system type of the object recommendation system or the types of objects that can be recommended.

[0039] In an embodiment of the present invention, the object recommendation system can obtain the object recommendation request for the object to be recommended sent by the user in a variety of ways. Optionally, the user can actively send an object recommendation request for the object to be recommended to the object recommendation system. For example, if the information of the product to be purchased is entered in the product browsing page, the object recommendation system background can obtain the product recommendation request for the product to be recommended actively sent by the user. Optionally, the object recommendation system can also actively obtain the object recommendation request for the object to be recommended without the user's perception. For example, if the user randomly browses audio and video through browsing operations (such as swiping up or swiping down, etc.) in the audio and video browsing interface, the object recommendation system background can use the user's browsing operations in the current audio and video browsing interface as the object recommendation request for the object to be recommended sent by the user.

[0040] It should be noted that the object recommendation request for an object to be recommended may include, but is not limited to, object-related data and the user's multi-dimensional recommendation-related factors. The object-related data may be used to describe the user's recommendation requirements for the recommended object, and the user's multi-dimensional recommendation-related factors may be determined from the user's multi-dimensional perspective and may be used as a reference for relevant factors in making object recommendations to the user.

[0041] S120 : Generate a multi-dimensional co-occurrence matrix matching the object recommendation request according to the multi-dimensional recommendation association factors and the object recommendation association data.

[0042] Among them, the object recommendation associated data can be related data stored by the object recommendation system for object recommendation. Exemplarily, the object recommendation associated data can include but is not limited to the object name, type, identifier and attribute description of the candidate recommended object in the object recommendation system. As long as it can be used as a reference for the object recommendation system to recommend objects, the embodiment of the present invention does not limit the specific data content of the object recommendation associated data. The multi-dimensional co-occurrence matrix can be a plurality of types of co-occurrence matrices generated according to different recommendation association factors. The matrix elements in the co-occurrence matrix can reflect the co-occurrence degree of two objects. It can be understood that the higher the co-occurrence degree, the higher the degree of matching between the two objects.

[0043] Correspondingly, after the object recommendation system obtains the object recommendation request initiated by the user for the recommended object, it can obtain the user's multi-dimensional recommendation association factors based on the object recommendation request, and obtain the locally stored object recommendation association data, so as to generate a multi-dimensional co-occurrence matrix matching the object recommendation request based on the user's multi-dimensional recommendation association factors and the object recommendation association data.

[0044] It is understandable that the multi-dimensional co-occurrence matrix can analyze the co-occurrence between the object to be recommended and the alternative recommended objects in the object recommendation system from the perspective of multiple reference factors, thereby improving the accuracy of the matching between the object to be recommended and the alternative recommended objects.

[0045] S130: Generate an object similarity matrix that matches the object recommendation request according to the to-be-recommended object association data and the object recommendation association data.

[0046] The object similarity matrix can represent the degree of similarity between the recommended object requested by the user and the candidate recommended objects in the object recommendation system. The matrix elements in the object similarity matrix can reflect the similarity between the two objects. It can be understood that a higher similarity indicates a higher degree of match between the two objects.

[0047] Correspondingly, after the object recommendation system obtains the object recommendation request initiated by the user for the recommended object, it can also obtain the object recommendation association data to be recommended based on the object recommendation request to determine the user's recommendation requirements for the recommended object, and obtain the locally stored object recommendation association data to generate an object similarity matrix matching the object recommendation request based on the object recommendation association data and the object recommendation association data.

[0048] It should be noted that Figure 1 This is merely a schematic diagram of one implementation method. There is no specific order in which steps S120 and S130 should be performed. That is, step S120 may be performed first, followed by step S130. Alternatively, step S130 may be performed first, followed by step S120. Alternatively, both steps may be performed in parallel, which is not limited in this embodiment of the present invention.

[0049] S140 : Generate a multidimensional relationship matrix that matches the object recommendation request according to the multidimensional co-occurrence matrix and the object similarity matrix.

[0050] S150: Determine a recommended object that matches the object recommendation request according to the multi-dimensional relationship matrix.

[0051] Among them, the dimensional relationship matrix can be multiple different types of relationship matrices determined based on multiple types of co-occurrence matrices and object similarity matrices, which can represent the degree of matching between the recommended object requested by the user and the alternative recommended objects in the object recommendation system.

[0052] In an embodiment of the present invention, after generating a multidimensional co-occurrence matrix and an object similarity matrix based on a user's object recommendation request, a multidimensional relationship matrix matching the object recommendation request can be further generated based on the multidimensional co-occurrence matrix and the object similarity matrix. It is understood that the multidimensional relationship matrix can reflect the degree of match between the object to be recommended and all candidate recommendation objects in the object recommendation system. Therefore, the recommended object matching the object recommendation request can be determined based on the multidimensional relationship matrix finally generated. For example, the candidate recommendation object with the highest degree of match with the object to be recommended can be selected based on the multidimensional relationship matrix as the recommended object matching the user's object recommendation request.

[0053] It can be seen that the multidimensional relationship matrix used to determine the recommended objects that match the object recommendation request is generated by the multidimensional co-occurrence matrix and the object similarity matrix. The multidimensional co-occurrence matrix itself can analyze the co-occurrence between the recommended object and the candidate recommended objects in the object recommendation system from the perspective of multiple reference factors, which can effectively improve the accuracy of the match between the recommended object and the candidate recommended objects, thereby improving the accuracy of the matching degree in the multidimensional co-occurrence matrix. Therefore, using the multidimensional relationship matrix to determine the recommended objects that match the object recommendation request can achieve accurate determination of the recommended objects, thereby improving the accuracy of object recommendations, avoiding recommending objects that the user is not interested in, and thus improving the user experience.

[0054] The embodiment of the present invention generates a multidimensional co-occurrence matrix for matching object recommendation requests based on the user's multidimensional recommendation association factors and object recommendation association data included in the object recommendation request for the object to be recommended sent by the user, and generates an object similarity matrix for matching object recommendation requests based on the object association data to be recommended and the object recommendation association data included in the object recommendation request, so as to generate a multidimensional relationship matrix for matching object recommendation requests based on the multidimensional co-occurrence matrix and the object similarity matrix, and finally determines the recommended object matching the object recommendation request based on the generated multidimensional relationship matrix, thereby solving the problem of low recommendation accuracy existing in existing object recommendation methods, and can improve the accuracy of object recommendation, thereby improving user experience.

