CN112837118A

CN112837118A - Commodity recommendation method and device for enterprise users

Info

Publication number: CN112837118A
Application number: CN202110083712.4A
Authority: CN
Inventors: 尹玉申
Original assignee: Beijing Dianzhi Technology Co ltd
Current assignee: Beijing Dianzhi Technology Co ltd
Priority date: 2021-01-21
Filing date: 2021-01-21
Publication date: 2021-05-25
Anticipated expiration: 2041-01-21
Also published as: WO2022156529A1; CN112837118B

Abstract

The invention provides a commodity recommendation method and device for enterprise users, which can be applied to an enterprise purchasing platform, wherein the enterprise purchasing platform receives a purchasing demand input by a user, the purchasing demand comprises the industry to which an purchasing enterprise belongs and/or the application scene of purchased commodities, and the recommended commodities are obtained by matching according to the purchasing demand and a knowledge graph of a comprehensive commodity pool, the knowledge graph comprises the industry and/or the application scene of each commodity in the comprehensive commodity pool, and the recommended commodities are displayed to the user. The method recommends the required commodity for the enterprise user from the comprehensive commodity pool according to the industry to which the purchasing enterprise belongs and/or the application scene of the purchased commodity, and can accurately recommend the required commodity for the enterprise user.

Description

Commodity recommendation method and device for enterprise users

Technical Field

The invention relates to the internet technology, in particular to a commodity recommendation method and device for enterprise users.

Background

Enterprise purchasing is the most main purchasing mode in the market economy nowadays and is characterized by purchasing large quantities of commodities. In order to facilitate enterprise purchasing and improve enterprise purchasing efficiency, a supplier or a third-party manufacturer provides a purchasing platform for an enterprise to selectively purchase purchased commodities, the enterprise purchasing platform can provide a differentiated and customized exclusive comprehensive purchasing solution for enterprise users, and a common purchasing platform is, for example, Beijing Dong Huacai.

Taking Beijing Dong Hui collection as an example, an enterprise user inputs a commodity code or a keyword to inquire a required commodity through a search function provided by a platform, the platform searches commodities in a commodity pool corresponding to the enterprise user according to search information input by the user, different enterprises have different commodity pools, if the commodities are searched in the commodity pool corresponding to the enterprise user, a search result is displayed to the user in a commodity list mode, if the commodities are not searched, the user needs to inform an operator of the platform of a purchase demand through a telephone, the operator of the platform adds the commodities which possibly meet the purchase demand of the user into the commodity pool corresponding to the user according to the purchase demand of the user, and the user can purchase the commodities on order.

However, in many cases, the user cannot search for the desired item to be purchased.

Disclosure of Invention

The invention provides a commodity recommendation method and device for enterprise users, which are used for solving the problem that commodities cannot be recommended for the enterprise users through an enterprise purchasing platform in the prior art.

In a first aspect, the present invention provides a method for recommending a commodity for an enterprise user, including:

receiving a purchasing requirement input by a user, wherein the purchasing requirement comprises an industry to which a purchasing enterprise belongs and/or an application scene of purchasing commodities;

according to the purchasing demand and a knowledge graph of the comprehensive commodity pool, matching to obtain recommended commodities, wherein the knowledge graph comprises industries and/or application scenes of all commodities in the comprehensive commodity pool;

and displaying the recommended commodity to a user.

Optionally, the method further includes:

determining a first commodity which does not belong to a commodity pool corresponding to the purchasing enterprise from the recommended commodities, wherein the commodity pool corresponding to the purchasing enterprise is a subset of the comprehensive commodity pool;

adding a second commodity into a commodity pool corresponding to the purchasing enterprise, wherein the second commodity belongs to the first commodity;

and receiving the purchasing operation of the user on the commodities in the commodity pool corresponding to the purchasing enterprise.

Optionally, the adding the second commodity into the commodity pool corresponding to the purchasing enterprise includes:

displaying an add control for the first item;

receiving a first operation of a user on an adding control of the second commodity;

and adding the second commodity into a commodity pool corresponding to the purchasing enterprise according to the first operation.

