CN116402569A

CN116402569A - Commodity recommendation method, device and system based on knowledge graph and storage medium

Info

Publication number: CN116402569A
Application number: CN202310223174.3A
Authority: CN
Inventors: 吴菁; 周研; 张晨
Original assignee: Zhejiang Create Link Technology Co ltd
Current assignee: Zhejiang Create Link Technology Co ltd
Priority date: 2023-03-09
Filing date: 2023-03-09
Publication date: 2023-07-07

Abstract

The embodiment of the invention discloses a commodity recommendation method, device and system based on a knowledge graph and a storage medium. As a simple example, the method provided by the embodiment of the invention only defines similar users from the same purchasing behavior, in addition, the product features, brand features, user labels and the like can be used for defining the user similarity, and are constructed into a graph model as key association entities of similarity recommendation, clustering is carried out through a clustering algorithm, and clustering conditions among different features are found in an unsupervised learning mode, so that accurate recommendation based on multidimensional features is realized. Different from the relational database, the schema of the graph database has strong flexibility, directly reflects business logic, and analysts can flexibly add different kinds of new relations, new nodes and new labels to form new subgraphs according to scene and business requirement changes, so that new recommendation strategies can be dynamically adjusted without worrying about destroying the functions of the existing query or application programs.

Description

Commodity recommendation method, device and system based on knowledge graph and storage medium

Technical Field

The invention relates to the technical field of computer software, in particular to a commodity recommendation method, device and system based on a knowledge graph and a storage medium.

Background

With the improvement of consumer living standard, the popularity of the mobile internet is enhanced, and the development of an electronic commerce system is more mature. The e-commerce industry is continually growing at a remarkable rate compared to the off-line retail industry.

The recommendation system is essentially a bridge between the clients and the commodities, and the basic task is to help the clients solve the problem of information overload, and accurately and quickly find the favorite and most likely to purchase objects of the clients from mass commodities. Therefore, two key points of the recommendation system, one is quasi and one is fast.

In the internet field, application scenarios common to online recommendation systems can be broadly divided into two categories: one is recommendation based on user dimension, namely recommendation is performed according to historical behaviors and interest preference of the user, such as recommendation song list of the networkcloud home page, discovery of the small red book home page and the like; the other is the recommendation of the dimension of the item, namely the recommendation is made according to the object currently browsed by the user, such as the "find similar" function of the Taobao specific commodity. Whether based on recommendations of user dimensions or based on recommendations of item dimensions, the recommendation process is essentially a process of information filtering: the recommendation system filters out products unlikely to be interested by the user before the user exits the page by analyzing the historical purchase and current behavior patterns of the user, and returns the most relevant Top-N product list according to the priority, and the general flow is shown in figure 1.

The electronic commerce platform is used as a provider of online shopping service, and the main requirement is to realize sales transformation. Various recommendation systems have been developed to help customers accurately and quickly find the most favorite and most likely to purchase items from a large number of items, thereby increasing customer viscosity and improving customer experience. The current product catalog change period is short and the change speed is high; the potential customer service requirement is high and the waiting tolerance is low; the recommendation dimension is complex and the object association degree is high under the scene marketing background. The rapid understanding of the customer multidimensional image, the real-time correlation of the customer's immediate interest with the constantly iterated product information, and the formation of the sceneries, the customization recommendation, become a great challenge.

One scenario similar to merchandise recommendations is where a brand party wishes to achieve accurate marketing of a product through a social network of users. The user can directly place an order for the goods recommended by the friends, compared with the traditional consumption process of searching, browsing and placing an order, the purchasing decision chain is shorter, and the consumers can be influenced more directly. By means of big data technology and mobile social media, social applications exhibit significant mobilization, localization features, which are good business marketing diversion entries. Because of the natural social nature of purchasing decisions, social layout is also being performed in the fields of electronic commerce, games, video, internet finance, and even online education. How to effectively enhance acquaintance relation propagation effect of a social network through a big data technology, increase brand exposure, pull user growth and promote user viscosity becomes a key of social electronic commerce platform differentiation competitiveness.

In summary, whether the traditional e-commerce platform or the emerging social e-commerce platform exists complex associations among a large number of users, commodities and commodities, and how to analyze and mine the association relations becomes a key for improving commodity recommendation accuracy.