[0055] Example 2

[0056] Figure 2 This is a flowchart of an object recommendation method provided by the second embodiment of the present invention. This embodiment is specific based on the above embodiment. In this embodiment, multiple specific optional methods for generating a multi-dimensional co-occurrence matrix, an object similarity matrix, and a multi-dimensional relationship matrix are provided. At the same time, a specific optional implementation method for determining the recommended object matching the object recommendation request based on the multi-dimensional relationship matrix is ​​also provided. Accordingly, Figure 2 As shown, the method of this embodiment may include:

[0057] S210: Obtain an object recommendation request for an object to be recommended sent by a user.

[0058] The object recommendation request includes the object-to-be-recommended object association data and the user's multi-dimensional recommendation association factors.

[0059] In an optional embodiment of the present invention, the object to be recommended may include a device to be recommended, and the multi-dimensional recommendation association factors include a user demand identifier, a user region identifier, and a user device operating rate.

[0060] Among them, the device to be recommended can request the recommendation of new equipment or second-hand equipment, etc., as long as the recommendation requirements are met. The embodiment of the present invention does not limit the device type of the recommended device. The user demand identifier can be an identifier of the user's recommendation requirements. Exemplarily, the user demand identifier can be an identifier of the user's industry or the user's position identifier, etc. The user region identifier can be an identifier of the user's region. The user device operating rate can be the operating rate of the user's existing equipment. Optionally, a corresponding device monitoring device (such as a device bracelet, etc.) can be installed on the user's existing equipment to monitor the device's operating time each month through the device monitoring device, thereby calculating the operating rate of the user's existing equipment as the user's device operating rate. The user device operating rate can reflect the user's frequency of use of the recommended equipment. The higher the user device operating rate, the more frequently the user uses the equipment, and the higher the user's recommendation demand for the equipment. It should be noted that before installing the corresponding device monitoring device on the user's existing equipment, it is first necessary to obtain authorization permission to install the corresponding device monitoring device on the user's existing equipment. At the same time, user-related information such as equipment start-up time and user equipment start-up rate are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data comply with relevant laws, regulations and standards of the relevant regions.

[0061] It can be seen that when the above-mentioned object recommendation method is applied to the field of equipment recommendation, by taking the user demand identification, user region identification and user equipment operating rate as the user's multi-dimensional recommendation association factors, the user's object recommendation needs can be comprehensively determined from a multi-dimensional perspective, thereby improving the accuracy of determining the user's object recommendation needs.

[0062] S220 : Generate a multi-dimensional co-occurrence matrix for matching the object recommendation request according to the user's multi-dimensional recommendation association factors and the object recommendation association data.

[0063] Accordingly, step S220 may specifically include the following operations:

[0064] S221 : Generate a requirement-object co-occurrence matrix matching the object recommendation request according to the user requirement identifier and the object recommendation association data.

[0065] In an optional embodiment of the present invention, optionally, generating a demand-object co-occurrence matrix matching the object recommendation request based on the user demand identifier and the object recommendation association data may include: calculating the demand co-occurrence degree between the user demand identifier and the object to be recommended based on the object recommendation association data; and constructing a demand-object co-occurrence matrix matching the object recommendation request based on the mapping relationship between the user demand identifier, the object to be recommended and the demand co-occurrence degree.

[0066] Among them, the demand-object co-occurrence matrix can be a co-occurrence matrix constructed from the perspective of user demand, and its matrix elements can be the demand co-occurrence degree between the user demand identifier and the device to be recommended. The demand co-occurrence degree is also the co-occurrence degree between the user demand identifier and the device to be recommended. In an embodiment of the present invention, optionally, a matching calculation can be performed based on the user demand identifier in the user's multi-dimensional recommendation association factors and the object recommendation association data stored in the object recommendation system to obtain the demand co-occurrence degree between the user demand identifier and each device to be recommended, thereby constructing a demand-object co-occurrence matrix that matches the object recommendation request based on the demand co-occurrence degree between the user demand identifier and each device to be recommended.

[0067] For example, assuming that the object recommendation system can recommend three types of devices proposed by the user, and assuming that the user demand identifier is the user industry identifier, the demand object co-occurrence matrix constructed and generated by the object recommendation system can be shown in Table 1:

[0068] Table 1 Requirement object co-occurrence matrix list

[0069] Equipment to be recommended 1 Equipment to be recommended 2 Equipment to be recommended 3 Industry 1 0 0 0.23 Industry 2 0.25 0.11 0 Industry 3 0 0.31 0.16

[0070] S222: Generate a regional object co-occurrence matrix matching the object recommendation request according to the user region identifier and the object recommendation association data.

[0071] In an optional embodiment of the present invention, generating a regional object co-occurrence matrix that matches the object recommendation request based on the user regional identifier and the object recommendation association data may include: calculating the regional co-occurrence degree between the user regional identifier and the object to be recommended based on the object recommendation association data; and constructing a regional object co-occurrence matrix that matches the object recommendation request based on a mapping relationship between the user regional identifier, the object to be recommended, and the regional co-occurrence degree.

[0072] The region-object co-occurrence matrix may be a co-occurrence matrix constructed from the perspective of the user's region, and its matrix elements may be the region co-occurrence degree between the user's region and the device to be recommended. The region co-occurrence degree is also the co-occurrence degree between the user's region and the device to be recommended.

[0073] In an embodiment of the present invention, optionally, a matching calculation can be performed based on the user's regional identifier in the user's multi-dimensional recommendation association factors and the object recommendation association data stored in the object recommendation system to obtain the regional co-occurrence degree between the user's regional identifier and each device to be recommended, thereby constructing a regional object co-occurrence matrix that matches the object recommendation request based on the regional co-occurrence degree between the user's regional identifier and each device to be recommended.

[0074] For example, assuming that the object recommendation system can recommend three types of devices proposed by the user, the regional object co-occurrence matrix generated by the object recommendation system can be shown in Table 2:

[0075] Table 2 Regional object co-occurrence matrix list

[0076] Equipment to be recommended 1 Equipment to be recommended 2 Equipment to be recommended 3 Region 1 0 1 0.2 Region 2 0.15 0.2 0 Region 3 1 0.3 0.1

[0077] S223 : Generate an operating rate object co-occurrence matrix matching the object recommendation request according to the user device operating rate and the object recommendation associated data.

[0078] In an optional embodiment of the present invention, generating a start-up rate object co-occurrence matrix matching the object recommendation request based on the user device start-up rate and the object recommendation associated data may include: calculating the start-up rate co-occurrence degree between the user device start-up rate and the object to be recommended based on the object recommendation associated data; and constructing a start-up rate object co-occurrence matrix matching the object recommendation request based on the mapping relationship between the user device start-up rate, the object to be recommended and the start-up rate co-occurrence degree.