Optionally, the method further includes:

collecting historical order data, wherein the historical order data comprises identification of commodities, industries to which purchasing enterprises of historical orders belong and/or application scenes of the historical orders;

determining the industry of the commodity in the historical order according to the industry to which the purchasing enterprise of the historical order belongs;

determining the industry of the commodities in the comprehensive commodity pool according to the industry of the commodities in the historical order;

determining the application scene of the commodities in the comprehensive commodity pool according to the application scene of the commodities in the historical order;

and constructing the knowledge graph according to the industry and application scenes of the commodities in the comprehensive commodity pool.

Optionally, the method further includes:

performing feature extraction on the historical orders of the commodities in the comprehensive commodity pool through a machine learning method to obtain features of purchased commodities of all industries and features of purchased commodities of all enterprises, wherein the features of the purchased commodities comprise at least one of the following data: the category of the purchased commodity, the price of the purchased commodity, the combinability of the purchased commodity and the repurchase rate of the purchased commodity;

and updating the knowledge graph according to the characteristics of the purchased commodities in each industry.

Optionally, the matching according to the purchasing demand and the knowledge graph of the comprehensive commodity pool to obtain recommended commodities includes:

and matching to obtain recommended commodities according to the characteristics of the purchased commodities of the purchasing enterprise, the purchasing demand and the knowledge graph of the comprehensive commodity pool.

Optionally, the knowledge graph further includes historical trading total amount of the commodity and historical repurchase rate of the commodity.

Optionally, the matching according to the characteristics of the purchased commodities of the purchasing enterprise, the purchasing demand and the knowledge graph of the comprehensive commodity pool to obtain recommended commodities includes:

matching the industry to which the purchasing enterprise belongs and/or the application scene of the purchased commodity with the industry and/or the scene of each commodity in the knowledge graph to obtain a first candidate commodity;

matching from the comprehensive commodity pool according to the characteristics of the purchased commodities of the purchasing enterprise to obtain a second candidate commodity;

obtaining a third candidate commodity according to the historical trading total amount of each commodity and the historical repurchase rate of the commodity in the knowledge graph;

and obtaining the recommended commodity according to the first candidate commodity, the second candidate commodity and the third candidate commodity.

In a second aspect, the present invention provides a method for recommending a commodity for an enterprise user, including:

when no commodity is searched in a commodity pool corresponding to a purchasing enterprise or the number of searched commodities is less than a preset value according to a keyword of the purchasing commodity input by a user, extracting attribute information of the purchasing commodity from the keyword of the purchasing commodity;

matching to obtain recommended commodities according to the attribute information of the purchased commodities and a knowledge graph of a comprehensive commodity pool, wherein the knowledge graph comprises the attribute information of each commodity in the comprehensive commodity pool, and the commodity pool corresponding to the purchasing enterprise is a subset of the comprehensive commodity pool;

and displaying the recommended commodity to a user.

Optionally, the method further includes:

determining a first commodity which does not belong to a commodity pool corresponding to the purchasing enterprise from the recommended commodities;

displaying an add control for the first item;

Optionally, the method further includes:

extracting the attributes and attribute values of the commodities in the comprehensive commodity pool;

correcting the attributes of the commodities;

clustering and fusing the corrected attributes of the commodities to obtain target attributes of the commodities;

determining a form of an attribute value of the target attribute;

and constructing the knowledge graph according to the target attribute and the attribute value of the commodity.

Optionally, the knowledge graph further includes historical transaction total amount of the commodity and the repurchase rate of the commodity.

Optionally, the matching according to the attribute information of the purchased commodity and the knowledge graph of the comprehensive commodity pool to obtain the recommended commodity includes:

matching the attribute information of the purchased commodity with the attribute information of each commodity in the knowledge graph to obtain a first candidate commodity;

obtaining a second candidate commodity according to the historical trading total amount of each commodity and the repurchase rate of the commodity in the knowledge graph;

and obtaining the recommended commodity according to the first candidate commodity and the second candidate commodity.