Disclosure of Invention

Aiming at the technical defects in the prior art, the embodiment of the invention aims to provide a commodity recommendation method, device, system and storage medium based on a knowledge graph, so as to solve the technical problems in the background art.

In order to achieve the above object, in a first aspect, an embodiment of the present invention provides a commodity recommendation method based on a knowledge graph, including:

acquiring historical consumption behavior data of all users, and constructing a commodity recommendation graph model of all users;

inquiring commodity recommendation graph models of target users and commodity recommendation graph models of other users, and acquiring similar consumption users with the consumption similarity of L before sequencing with the target users in a preset time period; the similar consumption users are other users who purchase the same commodity with the target user, and the consumption similarity is the number of the same commodity purchased by the similar consumption users and the target user;

inquiring all the commodities purchased by the similar consumer users, removing the commodities purchased by the target user from the commodities, and obtaining a first commodity to be recommended;

sorting the first commodities to be recommended according to the number of the purchasers from more than one, selecting the first commodities to be recommended of M before sorting as second commodities to be recommended, and recommending the second commodities to be recommended of N before sorting as first commodities to be recommended to the target user;

calculating the interest intensity of the similar consumption user on the second commodity to be recommended, sorting the second commodity to be recommended according to the interest intensity from large to small, and recommending the second commodity to be recommended of R before sorting to the target user as the second commodity to be recommended.

Further, the historical consumption behavior data comprises user basic information, order basic information, user behavior data and commodity basic information.

Further, the calculating the interest intensity of the similar consumption user in the second commodity to be recommended includes:

calculating the interest intensity of each similar consumption user on the second commodity to be recommended according to the purchase quantity of each similar consumption user for purchasing the second commodity to be recommended;

and summing the interest intensity of all similar consumption users on the second commodity to be recommended to obtain the interest intensity of the similar consumption users on the second commodity to be recommended.

Further, the interest intensity I of each similar consumption user on the second commodity to be recommended is calculated according to the following formula:

i=log (number purchased) +1.

In a second aspect, an embodiment of the present invention further provides a commodity recommendation device based on a knowledge graph, including:

the diagram model construction module is used for acquiring historical consumption behavior data of all users and constructing commodity recommendation diagram models of all users;

the first acquisition module is used for inquiring the commodity recommendation graph model of the target user and commodity recommendation graph models of other users and acquiring similar consumption users L before the consumption similarity of the target user is ranked within a preset time period; the similar consumption users are other users who purchase the same commodity with the target user, and the consumption similarity is the number of the same commodity purchased by the similar consumption users and the target user;

the second acquisition module is used for inquiring all the commodities purchased by the similar consumer users, removing the commodities purchased by the target user from the commodities and acquiring a first commodity to be recommended;

the first recommending module is used for sequencing the first commodities to be recommended according to the number of the purchasers from more than one, selecting the first commodities to be recommended of M before sequencing as second commodities to be recommended, and recommending the second commodities to be recommended of N before sequencing as first commodities to the target user;

the second recommending module is used for calculating the interest intensity of the similar consumption user on the second commodity to be recommended, sequencing the second commodity to be recommended according to the interest intensity from big to small, and recommending the second commodity to be recommended of R before sequencing to the target user as the second commodity to be recommended.

i=log (number purchased) +1.

In a third aspect, an embodiment of the present invention further provides a commodity recommendation system based on a knowledge graph, including a processor, an input device, an output device, and a memory, where the processor, the input device, the output device, and the memory are connected to each other, where the memory is configured to store a computer program, and the computer program includes program instructions, and the processor is configured to invoke the program instructions to perform the method according to the first aspect.

In a fourth aspect, embodiments of the present invention also provide a computer readable storage medium storing a computer program comprising program instructions which, when executed by a processor, cause the processor to perform the method according to the first aspect.

As a simple example, the method provided by the embodiment of the invention only defines similar users from the same purchasing behavior, in addition, the product features, brand features, user labels and the like can be used for defining the user similarity, and are constructed into a graph model as key association entities of similarity recommendation, clustering is carried out through a clustering algorithm, and clustering conditions among different features are found in an unsupervised learning mode, so that accurate recommendation based on multidimensional features is realized.