[0079] The operating rate object co-occurrence matrix may be a co-occurrence matrix constructed from the perspective of the user device operating rate, and its matrix elements may be the operating rate co-occurrence degree between the user device operating rate and the device to be recommended. The operating rate co-occurrence degree is also the co-occurrence degree between the user device operating rate and the device to be recommended.

[0080] In an embodiment of the present invention, optionally, a matching calculation can be performed based on the user device operating rate in the user's multi-dimensional recommendation association factors and the object recommendation association data stored in the object recommendation system to obtain the co-occurrence degree between the user device operating rate and the operating rate of each device to be recommended, thereby constructing a operating rate object co-occurrence matrix that matches the object recommendation request based on the co-occurrence degree between the user device operating rate and the operating rate of each device to be recommended.

[0081] For example, assuming that the object recommendation system can recommend three types of equipment proposed by the user, the object recommendation system can generate a co-occurrence matrix of operating rate objects as shown in Table 3:

[0082] Table 3 Co-occurrence matrix of operating rate objects

[0083] Equipment to be recommended 1 Equipment to be recommended 2 Equipment to be recommended 3 Operating rate 1 1 0 0.5 Operating rate 2 0.15 0.25 0.2 Operating rate 3 1 0.5 0.2

[0084] It should be noted that Figure 2 This is only a schematic diagram of an implementation method, and there is no specific order in which step S221, step S222, and step S223 are executed. That is, the three steps can be executed serially, in parallel, or randomly, and the embodiment of the present invention does not limit this.

[0085] In an embodiment of the present invention, a multi-dimensional co-occurrence matrix can be constructed based on the object database of the object recommendation system. Optionally, if the recommended object is a device, the object database can be a device database, which can collect and store data for candidate recommended devices, including: device name, device ID, device photo, device nameplate, device age, device type, device attributes, etc.

[0086] Alternatively, the Pairs algorithm and the Stripes algorithm (a presentation layer framework for quickly building web applications) can be used to calculate co-occurrences. The Pairs algorithm generates more key-value pairs than the Stripes algorithm. Furthermore, the Pairs algorithm captures each individual co-occurrence event, while the Stripes algorithm captures all co-occurrence events for a given event. Because both algorithms produce exchange and association results, they can be implemented using combiners. Optionally, the Reducer of each Mapper can be used as a combiner. Specifically, the MapReduce algorithm (a programming model) of Hadoop (a distributed system infrastructure) can be used for parallel computing of large-scale datasets. It consists of two phases: Map and Reduce. In the Map phase, the map() function takes a key / value pair as input and generates another series of key / value pairs. The MapReduce framework automatically aggregates this intermediate data according to the key value, and data with the same key value is uniformly processed by the reduce() function. The reduce() function takes a key and its corresponding value list as input, merges the values ​​with the same key, and produces another series of key / value pairs as the final output.

[0087] S230: Extracting a first multi-dimensional feature of the object to be recommended based on the associated data of the object to be recommended.

[0088] The first multidimensional feature may be a multidimensional feature of the object to be recommended.

[0089] S240: Extract second multi-dimensional features of candidate recommended objects according to the object recommendation associated data.

[0090] The candidate recommendation object may be a recommendable object stored in the object recommendation system. The second multi-dimensional feature may be a multi-dimensional feature of the candidate recommendation object.

[0091] S250: Calculate the feature similarity between the first multidimensional feature and the second multidimensional feature as the object similarity between the to-be-recommended object and the candidate recommended object.

[0092] S260: Construct the object similarity matrix according to the mapping relationship between the objects to be recommended, the candidate recommended objects, and the object similarities.

[0093] The feature similarity is the similarity between the first multidimensional feature and the second multidimensional feature, and the object similarity is the similarity between the object to be recommended and the candidate recommended objects.

[0094] Optionally, if the object to be recommended is a device to be recommended, when generating an object similarity matrix for matching the object recommendation request, a first multidimensional feature of the device to be recommended can be determined based on the object-recommendation association data of the device to be recommended, and a second multidimensional feature of the candidate device to be recommended can be determined based on the object recommendation association data of the candidate device to be recommended. Accordingly, after obtaining the first multidimensional feature of the device to be recommended and the second multidimensional feature of the candidate device to be recommended, an object similarity matrix for matching the object recommendation request can be generated based on the first and second multidimensional features.

[0095] Exemplarily, the first multidimensional feature and the second multidimensional feature may be various device feature data such as the name, type, identification, and performance indicators of the device. The embodiment of the present invention does not limit the feature dimensions and feature types of the first multidimensional feature and the second multidimensional feature.

[0096] In an optional embodiment of the present invention, the calculation of the feature similarity between the first multidimensional feature and the second multidimensional feature may include: calculating the feature intersection and feature union of the first multidimensional feature and the second multidimensional feature; and using the ratio of the feature intersection and the feature union as the feature similarity of the object similarity matrix.

[0097] The feature intersection is the intersection of the first multidimensional feature and the second multidimensional feature, and the feature union is the union of the first multidimensional feature and the second multidimensional feature.

[0098] For example, the Jaccard similarity calculation formula may be used to calculate the feature similarity value between the first multidimensional feature and the second multidimensional feature. The formula may be expressed as follows: Among them, J(M,N) represents the feature similarity between the recommended object M and the candidate recommended object N, that is, the object similarity, A M Represents the first multidimensional feature of the object to be recommended, A N The second multidimensional feature representing the candidate recommendation object.

[0099] For example, assuming that the device database of the object recommendation system stores three candidate recommended devices, the object similarity matrix constructed and generated by the object recommendation system may be as shown in Table 4:

[0100] Table 4 Object similarity matrix list

[0101] Alternative recommended equipment 1 Alternative recommended equipment 2 Alternative recommended equipment 3 Equipment to be recommended 1 0 1 1 Equipment to be recommended 2 1 0 1 Equipment to be recommended 3 1 1 0

[0102] It should be noted that the similarity in Table 4 is only represented by 0 and 1 for example, and the actual similarity value can be between 0 and 1 (including 0 and 1), and the embodiment of the present invention is not limited to this.

[0103] S270 : Generate a multi-dimensional relationship matrix for matching the object recommendation request according to the multi-dimensional co-occurrence matrix and the object similarity matrix.

[0104] Accordingly, step S270 may specifically include the following parallel operations:

[0105] S271 : Generate a requirement-object relationship matrix for matching the object recommendation request according to the requirement-object co-occurrence matrix and the object similarity matrix.