Optionally, the method further includes:

performing feature extraction on the historical orders of the commodities in the comprehensive commodity pool through a machine learning method to obtain features of purchased commodities of each enterprise, wherein the features of the purchased commodities comprise at least one of the following data: the category of the purchased commodity, the price of the purchased commodity, the combinability of the purchased commodity and the repurchase rate of the purchased commodity;

the matching according to the attribute information of the purchased commodity and the knowledge graph of the comprehensive commodity pool to obtain the recommended commodity comprises the following steps:

matching to obtain a third candidate commodity from the comprehensive commodity pool according to the characteristics of the purchased commodity of the purchasing enterprise;

and obtaining the recommended commodity according to the first candidate commodity and the third candidate commodity.

In a third aspect, the present invention provides a commodity recommendation apparatus for an enterprise user, including:

the system comprises a receiving module, a judging module and a display module, wherein the receiving module is used for receiving a purchasing demand input by a user, and the purchasing demand comprises the industry of a purchasing enterprise and/or the application scene of a purchasing commodity;

the matching module is used for matching to obtain recommended commodities according to the purchasing demands and a knowledge graph of the comprehensive commodity pool, wherein the knowledge graph comprises industries and/or application scenes of all commodities in the comprehensive commodity pool;

and the display module is used for displaying the recommended commodities to the user.

In a fourth aspect, the present invention provides a commodity recommendation apparatus for an enterprise user, including:

the attribute extraction module is used for extracting attribute information of the purchased commodities from the keywords of the purchased commodities when the commodities cannot be searched in a commodity pool corresponding to a purchased enterprise according to the keywords of the purchased commodities input by the user or the number of the searched commodities is less than a preset value;

the matching module is used for matching to obtain recommended commodities according to the attribute information of the purchased commodities and a knowledge graph of the comprehensive commodity pool, the knowledge graph comprises the attribute information of each commodity in the comprehensive commodity pool, and the commodity pool corresponding to the purchasing enterprise is a subset of the comprehensive commodity pool;

In a fifth aspect, the present invention provides an electronic device, comprising: at least one processor and memory;

the memory stores computer-executable instructions;

the at least one processor executing the computer-executable instructions stored by the memory causes the at least one processor to perform the method according to any one of the first, second and alternative aspects of the invention.

In a sixth aspect, the present invention provides a computer readable storage medium having stored thereon computer executable instructions for implementing a method according to any one of the first, second and alternative aspects of the present invention when executed by a processor.

In a seventh aspect, the present invention provides a computer program product comprising a computer program which, when executed by a processor, performs the method of any one of the first, second and alternative aspects of the present invention.

The commodity recommendation method and device for the enterprise users can be applied to an enterprise purchasing platform, the enterprise purchasing platform receives purchasing demands input by users, the purchasing demands comprise industries to which purchasing enterprises belong and/or application scenes of purchased commodities, recommended commodities are obtained through matching according to the purchasing demands and a knowledge graph of a comprehensive commodity pool, the knowledge graph comprises industries and/or application scenes of all commodities in the comprehensive commodity pool, and the recommended commodities are displayed to the users. The method recommends the required commodity for the enterprise user from the comprehensive commodity pool according to the industry to which the purchasing enterprise belongs and/or the application scene of the purchased commodity, and can accurately recommend the required commodity for the enterprise user.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure.

Fig. 1 is a flowchart of a commodity recommendation method for an enterprise user according to an embodiment of the present invention;

FIG. 2 is a schematic view of a merchandise search page;

FIG. 3 is a schematic diagram of a smart referral selection page;

FIG. 4 is a schematic diagram of an advanced screening interface for recommended merchandise;

fig. 5 is a flowchart of a commodity recommendation method for an enterprise user according to a second embodiment of the present invention;

FIG. 6 is a diagram illustrating a result of merchandise recommendation;

fig. 7 is a flowchart of a commodity recommendation method for an enterprise user according to a third embodiment of the present invention;

fig. 8 is a flowchart of a commodity recommendation method for an enterprise user according to a fourth embodiment of the present invention;

FIG. 9 is a schematic diagram of attribute information of a product;

fig. 10 is a flowchart of a commodity recommendation method for an enterprise user according to a fifth embodiment of the present invention;

fig. 11 is a flowchart of a commodity recommendation method for an enterprise user according to a sixth embodiment of the present invention;

fig. 12 is a schematic structural diagram of a commodity recommendation device for an enterprise user according to a seventh embodiment of the present invention;

fig. 13 is a schematic structural diagram of a commodity recommendation device for an enterprise user according to an eighth embodiment of the present invention;

fig. 14 is a schematic structural diagram of an electronic device according to a ninth embodiment of the present application.