Different from the relational database, the schema of the graph database has strong flexibility, directly reflects business logic, and analysts can flexibly add different kinds of new relations, new nodes and new labels to form new subgraphs according to scene and business requirement changes, so that new recommendation strategies can be dynamically adjusted without worrying about destroying the functions of the existing query or application programs. The model flexibility of the graph database avoids the trouble of the first-stage thinking of projects and the trouble of including each detail, so that a user can flexibly change a data model according to the business development of a company and the scene change of the customer, the efficient dynamic association of the customer, the product and the scene is realized, and the cost and the development period of system iteration are greatly reduced.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below.

FIG. 1 is a schematic diagram of a recommendation flow of a conventional recommendation system according to the background of the invention;

FIG. 2 is a schematic diagram of an architecture of a conventional recommendation system according to an embodiment of the present invention;

FIG. 3 is a flowchart of a commodity recommendation method based on a knowledge graph according to an embodiment of the present invention;

FIG. 4 is a schematic structural diagram of a commodity recommendation graph model according to an embodiment of the present invention;

FIG. 5 is a schematic view of user consumption preferences provided by an embodiment of the present invention;

FIG. 6 is a schematic diagram of similar user product recommendations provided by an embodiment of the present invention;

fig. 7 is a schematic structural diagram of a commodity recommendation device based on a knowledge graph according to an embodiment of the present invention;

fig. 8 is a schematic structural diagram of a commodity recommendation system based on a knowledge graph according to an embodiment of the present invention.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

It should be understood that the terms "comprises" and "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

It is also to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

It should be further understood that the term "and/or" as used in the present specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations.

As used in this specification and the appended claims, the term "if" may be interpreted as "when..once" or "in response to a determination" or "in response to detection" depending on the context. Similarly, the phrase "if a determination" or "if a [ described condition or event ] is detected" may be interpreted in the context of meaning "upon determination" or "in response to determination" or "upon detection of a [ described condition or event ]" or "in response to detection of a [ described condition or event ]".

It is noted that unless otherwise indicated, technical or scientific terms used herein should be given the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

With the advent of new retail and scene marketing age offline, the dimensions that the recommendation system needs to consider increase: time, location, weather, etc. can be important factors that affect the customer's shopping experience. The attention duration of the shopper is shorter and shorter, and potential customers are likely to lose if they cannot make multidimensional real-time recommendations to miss the appropriate opportunities according to their current behavior, whether online or offline.

The traditional recommendation system catches the fly to the elbow in the face of more frequently iterated product catalogs, more complex and multiple marketing scenes and recommendation dimensions, and more urgent recommendation real-time requirements. In a recommendation system, user behavior data for items is the most central data that directly indicates items that a user may be interested in. The existing architecture separates the characteristic information of the user and the product, and generates different characteristic information of the user/product according to a specific model, as shown in fig. 2.

When searching for a product which is likely to be interested by a certain customer, the system establishes a similarity matrix about the user and the product based on a recommendation algorithm of user behavior analysis, and selects an object set which is likely to be interested by the user. However, as the number of users and the number of articles are continuously enlarged, the cost of calculating the similarity matrix of the users/products by the system is very high, the timeliness of recommendation cannot be ensured, and instant update on the recommended content is difficult to achieve when the behavior of the users changes.

From data storage to recall and further to a sequencing link, the more concentrated the commodity set screened by the system is, the higher the accuracy is, the better the recommendation effect is, but the difficulty is increased. The recall speed is improved, products related to the current interests of customers and consumption scenes are found out from mass commodities in real time by a help system, and the method is an advantage of the graph database.

The key to the graph database technology is that it focuses not only on the stored data itself (users, commodities, categories, etc.), but also on the associations between stored data (what commodities the user has purchased, who the user likes, what the user has purchased before what).

The graph database may recommend customers based on various dimensions, such as customers, customer purchases of products, relationships between customers, and inherent correlations between products. For a given customer, the recommendation system can analyze the purchase behaviors of all customers so as to find other customers similar to the purchase behaviors, and generate a recommendation list of 'the customers buying the product also like … …'; and a recommendation list of 'your friends like also..you' can be generated for the target clients through the friend relation among the clients. More importantly, the graph database can simultaneously make recommendations in combination with multiple dimensions, such as "like …" with friends who have purchased similar products.