[0106] In an optional embodiment of the present invention, generating a requirement-object relationship matrix matching the object recommendation request based on the requirement-object co-occurrence matrix and the object similarity matrix may include: calculating the requirement recommendation score between the user requirement identifier and the alternative recommendation object based on the matrix elements of the requirement-object co-occurrence matrix and the matrix elements of the object similarity matrix; and constructing a requirement-object relationship matrix matching the object recommendation request based on the mapping relationship between the user requirement identifier, the alternative recommendation object and the requirement recommendation score.

[0107] The demand recommendation score can represent the degree of match between the user's demand identification and the candidate recommended objects. It is understood that a higher demand recommendation score indicates a closer match between the user's demand identification and the candidate recommended objects. The demand-object relationship matrix can be a relationship matrix constructed from the perspective of user needs, and its matrix elements can be the degree of match between the user's needs and the candidate recommended devices.

[0108] In an embodiment of the present invention, optionally, a matching calculation can be performed based on the demand object co-occurrence matrix and the object similarity matrix obtained by the above calculation to generate a demand object relationship matrix that matches the object recommendation request. Specifically, the matrix elements of the demand object co-occurrence matrix and the object similarity matrix can be multiplied and added to obtain the value of each matrix element of the demand object relationship matrix. Optionally, when the object to be recommended is the device to be recommended, taking the user demand as the industry to which the user belongs as an example, each matrix element of the demand object relationship matrix can be the demand recommendation score for the industry to which the user belongs and the corresponding device. It can be understood that the higher the demand recommendation score, the more the industry needs the device.

[0109] For example, assuming the demand object co-occurrence matrix shown in Table 1 and the object similarity matrix shown in Table 4, the co-occurrence degree between industry 1 and each device to be recommended can be multiplied by the similarity between industry 1 and each candidate device to obtain the demand recommendation score for industry 1 and each candidate device. Specifically, the demand recommendation score for industry 1 and candidate device 1 can be: 0*0+0*1+0.23*1=0.23, the demand recommendation score for industry 1 and candidate device 2 can be: 0*1+0*0+0.23*1=0.23, and the demand recommendation score for industry 1 and candidate device 3 can be: 0*1+0*1+0.23*0=0. The demand recommendation score for industry 2 and candidate device 2 can be: 0.25*1+0.11*0+0*1=0.25, and the demand recommendation score for industry 2 and candidate device 3 can be: 0.25*1+0.11*1+0*0=0.36. The demand recommendation score for industry 3 and alternative recommended device 1 can be: 0*0+0.31*1+0.16*1=0.47, the demand recommendation score for industry 3 and alternative recommended device 2 can be: 0*1+0.31*0+0.16*1=0.16, and the demand recommendation score for industry 3 and alternative recommended device 3 can be: 0*0+0.31*1+0.16*1=0.31. Thus, we can obtain the demand object relationship matrix shown in Table 5.

[0110] Table 5 Requirements object relationship matrix list

[0111] Alternative recommended equipment 1 Alternative recommended equipment 2 Alternative recommended equipment 3 Industry 1 0.23 0.23 0 Industry 2 0.11 0.25 0.36 Industry 3 0.47 0.16 0.31

[0112] S272: Generate a region-object relationship matrix matching the object recommendation request according to the region-object co-occurrence matrix and the object similarity matrix.

[0113] In an optional embodiment of the present invention, generating a regional object relationship matrix that matches the object recommendation request based on the regional object co-occurrence matrix and the object similarity matrix may include: calculating the regional recommendation score between the user regional identifier and the candidate recommendation object based on the matrix elements of the regional object co-occurrence matrix and the matrix elements of the object similarity matrix; and constructing a regional object relationship matrix that matches the object recommendation request based on the mapping relationship between the user regional identifier, the candidate recommendation objects, and the regional recommendation score.

[0114] The region recommendation score can represent the degree of match between the user's region and the candidate recommended objects. It can be understood that a higher region recommendation score indicates a closer match between the user's region and the candidate recommended objects. The region-object relationship matrix can be a relationship matrix constructed from the perspective of user needs, with its elements representing the degree of match between the user's region and the candidate recommended devices.

[0115] In this embodiment of the present invention, a matching calculation can optionally be performed based on the region-object co-occurrence matrix and object similarity matrix calculated above to generate a region-object relationship matrix that matches the object recommendation request. Specifically, the matrix elements of the region-object co-occurrence matrix and the object similarity matrix can be multiplied and added together to obtain a region recommendation score for each matrix element in the region-object relationship matrix. The calculation method for the region-object relationship matrix is ​​similar to that for the requirement-object relationship matrix and will not be further elaborated.

[0116] S273 : Generate an operating rate object relationship matrix that matches the object recommendation request according to the operating rate object co-occurrence matrix and the object similarity matrix.

[0117] In an optional embodiment of the present invention, generating a start-up rate object relationship matrix matching the object recommendation request based on the start-up rate object co-occurrence matrix and the object similarity matrix may include: calculating the start-up rate recommendation score between the user device start-up rate and the alternative recommended object based on the matrix elements of the start-up rate object co-occurrence matrix and the matrix elements of the object similarity matrix; and constructing a start-up rate object relationship matrix matching the object recommendation request based on the mapping relationship between the user device start-up rate, the alternative recommended object and the start-up rate recommendation score.

[0118] The recommended operating rate score can represent the degree of match between the user's operating rate and the candidate recommended objects. It can be understood that a higher recommended operating rate score indicates a closer match between the user's operating rate and the candidate recommended objects. The operating rate object relationship matrix can be a relationship matrix constructed from the perspective of the user's device operating rates, and its matrix elements can be the degree of match between the user's device operating rates and the candidate recommended devices.

[0119] In an embodiment of the present invention, optionally, a matching calculation can be performed based on the operating rate object co-occurrence matrix and the object similarity matrix calculated above to generate an operating rate object relationship matrix that matches the object recommendation request. Specifically, the matrix elements of the operating rate object co-occurrence matrix and the object similarity matrix can be multiplied and added together to obtain the operating rate recommendation score for each matrix element of the operating rate object relationship matrix. The calculation method of the operating rate object relationship matrix is ​​the same as that of the demand object relationship matrix, and will not be repeated here.

[0120] It should be noted that Figure 2 This is only a schematic diagram of an implementation method, and there is no specific order in which step S261, step S262, and step S263 are executed. That is, the three steps can be executed serially, in parallel, or randomly, and the embodiment of the present invention does not limit this.

[0121] S280: Determine a recommended object that matches the object recommendation request according to the multi-dimensional relationship matrix.