With the foregoing drawings in mind, certain embodiments of the disclosure have been shown and described in more detail below. These drawings and written description are not intended to limit the scope of the disclosed concepts in any way, but rather to illustrate the concepts of the disclosure to those skilled in the art by reference to specific embodiments.

Detailed Description

Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the exemplary embodiments below are not intended to represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.

The invention provides a commodity recommendation method which is applied to an enterprise purchasing platform, is convenient for enterprises to purchase commodities and improves purchasing efficiency of the enterprises.

Fig. 1 is a flowchart of a commodity recommendation method for an enterprise user according to an embodiment of the present invention, and as shown in fig. 1, the method according to this embodiment includes the following steps:

s101, receiving a purchasing requirement input by a user, wherein the purchasing requirement comprises the industry of a purchasing enterprise and/or the application scene of a purchasing commodity.

The enterprise procurement platform provides the item search page shown in fig. 2, which includes a search box 11a, an intelligent referral control 11b, my shopping cart 11c, and a list of all items 11 d. The search box 11a is used for inputting keywords of commodities to search, the intelligent recommended item control 11b is used for triggering intelligent search, the my shopping cart 11c is used for storing commodities purchased by a user, and the whole commodity list 11d is used for viewing a commodity pool corresponding to an enterprise.

In the enterprise purchasing platform, a corresponding commodity pool is set for each logged-in enterprise user, the commodity pools corresponding to different enterprise users are different, the commodity pool corresponding to an enterprise is configured for the enterprise when the enterprise applies for registration, and after each enterprise user logs in the enterprise purchasing platform, only commodities in the commodity pool can be seen. The categories or kinds of commodities in the commodity pool corresponding to the enterprise are displayed in the whole commodity list 11d, and the categories of commodities include food, beverages, mobile phone communication, household appliances, computer office, household cleaning/paper products, kitchen ware, personal care, furniture, mother and infant, clothes, industrial products and the like.

After the user clicks the intelligent recommendation control 11b, the intelligent recommendation selection page shown in fig. 3 is displayed and overlaid on the page shown in fig. 2, or the user jumps to the intelligent recommendation selection page shown in fig. 3, as shown in fig. 3, the page displays industry information and a purchasing application scene, the industry information includes a plurality of industry controls for the user to select, for example, government, operator, finance, energy, message and internet, manufacturing industry, and the like, the user selects the industry to which the purchasing enterprise belongs by clicking the industry controls, and optionally, one or more industries can be selected for the purchasing enterprise.

The purchase application scene comprises a plurality of application scenes for a user to select, and the application scenes comprise office collection, marketing, festival and welfare, labor protection and protection, birthday care, instant incentive and the like. Optionally, the application scenario may further include a plurality of sub-scenarios, for example, a plurality of festivals are included in the annual festival welfare: the method comprises the following steps that after a user clicks a certain application scene, if a plurality of sub-scenes are included in the application scene, the sub-scenes are displayed to the user for the user to select.

After selecting the industry to which the purchasing enterprise belongs and/or the application scene of the purchased commodity, the user clicks the intelligent recommended commodity viewing control shown in fig. 3, and then the purchasing platform of the enterprise generates a purchasing demand.

And S102, matching to obtain recommended commodities according to the purchasing demand and a knowledge graph of the comprehensive commodity pool, wherein the knowledge graph comprises industries and/or application scenes for obtaining various commodities in the comprehensive commodity pool.