These similar customers and friends like products are also the most likely products of interest to the current customer. Through multi-hop associated query of the graph database, the recommendation system can extract required data in real time to complete quick recall. The recalled products are then ranked according to their relevance to the current customer. Through a recommendation algorithm based on the graph, the relevance of the product and the client can be accurately obtained by comprehensively comparing the distance (path length), the density (path number) and the preference degree (weight on the upper edge of the path) of the relation between the client and the product. Meanwhile, by properly weighting the specific type edges, the influence difference of the specific relationship in different recommendation scenes can be highlighted, for example, the friend relationship edges can be weighted to reflect the strong influence effect of the social relationship in the social marketing scene.

In the embodiment, shopping behaviors of the user are taken as recommendation dimensions, and how the personalized recommendation is performed by the recommendation system through the graph technology is demonstrated.

In a first aspect, as shown in fig. 3, an embodiment of the present invention provides a flowchart of a commodity recommendation method based on a knowledge graph. The method specifically comprises the following steps:

s100: and acquiring historical consumption behavior data of all users, and constructing a commodity recommendation graph model of all users.

Specifically, the historical consumption behavior data comprises user basic information, order basic information, user behavior data and commodity basic information. Wherein, the user basic information comprises user numbers, user names and the like; the order basic information comprises order numbers, transaction time and the like; the user behavior data comprise the user adding shopping carts, completing orders and the like; the commodity basic information includes commodity number, commodity name, price, class number, class name, and the like.

The constructed commodity recommendation graph model is shown in fig. 4, and the point types in the graph model are shown in table 1-1.

TABLE 1-1 Point types in the graph model

Point type	Attributes of
		User' s	User number, name
Goods commodity	Commodity number, name, price
		Order form	Order number
Products and products	Item number, name

The edge types in the graph model are shown in tables 1-2.

Table 1-2 edge types in graph model

Edge type	Type of starting point	Type of termination point	Attributes of
				Shopping cart	User' s	Goods commodity	Time and quantity of purchase
Completion of	User' s	Order form	Time of order
				Included	Order form	Goods commodity	Quantity of
Belonging to	Goods commodity	Products and products	/

The following are a set of examples of merchandise recommendations using the graphical model described above. Fig. 5 intercepts a partial data in the commodity recommendation graph model of the consumer's "mins", and the consumer ' mins ' consumption preference can be intuitively seen from fig. 5. The Xiaoming completes 3 orders in the double eleven activities of the electronic commerce (11 months 1 day to 11 months 11 days). First, the Xiaoming purchased two brands of toothbrushes in an order of 11 months and 9 days, and then continued to purchase some protective articles the next day, and finally purchased a bath ball on the double eleven days.

In addition to "small Ming", the system background may also have records of purchases by other hundreds of thousands of users. Then the recommender system may use the historical consumption behavior of other users to influence the future consumption decisions of the small.

S200: inquiring commodity recommendation graph models of target users and commodity recommendation graph models of other users, and acquiring similar consumption users with the consumption similarity of L before sequencing with the target users in a preset time period; the similar consumption users are other users who purchase the same commodity with the target user, and the consumption similarity is the quantity of the same commodity purchased by the similar consumption users and the target user.

When the user has a purchase intention but does not have an explicit purchasing goal, his attention can be captured by providing the user with goods of his interest. The recommendation system can make a recommendation for the user who purchases the same commodity in a large amount with a higher probability that the user is similar to the user in terms of living needs and personal interests.

Taking fig. 6 as an example, the purchase behavior of the small just and small red is similar to that of the small bright, and the recommendation system can recommend other commodities purchased by the small just and small red to the small bright, so that the attention of the small bright is attracted. The XiaoMing is a target user of the recommendation system, and the Xiaojust and the Xiaohong are similar consumption users with similar consumption behaviors with the XiaoMing.

In this example, the query obtains a population of users whose consumption behavior is similar to that of Ming in the past half year, and takes the top 10 users with the largest number of the same items purchased as similar consumption users of Ming.

S300: and inquiring all the commodities purchased by the similar consumption users, removing the commodities purchased by the target users from the commodities, and obtaining a first commodity to be recommended.