[0122] In an optional embodiment of the present invention, determining the recommended object that matches the object recommendation request based on the multi-dimensional relationship matrix may include: taking the alternative recommended object with the highest demand recommendation score in the demand object relationship matrix as the first alternative recommended object; taking the alternative recommended object with the highest region recommendation score in the region object relationship matrix as the second alternative recommended object; taking the alternative recommended object with the highest operating rate recommendation score in the operating rate object relationship matrix as the third alternative recommended object; and determining the recommended object that matches the object recommendation request based on the first alternative recommended object, the second alternative recommended object and the third alternative recommended object.

[0123] The first candidate recommended object may be the candidate recommended object that best matches the object recommendation request, determined based on the demand object relationship matrix. The second candidate recommended object may be the candidate recommended object that best matches the object recommendation request, determined based on the region object relationship matrix. The third candidate recommended object may be the candidate recommended object that best matches the object recommendation request, determined based on the operating rate object relationship matrix.

[0124] Because the object recommendation system constructs multiple relationship matrices based on the user's object recommendation request, the best-matching candidate recommendation object can be calculated for each relationship matrix, that is, the first candidate recommendation object, the second candidate recommendation object, and the third candidate recommendation object. Optionally, the first candidate recommendation object, the second candidate recommendation object, and the third candidate recommendation object can be the candidate recommendation object with the highest score in each relationship matrix. As shown in Table 5, the first candidate recommendation device for user 1 in industry 1 can be candidate recommendation device 1 and candidate recommendation device 2, and the second candidate recommendation device for user 2 in industry 2 can be candidate recommendation device 3.

[0125] Furthermore, the recommended objects that match the object recommendation request can be finally determined based on the candidate recommended objects determined by the various relationship matrices. For example, the first candidate recommended object, the second candidate recommended object, and the third candidate recommended object can all be used as recommended objects, or one or more can be selected from the first candidate recommended object, the second candidate recommended object, and the third candidate recommended object as recommended objects. Alternatively, new candidate recommended objects can be recalculated as recommended objects based on the first candidate recommended object, the second candidate recommended object, and the third candidate recommended object. The embodiment of the present invention does not limit the method for determining the recommended objects.

[0126] In an optional embodiment of the present invention, determining the recommended object that matches the object recommendation request based on the first alternative recommendation object, the second alternative recommendation object and the third alternative recommendation object may include: obtaining the recommendation weights of the demand object relationship matrix, the region object relationship matrix and the operating rate object relationship matrix; calculating the comprehensive recommendation score between the user and the alternative recommendation object based on the recommendation weights of each relationship matrix and the matrix elements of the multi-dimensional relationship matrix; constructing a comprehensive relationship matrix that matches the object recommendation request based on the mapping relationship between the user, the alternative recommendation object and the comprehensive recommendation score; taking the alternative recommendation object with the highest comprehensive recommendation score in the comprehensive relationship matrix as the fourth alternative recommendation object; and determining the recommended object that matches the object recommendation request based on the first alternative recommendation object, the second alternative recommendation object, the third alternative recommendation object and the fourth alternative recommendation object.

[0127] The recommendation weight can be the weight of each relationship matrix in the recommendation. The comprehensive recommendation score can represent the degree of match between the user and the candidate recommendation object. It is understood that a higher comprehensive recommendation score indicates a closer match between the user and the candidate recommendation object. The comprehensive relationship matrix can be a total relationship matrix calculated based on the recommendation weights of each relationship matrix and each relationship matrix. The fourth candidate recommendation object can be the candidate recommendation object obtained based on the comprehensive relationship matrix.

[0128] To further improve the accuracy of object recommendations, after obtaining the multidimensional relationship matrix, the recommendation weights for each relationship matrix can be determined based on their importance. Furthermore, a comprehensive recommendation score for the combined relationship matrix is ​​calculated for each relationship matrix and its matching recommendation weights. This allows the mapping between the user, candidate recommendation objects, and the comprehensive recommendation scores to construct a comprehensive relationship matrix that matches the object recommendation request. Optionally, the comprehensive recommendation score can be the matching score between the user and each candidate recommendation device.

[0129] For example, assume that the recommended weight of the configuration demand object relationship matrix is ​​0.5, the recommended weight of the region object relationship matrix is ​​0.25, and the recommended weight of the operating rate object relationship matrix is ​​0.25. That is, the ratio of the recommended weights of the demand object relationship matrix, the region object relationship matrix, and the operating rate object relationship matrix is ​​2:1:1. The matching score between industry 1 and equipment 1 is 0.23, the matching score between industry 1 and equipment 2 is 0.23, and the matching score between industry 1 and equipment 3 is 0. The matching score between region 1 and equipment 1 is 0.2, the matching score between region 1 and equipment 2 is 0.3, and the matching score between region 1 and equipment 3 is 0.5. The matching score between operating rate 1 and equipment 1 is 0.3, the matching score between operating rate 1 and equipment 2 is 0.4, and the matching score between operating rate 1 and equipment 3 is 0. In the comprehensive relationship matrix, the matching score between user 1 and candidate recommended device 1 is: 0.23*0.5+0.2*0.25+0.3*0.25=0.24, the matching score between user 1 and candidate recommended device 2 is: 0.23*0.5+0.3*0.25+0.4*0.25=0.29, and the matching score between user 1 and candidate recommended device 3 is: 0*0.5+0.5*0.25+0*0.25=0.125. Similarly, we can obtain the matching score values ​​between user 2, user 3, and other users and each candidate recommended device in the comprehensive relationship matrix.

[0130] Accordingly, after obtaining the comprehensive relationship matrix, the candidate recommended object with the highest comprehensive recommendation score with the user in the comprehensive relationship matrix can be selected as the fourth candidate recommended object for the user. Taking the above example as an example, for user 1, candidate recommended device 2 can be selected as the fourth candidate recommended device for user 1.

[0131] Correspondingly, the recommended object that matches the object recommendation request is determined based on the first alternative recommended object, the second alternative recommended object, the third alternative recommended object and the fourth alternative recommended object. The first alternative recommended object, the second alternative recommended object, the third alternative recommended object and the fourth alternative recommended object may all be used as the recommended object that matches the object recommendation request, or only the fourth alternative recommended object may be used as the recommended object that matches the object recommendation request. The embodiment of the present invention does not limit this.

[0132] In this embodiment of the present invention, if a user's object recommendation request does not include the user's multi-dimensional recommendation-related factors, for example, if the user is a new user and the object recommendation system does not have any information about the new user, the object recommendation system may use random recommendation to recommend objects to the new user. Optionally, random recommendation may involve randomly selecting a set number of candidate recommended objects with the same probability as the recommended objects. For example, all devices in the device library may randomly select four pieces of data with the same probability to recommend them to the new user.