The knowledge graph is constructed according to historical order data of enterprises, and the industries and/or application scenes of the commodities in the comprehensive commodity pool are obtained by obtaining the industries to which the enterprises belong in the historical orders and the application scenes of the commodities in the historical orders. The comprehensive commodity pool is the largest commodity pool provided by the enterprise purchasing platform, the comprehensive commodity pool does not belong to or correspond to a certain enterprise, and the commodity pool corresponding to each enterprise is a subset of the comprehensive commodity pool.

The method can add the industry and application scenes of the commodities in the information of the commodities to obtain the knowledge map, and can understand that each commodity can have one or more industries and also can have one or more application scenes.

When an item has one or more industries, a weight may be defined for each industry, with the corresponding weight reflecting the likelihood that the item will be purchased in that industry. Similarly, when a commodity has a plurality of application scenes, each application scene also has a weight, and the weight corresponding to the application scene reflects the possibility that the commodity is purchased in the application scene.

Optionally, the knowledge graph may be updated, for example, as time changes, new application scenarios or application scenario changes may occur for the goods, or the goods may be applied to more industries, at which time application scenarios or industries may be added, deleted, or modified for the goods in the knowledge graph.

And matching the application scenes of the industries and/or purchased commodities of the purchasing enterprises included in the purchasing demands with the industries and/or application scenes of the commodities in the knowledge graph to obtain recommended commodities. And if the industry and the application scene of a certain commodity in the knowledge graph are the same as those in the purchasing requirement, the commodity is a recommended commodity.

And S103, displaying the recommended commodity to the user.

The enterprise purchasing platform displays the recommended commodities to the user, and the platform can sort the recommended commodities and then display the sorted recommended commodities to the user. Optionally, the knowledge graph may further include historical total volume of trades of the commodities and/or historical rate of repurchasing of the commodities, when the recommended commodities are sorted, the recommended commodities may be sorted according to the historical total volume of trades of the commodities and/or the historical rate of repurchasing of the commodities, the commodities with high historical total volume of trades and the historical rate of repurchasing of the commodities are hot commodities generally, the hot commodities are preferentially displayed to the user, the user can find the commodity required to be purchased, and user experience is improved.

Optionally, the characteristics of the purchased commodities of each enterprise can be acquired, and then the commodities with high enterprise purchasing possibility can be preferentially displayed to the user according to the characteristics of the purchased commodities of the enterprise, so that the user can find the commodities required to be purchased, and the user experience is improved.

In the embodiment, commodities are recommended for the user from the comprehensive commodity pool according to the industry to which the purchasing enterprise belongs and/or the application scene of the purchased commodities, so that the condition that the commodities cannot be matched can be avoided.

The quantity of recommended commodities is possibly large, so that enterprise users can conveniently and quickly find required products, optionally, a high-level screening function is provided, and the users can find the commodities required to be purchased from the recommended products through the high-level screening function. Fig. 4 is a schematic diagram of an advanced screening interface of recommended commodities, and as shown in fig. 4, a user can select the following conditions from the recommended commodities through an advanced screening function: inventory of goods, user preferences, invoice range, goodness, number of picks, etc. The commodity inventory comprises an unlimited plurality of province options, the user preference comprises a gender option and an age option, and the age option comprises an unlimited plurality of age group options. The invoice range is not limited, the options of issuing value-added invoice commodities and the like are supported, and the good rating rate comprises not limited and a plurality of rating range options.

The user can quickly find the required purchased commodities from a large number of recommended commodities through options in the advanced screening function, and the purchasing efficiency is improved.

In this embodiment, the enterprise purchasing platform receives a purchasing requirement input by a user, where the purchasing requirement includes an industry to which the purchasing enterprise belongs and/or an application scenario of the purchased commodity, and matches the purchasing requirement and a knowledge graph of the comprehensive commodity pool to obtain a recommended commodity, where the knowledge graph includes the industry and/or the application scenario of each commodity in the comprehensive commodity pool, and displays the recommended commodity to the user. The method recommends the required commodity for the enterprise user from the comprehensive commodity pool according to the industry to which the purchasing enterprise belongs and/or the application scene of the purchased commodity, and can accurately recommend the required commodity for the enterprise user.