S400: and sorting the first commodities to be recommended according to the number of the purchasers from more than one, selecting the first commodities to be recommended of M before sorting as second commodities to be recommended, and recommending the second commodities to be recommended of N before sorting as first commodities to be recommended to the target user.

Specifically, the first commodity to be recommended is inquired and purchased by a plurality of similar consumers at the same time, and the first commodity to be recommended is ranked according to the number of the purchases from more to less. In this embodiment, the first commodity to be recommended before the first 10 is selected as the second commodity to be recommended, and the second commodity to be recommended before the first 3 is selected as the first commodity to be recommended to the target user.

S500: calculating the interest intensity of the similar consumption user on the second commodity to be recommended, sorting the second commodity to be recommended according to the interest intensity from large to small, and recommending the second commodity to be recommended of R before sorting to the target user as the second commodity to be recommended.

Specifically, the calculating the interest intensity of the similar consumer user in the second commodity to be recommended includes:

s501: and calculating the interest intensity of each similar consumption user in the second commodity to be recommended according to the purchase quantity of each similar consumption user in purchasing the second commodity to be recommended.

Specifically, the interest intensity I of each similar consumer to the second commodity to be recommended is calculated according to the following formula:

i=log (number purchased) +1.

S502: and summing the interest intensity of all similar consumption users on the second commodity to be recommended to obtain the interest intensity of the similar consumption users on the second commodity to be recommended.

In this embodiment, the second commodity to be recommended of the first 3 items is recommended to the target user as a second recommended commodity.

As can be seen from FIG. 6, as shown in tables 1-3, both small just and small red purchased K brand wet wipes and the number of purchases was high, so this was the preferred recommendation. Both the brand F liquid laundry detergent and the brand I liquid foundation are purchased by only one of them, but the liquid laundry detergent is purchased in a greater quantity than the liquid foundation, i.e., the target user may be interested in it more often, so it is also recommended preferentially. And finally, comprehensively reflecting the recommended degree of the product by the interestingness index.

Tables 1-3 Commodity query results based on similar user recommendations

Trade name	Total number of buyers	Total purchase quantity	Total interest level
				K brand wet tissue	2	6	4.20
F brand laundry detergent	1	3	2.10
				Brand I foundation liquid	1	2	1.69

As a simple example, the embodiment defines similar users only from the same purchasing behavior, in addition, product features, brand features, user labels and the like can be used for defining user similarity, and are constructed into a graph model as key associated entities of similarity recommendation, clustering is performed through a clustering algorithm, and clustering conditions among different features are found in an unsupervised learning mode, so that accurate recommendation based on multidimensional features is achieved.

In a second aspect, the embodiment of the invention also provides a commodity recommendation device based on the knowledge graph. As shown in fig. 7, the apparatus may include:

the graph model construction module 201 is configured to acquire historical consumption behavior data of all users, and construct a commodity recommendation graph model of all users;

the first obtaining module 202 is configured to query a commodity recommendation graph model of the target user and commodity recommendation graph models of other users, and obtain similar consuming users who are ranked with the consumption similarity of the target user in a preset time period; the similar consumption users are other users who purchase the same commodity with the target user, and the consumption similarity is the number of the same commodity purchased by the similar consumption users and the target user;

the second obtaining module 203 is configured to query all the commodities purchased by the similar consumer, and remove the commodities purchased by the target user from the commodities to obtain a first commodity to be recommended;

the first recommending module 204 is configured to sort the first articles to be recommended according to the number of the purchasers from more than one to less, select the first articles to be recommended of M before sorting as second articles to be recommended, and recommend the second articles to be recommended of N before sorting as first articles to the target user;

the second recommending module 205 is configured to calculate the interest intensity of the similar consumer user on the second to-be-recommended goods, rank the second to-be-recommended goods according to the interest intensity from large to small, and recommend the second to-be-recommended goods before ranking as the second to-be-recommended goods to the target user.

i=log (number purchased) +1.