[0133] The above technical solution, by using a multidimensional co-occurrence matrix and an object similarity matrix to generate a multidimensional relationship matrix that matches the user's object recommendation request, and ultimately determining the recommended objects that match the object recommendation request based on the generated multidimensional relationship matrix, can improve the accuracy, recall rate, coverage rate, diversity, novelty, surprise, and real-time performance of object recommendations, thereby achieving precise push notifications to users and promoting the transaction conversion rate of recommended objects. The object recommendation method provided by the embodiment of the present invention can not only reduce the time users spend selecting or browsing objects, but also improve the user experience, enhance user favorability, and increase user retention rate, and ultimately optimize asset disposal efficiency, minimize operating costs, increase cash flow rate, obtain more profits, and enhance market competitiveness.

[0134] It should be noted that the relevant information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data comply with relevant laws, regulations and standards in the relevant regions.

[0135] It should be noted that any arrangement and combination of the technical features in the above embodiments also falls within the protection scope of the present invention.

[0136] Example 3

[0137] Figure 3 is a schematic diagram of an object recommendation device provided by the third embodiment of the present invention, such as Figure 3 As shown, the apparatus includes: an object recommendation request acquisition module 310, a multi-dimensional co-occurrence matrix generation module 320, an object similarity matrix generation module 330, a multi-dimensional relationship matrix generation module 340, and a recommended object determination module 350, wherein:

[0138] The object recommendation request acquisition module 310 is configured to acquire an object recommendation request sent by a user; wherein the object recommendation request includes object-related data of the object to be recommended and the multi-dimensional recommendation-related factors of the user;

[0139] A multi-dimensional co-occurrence matrix generating module 320 is configured to generate a multi-dimensional co-occurrence matrix matching the object recommendation request based on the user's multi-dimensional recommendation association factors and the object recommendation association data;

[0140] An object similarity matrix generating module 330 is configured to generate an object similarity matrix matching the object recommendation request based on the to-be-recommended object association data and the object recommendation association data;

[0141] A multi-dimensional relationship matrix generating module 340 is configured to generate a multi-dimensional relationship matrix matching the object recommendation request based on the multi-dimensional co-occurrence matrix and the object similarity matrix;

[0142] The recommended object determination module 350 is configured to determine a recommended object that matches the object recommendation request according to the multi-dimensional relationship matrix.

[0143] The embodiment of the present invention generates a multidimensional co-occurrence matrix for matching object recommendation requests based on the user's multidimensional recommendation association factors and object recommendation association data included in the object recommendation request for the object to be recommended sent by the user, and generates an object similarity matrix for matching object recommendation requests based on the object association data to be recommended and the object recommendation association data included in the object recommendation request, so as to generate a multidimensional relationship matrix for matching object recommendation requests based on the multidimensional co-occurrence matrix and the object similarity matrix, and finally determines the recommended object matching the object recommendation request based on the generated multidimensional relationship matrix, thereby solving the problem of low recommendation accuracy existing in existing object recommendation methods, and can improve the accuracy of object recommendation, thereby improving user experience.

[0144] Optionally, the object to be recommended includes the device to be recommended, and the multi-dimensional recommendation association factors include the user demand identifier, the user region identifier and the user device operating rate; the multi-dimensional co-occurrence matrix generation module 320 is specifically used to: calculate the demand co-occurrence between the user demand identifier and the object to be recommended based on the object recommendation association data; construct a demand-object co-occurrence matrix that matches the object recommendation request based on the mapping relationship between the user demand identifier, the object to be recommended and the demand co-occurrence; calculate the regional co-occurrence between the user region identifier and the object to be recommended based on the object recommendation association data; construct a regional object co-occurrence matrix that matches the object recommendation request based on the mapping relationship between the user region identifier, the object to be recommended and the regional co-occurrence; calculate the operating rate co-occurrence between the user device operating rate and the object to be recommended based on the object recommendation association data; construct a operating rate object co-occurrence matrix that matches the object recommendation request based on the mapping relationship between the user device operating rate, the object to be recommended and the operating rate co-occurrence.

[0145] Optionally, the object similarity matrix generation module 330 is specifically used to: extract the first multidimensional feature of the object to be recommended based on the associated data of the object to be recommended; extract the second multidimensional feature of the alternative recommended object based on the object recommendation associated data; calculate the feature similarity between the first multidimensional feature and the second multidimensional feature as the object similarity between the object to be recommended and the alternative recommended object; and construct the object similarity matrix based on the mapping relationship between the object to be recommended, the alternative recommended object and the object similarity.

[0146] Optionally, the object similarity matrix generation module 330 is specifically configured to: calculate a feature intersection and a feature union of the first multidimensional feature and the second multidimensional feature; and use a ratio of the feature intersection and the feature union as the feature similarity.

[0147] Optionally, the multi-dimensional relationship matrix generation module 340 is specifically used to: calculate the demand recommendation score between the user demand identifier and the alternative recommended object based on the matrix elements of the demand-object co-occurrence matrix and the matrix elements of the object similarity matrix; construct a demand-object relationship matrix that matches the object recommendation request based on the mapping relationship between the user demand identifier, the alternative recommended object and the demand recommendation score; calculate the regional recommendation score between the user regional identifier and the alternative recommended object based on the matrix elements of the regional object co-occurrence matrix and the matrix elements of the object similarity matrix; construct a regional object relationship matrix that matches the object recommendation request based on the mapping relationship between the user regional identifier, the alternative recommended object and the regional recommendation score; calculate the operating rate recommendation score between the user device operating rate and the alternative recommended object based on the matrix elements of the operating rate object co-occurrence matrix and the matrix elements of the object similarity matrix; and construct a operating rate object relationship matrix that matches the object recommendation request based on the mapping relationship between the user device operating rate, the alternative recommended object and the operating rate recommendation score.

[0148] Optionally, the recommended object determination module 350 is specifically used to: use the alternative recommended object with the highest demand recommendation score in the demand object relationship matrix as the first alternative recommended object; use the alternative recommended object with the highest regional recommendation score in the regional object relationship matrix as the second alternative recommended object; use the alternative recommended object with the highest operating rate recommendation score in the operating rate object relationship matrix as the third alternative recommended object; and determine the recommended object that matches the object recommendation request based on the first alternative recommended object, the second alternative recommended object and the third alternative recommended object.