Fig. 5 is a flowchart of a product recommendation method for an enterprise user according to a second embodiment of the present invention, and as shown in fig. 5, the method according to this embodiment further includes the following steps based on the first embodiment:

and S104, determining the first commodity which does not belong to the commodity pool corresponding to the purchasing enterprise from the recommended commodities, wherein the commodity pool corresponding to the purchasing enterprise is a subset of the comprehensive commodity pool.

The recommended commodity is recommended from the comprehensive commodity pool, so that part of the recommended commodity may not belong to the commodity pool corresponding to the purchasing enterprise. In the embodiment of the application, each enterprise user can only purchase the commodities in the commodity pool of the enterprise, so that the user can only purchase the commodities by adding the first commodities, which do not belong to the commodity pool corresponding to the purchasing enterprise, in the commodity pool corresponding to the purchasing enterprise.

And S105, adding the second commodity into a commodity pool corresponding to the purchasing enterprise, wherein the second commodity belongs to the first commodity.

In one implementation, after determining the first item, the enterprise procurement platform automatically adds the first item to a pool of items corresponding to the procurement enterprise. In another implementation manner, an adding control of a first commodity is displayed, a first operation of a user on the adding control of a second commodity is received, the first operation can be a touch operation, a click operation or a double-click operation on the adding control, and the enterprise purchasing platform adds the second commodity into a commodity pool corresponding to a purchasing enterprise according to the first operation. In the latter implementation manner, part of the first goods can be added to the goods pool corresponding to the purchasing enterprise according to the selection of the user, for example, when the user needs to purchase a certain goods, the user finds that the goods is not in the goods pool, and adds the goods to the goods pool through the above operation.

Fig. 6 is a schematic diagram of a commodity recommendation result, and as shown in fig. 6, a "apply for adding pool" control is not displayed on a purchase interface of a commodity in a commodity pool corresponding to a purchasing enterprise, and after a user clicks the "apply for adding pool" control, the commodity is added to the commodity pool corresponding to the purchasing enterprise.

And S106, receiving the purchasing operation of the user on the commodities in the commodity pool corresponding to the purchasing enterprise.

In the embodiment, the user can add the commodity to be purchased into the commodity pool corresponding to the purchasing enterprise through the display interface of the recommended commodity, so that the operation of adding the commodity into the commodity pool is simplified. In the prior art, an enterprise user cannot add commodities into a commodity pool, and needs an operator of an enterprise purchasing platform to add the commodities into the commodity pool in the background, so that the operation is complex.

Fig. 7 is a flowchart of a product recommendation method for an enterprise user according to a third embodiment of the present invention, where the method in this embodiment is used to construct a knowledge graph of an integrated product pool, and steps in this embodiment may be executed before step S101 in the first embodiment, as shown in fig. 7, the method in this embodiment includes the following steps:

s201, historical order data are collected, wherein the historical order data comprise identification of commodities, industries to which purchasing enterprises of historical orders belong and/or application scenes of the historical orders.

The identification of the commodity in the historical order can be the name of the commodity or the number of the commodity, and the like, wherein the industry to which the purchasing enterprise of the historical order belongs and/or the application scene of the historical order are added for constructing the knowledge graph. Optionally, the historical order data further includes the quantity of the goods purchased by the user, the amount of the goods purchased by the user, and the like.

S202, determining the industry of the commodities in the historical order according to the industry of the purchasing enterprise of the historical order.

S203, determining the industry of the commodities in the comprehensive commodity pool according to the industry of the commodities in the historical order.

For the same commodity or the same category of commodities, the industries are the same, for example, the industry of the mobile phone of the manufacturer A in the historical order is consumption and internet, and then the industries of the mobile phones of different manufacturers in the comprehensive commodity pool are determined to be consumption and internet industries. Or determining the industries of other commodities of the same category as the mobile phone as the industries of adding consumption and the Internet.

And S204, determining the application scene of the commodities in the comprehensive commodity pool according to the application scene of the commodities in the historical order.