Based on the same inventive concept, the embodiment of the invention provides a commodity recommendation system based on a knowledge graph. As shown in fig. 8, the system may include: one or more processors 101, one or more input devices 102, one or more output devices 103, and a memory 104, the processors 101, input devices 102, output devices 103, and memory 104 being interconnected by a bus 105. The memory 104 is used for storing a computer program comprising program instructions, the processor 101 being configured to invoke the program instructions for performing the method of the above-described knowledge-graph-based commodity recommendation method embodiment section.

It should be appreciated that in embodiments of the present invention, the processor 101 may be a central processing unit (Central Processing Unit, CPU), which may also be other general purpose processors, digital signal processors (Digital Signal Processor, DSPs), application specific integrated circuits (Application Specific Integrated Circuit, ASICs), off-the-shelf programmable gate arrays (Field-Programmable Gate Array, FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The input device 102 may include a keyboard or the like, and the output device 103 may include a display (LCD or the like), a speaker or the like.

The memory 104 may include read only memory and random access memory and provides instructions and data to the processor 101. A portion of the memory 104 may also include non-volatile random access memory. For example, the memory 104 may also store information of device type.

In a specific implementation, the processor 101, the input device 102, and the output device 103 described in the embodiments of the present invention may execute the implementation described in the embodiments of the commodity recommendation method based on the knowledge graph provided in the embodiments of the present invention, which is not described herein again.

It should be noted that, regarding the specific workflow of the commodity recommendation system based on the knowledge graph, reference may be made to the foregoing method embodiment section, and no further description is given here.

Further, an embodiment of the present invention also provides a readable storage medium storing a computer program, the computer program including program instructions that when executed by a processor implement: the commodity recommendation method based on the knowledge graph.

The computer readable storage medium may be an internal storage unit of the background server according to the foregoing embodiment, for example, a hard disk or a memory of the system. The computer readable storage medium may also be an external storage device of the system, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card) or the like, which are provided on the system. Further, the computer readable storage medium may also include both internal storage units and external storage devices of the system. The computer readable storage medium is used to store the computer program and other programs and data required by the system. The computer-readable storage medium may also be used to temporarily store data that has been output or is to be output.

Those of ordinary skill in the art will appreciate that the elements and algorithm steps described in connection with the embodiments disclosed herein may be embodied in electronic hardware, in computer software, or in a combination of the two, and that the elements and steps of the examples have been generally described in terms of function in the foregoing description to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention is essentially or a part contributing to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

While the invention has been described with reference to certain preferred embodiments, it will be understood by those skilled in the art that various changes and substitutions of equivalents may be made and equivalents will be apparent to those skilled in the art without departing from the scope of the invention. Therefore, the protection scope of the invention is subject to the protection scope of the claims.

Claims

1. The commodity recommendation method based on the knowledge graph is characterized by comprising the following steps of:

2. The method of claim 1, wherein the historical consumption behavior data comprises user basic information, order basic information, user behavior data, and commodity basic information.

3. The method for recommending commodities based on a knowledge graph according to claim 1, wherein said calculating the interest intensity of the similar consuming user in the second commodity to be recommended comprises:

4. The knowledge-graph-based commodity recommendation method as claimed in claim 3, wherein the interest intensity I of each similar consuming user in said second commodity to be recommended is calculated according to the following formula:

i=log (number purchased) +1.

5. A commodity recommendation device based on a knowledge graph, comprising:

6. The knowledge-based commodity recommendation apparatus as claimed in claim 5, wherein said historical consumption behavior data comprises user basic information, order basic information, user behavior data, commodity basic information.

7. The knowledge-based commodity recommendation apparatus according to claim 5, wherein said calculating the interest intensity of said similar consuming user in a second commodity to be recommended comprises:

8. The knowledge-based commodity recommendation apparatus according to claim 7, wherein the interest intensity I of each similar consuming user in said second commodity to be recommended is calculated according to the following formula:

i=log (number purchased) +1.

9. A knowledge-graph-based commodity recommendation system comprising a processor, an input device, an output device and a memory, the processor, the input device, the output device and the memory being interconnected, wherein the memory is adapted to store a computer program comprising program instructions, the processor being configured to invoke the program instructions to perform the method of any of claims 1-4.

10. A computer readable storage medium, characterized in that the computer readable storage medium stores a computer program comprising program instructions which, when executed by a processor, cause the processor to perform the method of any of claims 1-4.