[0149] Optionally, the recommended object determination module 350 is specifically used to: obtain the recommendation weights of the demand object relationship matrix, the region object relationship matrix and the operating rate object relationship matrix; calculate the comprehensive recommendation score between the user and the alternative recommendation object based on the recommendation weights of each relationship matrix and the matrix elements of the multi-dimensional relationship matrix; construct a comprehensive relationship matrix that matches the object recommendation request based on the mapping relationship between the user, the alternative recommendation object and the comprehensive recommendation score; use the alternative recommendation object with the highest comprehensive recommendation score in the comprehensive relationship matrix as the fourth alternative recommendation object; determine the recommended object that matches the object recommendation request based on the first alternative recommendation object, the second alternative recommendation object, the third alternative recommendation object and the fourth alternative recommendation object.

[0150] The object recommendation device can execute the object recommendation method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method. For technical details not fully described in this embodiment, please refer to the object recommendation method provided by any embodiment of the present invention.

[0151] Since the object recommendation device described above is a device that can execute the object recommendation method in the embodiment of the present invention, based on the object recommendation method described in the embodiment of the present invention, those skilled in the art can understand the specific implementation of the object recommendation device of this embodiment and its various variations. Therefore, how the object recommendation device implements the object recommendation method in the embodiment of the present invention will not be described in detail here. As long as those skilled in the art implement the device used in the object recommendation method in the embodiment of the present invention, it falls within the scope of protection of this application.

[0152] Example 4

[0153] Figure 4 This is a structural diagram of an electronic device provided in Example 4 of the present invention. Figure 4 A block diagram of an exemplary electronic device 12 suitable for implementing embodiments of the present invention is shown. Figure 4 The electronic device 12 shown is only an example and should not limit the functionality and scope of use of the embodiments of the present invention.

[0154] like Figure 4 As shown, the electronic device 12 is implemented as a general-purpose computing device. Components of the electronic device 12 may include, but are not limited to, one or more processors 16, a memory 28, and a bus 18 that connects various system components (including the memory 28 and the processor 16).

[0155] Bus 18 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus architectures. Examples of these architectures include, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MCA) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

[0156] The electronic device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the electronic device 12, including volatile and non-volatile media, removable and non-removable media.

[0157] The memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. The electronic device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 34 may be configured to read and write non-removable, non-volatile magnetic media ( Figure 4 Not shown, often called a "hard drive"). Although Figure 4 Not shown, a magnetic disk drive for reading and writing to a removable non-volatile magnetic disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a Compact Disc-Read Only Memory (CD-ROM), a Digital Video Disc-Read Only Memory (DVD-ROM), or other optical media) may be provided. In these cases, each drive may be connected to bus 18 via one or more data medium interfaces. Memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of various embodiments of the present invention.

[0158] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in memory 28. Such program modules 42 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may include an implementation of a network environment. Program modules 42 generally implement the functions and / or methods of the embodiments described herein.

[0159] The electronic device 12 may also communicate with one or more external devices 14 (e.g., keyboards, pointing devices, displays 24, etc.), and may also communicate with one or more devices that enable a user to interact with the electronic device 12, and / or any device that enables the electronic device 12 to communicate with one or more other computing devices (e.g., a network card, a modem, etc.). Such communication may be performed through an input / output (I / O) interface 22. Furthermore, the electronic device 12 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN) and / or a public network, such as the Internet) through a network adapter 20. As shown, the network adapter 20 communicates with other modules of the electronic device 12 via the bus 18. It should be understood that although Figure 4 Not shown, other hardware and / or software modules may be used in conjunction with the electronic device 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, (Redundant Arrays of Independent Disks, RAID) systems, tape drives, and data backup storage systems.

[0160] The processor 16 executes various functional applications and data processing by running the program stored in the memory 28, thereby realizing the object recommendation method provided by the embodiment of the present invention: obtaining an object recommendation request sent by a user; wherein the object recommendation request includes the object association data of the object to be recommended and the multi-dimensional recommendation association factors of the user; generating a multi-dimensional co-occurrence matrix matching the object recommendation request based on the multi-dimensional recommendation association factors of the user and the object recommendation association data; generating an object similarity matrix matching the object recommendation request based on the object association data to be recommended and the object recommendation association data; generating a multi-dimensional relationship matrix matching the object recommendation request based on the multi-dimensional co-occurrence matrix and the object similarity matrix; and determining the recommended object matching the object recommendation request based on the multi-dimensional relationship matrix.

[0161] Example 5

[0162] Embodiment 5 of the present invention also provides a computer storage medium storing a computer program, which, when executed by a computer processor, is used to execute the object recommendation method described in any of the above embodiments of the present invention: obtaining an object recommendation request sent by a user; wherein the object recommendation request includes object association data of an object to be recommended and multi-dimensional recommendation association factors of the user; generating a multi-dimensional co-occurrence matrix matching the object recommendation request based on the multi-dimensional recommendation association factors of the user and the object recommendation association data; generating an object similarity matrix matching the object recommendation request based on the object association data to be recommended and the object recommendation association data; generating a multi-dimensional relationship matrix matching the object recommendation request based on the multi-dimensional co-occurrence matrix and the object similarity matrix; and determining the recommended object matching the object recommendation request based on the multi-dimensional relationship matrix.

[0163] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination thereof. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM, or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by an instruction execution system, device or device or used in combination with it.

[0164] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0165] The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wire, optical cable, radio frequency (RF), etc., or any suitable combination of the foregoing.

[0166] Computer program code for performing the operations of the present invention may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0167] Note that the above are only preferred embodiments of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments, and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments and may include many other equivalent embodiments without departing from the concept of the present invention. The scope of the present invention is determined by the scope of the appended claims.

Claims

1. An object recommendation method, characterized in that: include: Obtaining an object recommendation request sent by a user; wherein the object recommendation request includes object-related data of the object to be recommended and the multi-dimensional recommendation-related factors of the user; generating a multidimensional co-occurrence matrix matching the object recommendation request according to the multidimensional recommendation association factors and the object recommendation association data; Generating a multi-dimensional co-occurrence matrix matching the object recommendation request according to the multi-dimensional recommendation association factors and the object recommendation association data includes: Calculating, based on the object recommendation association data, the co-occurrence of the plurality of multi-dimensional recommendation association factors between the plurality of multi-dimensional recommendation association factors and the object to be recommended; Constructing a plurality of multidimensional co-occurrence matrices matching the object recommendation request based on a mapping relationship between the plurality of multidimensional recommendation association factors, the objects to be recommended, and the co-occurrence degrees of the corresponding plurality of multidimensional recommendation association factors; generating an object similarity matrix matching the object recommendation request according to the to-be-recommended object association data and the object recommendation association data; generating a multidimensional relationship matrix matching the object recommendation request according to the multidimensional co-occurrence matrix and the object similarity matrix; Determine a recommended object that matches the object recommendation request according to the multi-dimensional relationship matrix.