The application scenes of the commodities in the comprehensive commodity pool can be determined according to the commodity classification categories and the application scenes of the commodities in the historical orders. Specifically, the category of the commodity in the historical order is determined, and then the application scene of the commodity of the same category in the comprehensive commodity pool is determined as the application scene of the commodity in the historical order.

And S205, constructing a knowledge graph according to the industry and application scenes of the commodities in the comprehensive commodity pool.

The industry and application scenes of the commodities in the comprehensive commodity pool are obtained through the steps S203 and S204, and the commerce and application scenes of the commodities are added into the information of the commodities to obtain the knowledge graph.

And S206, extracting the characteristics of the historical orders of the commodities in the comprehensive commodity pool by a machine learning method to obtain the characteristics of the purchased commodities in each industry and the characteristics of the purchased commodities in each enterprise.

And S207, updating the knowledge graph according to the characteristics of the purchased commodities in each industry.

Over time, the industries of the goods in the knowledge-graph may change, for example, a good is purchased in large quantities in a new industry, and the purchase in the previous industry is less heavily weighted, so that the information of the good in the knowledge-graph needs to be updated.

Alternatively, the knowledge graph may be updated periodically or event-triggered by steps S206 and S207. Wherein the characteristics of the industry or the user purchasing the commodity comprise at least one of the following data: category of the purchased goods, price of the purchased goods, combinability of the purchased goods, and repurchase rate of the purchased goods.

The price of the purchased product may include historical total Volume (GMV) of the product, and GMV generally refers to the total Volume of the product in a period of time, for example, the total Volume of a certain skin care product in 10 months and 15-30 days. The repurchase rate of the good may be the number of times the good is repeatedly purchased (also referred to as the repurchase rate), with more repurchase of the good reflecting greater customer loyalty to the brand and the repurchase rate of the good also reflecting the thermal digestibility of the good.

The combination performance of the commodities can reflect the relationship among the commodities, and some commodities need to be purchased simultaneously due to use requirements when purchased, for example, the combination of a computer and a mouse and the combination of a desk and a chair, so that the required commodities can be recommended to the user according to the combinability of the commodities, for example, when the user purchases a computer, the user recommends the computer and the mouse to the user simultaneously, and the recommended commodities are more targeted.

After the characteristics of the purchased commodities in each industry are obtained, the industry of each commodity in the knowledge map can be updated according to the characteristics of the purchased commodities in each industry. For example, the purchase price of the industry a for the commodity 001 is gradually increased according to the repurchase rate of the purchased commodity of the industry and/or the price data of the purchased commodity, and the purchase price is larger than the set threshold, and meanwhile, the repurchase rate of the industry a for the commodity 001 is increased, and the repurchase rate is larger than the set threshold, the industry a is increased to the industry of the commodity 001 in the knowledge graph. If the purchase price of the commodity 001 by the industry B is reduced and the repurchase rate is reduced, the weight of the industry B of the commodity 001 in the knowledge graph is reduced, or the industry B of the commodity 001 in the knowledge graph is deleted.

Or, if the combination of the commodities purchased in each industry is found to be that the commodities 002 and 003 are often sold in combination, the industry of the commodity 003 in the knowledge graph is updated according to the industry of the commodity 002, namely the industry of the commodity 002 is increased to the industry of the commodity 003.

Or according to the categories of the commodities purchased in each industry, if the commodities purchased in the industry A are found to be added with a new category compared with the commodities purchased in the prior industry, the scene of the commodities corresponding to the new category in the knowledge graph is updated.

If the characteristics of the purchased commodities of each enterprise are obtained through the machine learning algorithm in step S206, the recommended commodities can be obtained through matching according to the characteristics of the purchased commodities of the purchasing enterprise, the purchasing demand and the knowledge map of the comprehensive commodity pool when matching the commodities.

The characteristics of the purchasing enterprise for purchasing the goods include at least one of the following data: category of the purchased goods, price of the purchased goods, combinability of the purchased goods, and repurchase rate of the purchased goods. By combining the characteristics of purchasing the commodities of the enterprise for recommending the commodities for the enterprise users, the accuracy of recommending the commodities can be improved, and the satisfaction degree of the enterprise users can be improved.