2. The method according to claim 1, characterized in that The objects to be recommended include equipment to be recommended, and the multi-dimensional recommendation related factors include user demand identification, user region identification, and user equipment operating rate; Generating a multi-dimensional co-occurrence matrix matching the object recommendation request according to the multi-dimensional recommendation association factors and the object recommendation association data includes: Calculating the demand co-occurrence degree between the user demand identifier and the object to be recommended according to the object recommendation association data; Constructing a demand-object co-occurrence matrix that matches the object recommendation request according to a mapping relationship between the user demand identifier, the object to be recommended, and the demand co-occurrence degree; Calculating the regional co-occurrence between the user region identifier and the object to be recommended according to the object recommendation association data; Constructing a regional object co-occurrence matrix matching the object recommendation request according to a mapping relationship between the user region identifier, the object to be recommended, and the regional co-occurrence degree; Calculating the co-occurrence degree of the operating rate between the operating rate of the user device and the object to be recommended according to the object recommendation association data; An operating rate object co-occurrence matrix matching the object recommendation request is constructed according to a mapping relationship among the operating rate of the user equipment, the object to be recommended, and the operating rate co-occurrence degree.

3. The method according to claim 1, characterized in that The generating of an object similarity matrix matching the object recommendation request according to the to-be-recommended object association data and the object recommendation association data includes: Extracting a first multidimensional feature of the object to be recommended based on the associated data of the object to be recommended; extracting second multidimensional features of candidate recommended objects based on the object recommendation associated data; Calculating a feature similarity between the first multidimensional feature and the second multidimensional feature as an object similarity between the to-be-recommended object and the candidate recommended object; The object similarity matrix is ​​constructed according to the mapping relationship between the to-be-recommended objects, the candidate recommended objects, and the object similarities.

4. The method according to claim 3, characterized in that The calculating the feature similarity between the first multidimensional feature and the second multidimensional feature includes: Calculating a feature intersection and a feature union of the first multidimensional feature and the second multidimensional feature; The ratio of the feature intersection to the feature union is used as the feature similarity.

5. The method according to claim 2, characterized in that Generating a multidimensional relationship matrix matching the object recommendation request according to the multidimensional co-occurrence matrix and the object similarity matrix includes: Calculating a demand recommendation score between the user demand identifier and the candidate recommended object based on the matrix elements of the demand-object co-occurrence matrix and the matrix elements of the object similarity matrix; Constructing a demand-object relationship matrix that matches the object recommendation request according to a mapping relationship among the user demand identifier, the candidate recommended objects, and the demand recommendation score; Calculating a region recommendation score between the user region identifier and the candidate recommended object based on matrix elements of the region-object co-occurrence matrix and matrix elements of the object similarity matrix; Constructing a region-object relationship matrix matching the object recommendation request according to a mapping relationship among the user region identifier, the candidate recommended objects, and the region recommendation scores; Calculating a recommended operating rate score between the operating rate of the user device and the candidate recommended object based on the matrix elements of the operating rate object co-occurrence matrix and the matrix elements of the object similarity matrix; An operating rate object relationship matrix matching the object recommendation request is constructed according to a mapping relationship among the operating rate of the user equipment, the candidate recommended objects, and the operating rate recommendation scores.

6. The method according to claim 5, characterized in that The determining, according to the multi-dimensional relationship matrix, a recommended object that matches the object recommendation request includes: The candidate recommended object with the highest demand recommendation score in the demand object relationship matrix is ​​selected as the first candidate recommended object; The candidate recommended object with the highest regional recommendation score in the regional object relationship matrix is ​​selected as the second candidate recommended object; The candidate recommended object with the highest recommended score for the operating rate in the operating rate object relationship matrix is ​​selected as the third candidate recommended object; A recommended object matching the object recommendation request is determined according to the first candidate recommended object, the second candidate recommended object, and the third candidate recommended object.

7. The method according to claim 6, characterized in that The determining a recommended object matching the object recommendation request according to the first candidate recommended object, the second candidate recommended object, and the third candidate recommended object includes: Obtaining recommended weights of the demand object relationship matrix, the region object relationship matrix, and the operating rate object relationship matrix; Calculating a comprehensive recommendation score between the user and the candidate recommendation object based on the recommendation weights of each relationship matrix and the matrix elements of the multi-dimensional relationship matrix; Constructing a comprehensive relationship matrix matching the object recommendation request according to the mapping relationship between the user, the candidate recommendation objects, and the comprehensive recommendation scores; The candidate recommendation object with the highest comprehensive recommendation score in the comprehensive relationship matrix is ​​selected as the fourth candidate recommendation object; A recommended object that matches the object recommendation request is determined according to the first candidate recommended object, the second candidate recommended object, the third candidate recommended object, and the fourth candidate recommended object.

8. An object recommendation device, characterized in that: include: An object recommendation request acquisition module is configured to acquire an object recommendation request sent by a user; wherein the object recommendation request includes object-related data of the object to be recommended and the multi-dimensional recommendation-related factors of the user; A multi-dimensional co-occurrence matrix generating module, configured to generate a multi-dimensional co-occurrence matrix matching the object recommendation request based on the user's multi-dimensional recommendation association factors and the object recommendation association data; The multi-dimensional co-occurrence matrix generation module is further configured to: calculate, based on the object recommendation association data, the co-occurrence degrees of the multiple multi-dimensional recommendation association factors between the multiple multi-dimensional recommendation association factors and the object to be recommended; and construct, based on a mapping relationship between the multiple multi-dimensional recommendation association factors, the object to be recommended, and the corresponding co-occurrence degrees of the multiple multi-dimensional recommendation association factors, a plurality of multi-dimensional co-occurrence matrices matching the object recommendation request; An object similarity matrix generating module, configured to generate an object similarity matrix matching the object recommendation request based on the to-be-recommended object association data and the object recommendation association data; a multi-dimensional relationship matrix generating module, configured to generate a multi-dimensional relationship matrix matching the object recommendation request based on the multi-dimensional co-occurrence matrix and the object similarity matrix; The recommended object determination module is used to determine the recommended object that matches the object recommendation request according to the multi-dimensional relationship matrix.

9. An electronic device, characterized in that: The electronic device comprises: one or more processors; a storage device for storing one or more computer programs; When the one or more computer programs are executed by the one or more processors, the object recommendation method according to any one of claims 1 to 7 is implemented when the one or more processors execute the computer programs.

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

Images