Illustratively, matching application scenes of industries and/or purchased commodities to which a purchasing enterprise belongs with industries and/or scenes of commodities in a knowledge graph to obtain a first candidate commodity, then matching from a comprehensive commodity pool according to characteristics of the purchased commodities of the purchasing enterprise to obtain a second candidate commodity, and obtaining a recommended commodity by taking a union of the first candidate commodity and the second candidate commodity. Or matching the industry to which the purchasing enterprise belongs and/or the application scene of the purchased commodities with the industry and/or scene of each commodity in the knowledge graph to obtain a first candidate commodity, and when the number of the first candidate commodities is large, matching the first candidate commodity with the first candidate commodity according to the characteristics of the purchased commodities of the purchasing enterprise to obtain a recommended commodity.

Optionally, the knowledge graph further includes historical transaction total amount of the commodity and historical repurchase rate of the commodity, and correspondingly, according to the characteristics of the purchased commodity of the purchasing enterprise, the purchasing demand and the knowledge graph of the comprehensive commodity pool, the recommended commodity is obtained by matching, and may be: matching the industry to which the purchasing enterprise belongs and/or the application scene of the purchased commodity with the industry and/or scene of each commodity in the knowledge graph to obtain a first candidate commodity; matching from the comprehensive commodity pool according to the characteristics of the purchased commodities of the purchasing enterprise to obtain a second candidate commodity; obtaining a third candidate commodity according to the historical trading total amount of each commodity and the historical repurchase rate of the commodity in the knowledge map; and obtaining a recommended commodity according to the first candidate commodity, the second candidate commodity and the third candidate commodity. For example, a union of the first candidate item, the second candidate item, and the third candidate item may be taken as the recommended item.

Obtaining a third candidate commodity according to the historical trading total amount of each commodity and the historical repurchase rate of the commodity in the knowledge graph, wherein the third candidate commodity can be obtained by: according to the knowledge graph, some commodities with larger historical transaction total amount and larger historical repurchase rate are selected from the comprehensive commodity pool as second candidate products according to the historical transaction total amount and the historical repurchase rate, the commodities with larger historical transaction total amount and larger historical repurchase rate are usually some popular commodities, and the popular commodities can be recommended to the user together in the embodiment.

In another implementation manner, the application scenarios of industries and/or purchased commodities to which the purchasing enterprise belongs are matched with the industries and/or scenarios of the commodities in the knowledge graph to obtain a first candidate commodity, further, according to the characteristics of the purchased commodities of the purchasing enterprise, a fourth candidate commodity is obtained from the first candidate commodity in a matching manner, according to the historical transaction total amount and the historical repurchase rate of the commodities, a fifth candidate commodity is obtained from the first candidate commodity in a matching manner, and the union of the fourth candidate commodity and the fifth candidate commodity is taken as a recommended commodity.

In this embodiment, historical order data is collected, where the historical order data includes an identification of a commodity, an industry to which a purchasing enterprise of a historical order belongs, and/or an application scenario of the historical order, the industry of the commodity in the historical order is determined according to the industry to which the purchasing enterprise of the historical order belongs, the industry of the commodity in the historical order is determined according to the industry of the commodity in the historical order, the application scenario of the commodity in the comprehensive commodity pool is determined according to the application scenario of the commodity in the historical order, and a knowledge graph is constructed according to the industry and the application scenario of the commodity in the comprehensive commodity pool. And the historical orders of the commodities in the comprehensive commodity pool are subjected to feature extraction through a machine learning method to obtain the features of the commodities purchased in each industry, the knowledge map is updated according to the features of the commodities purchased in each industry, and the accuracy of the commodities recommended to the user can be improved by continuously updating the knowledge map.

Fig. 8 is a flowchart of a commodity recommendation method for an enterprise user according to a fourth embodiment of the present invention, and as shown in fig. 8, the method according to this embodiment includes the following steps:

s301, when no commodity is searched in a commodity pool corresponding to a purchasing enterprise according to the keyword of the purchasing commodity input by the user or the number of the searched commodities is less than a preset value, extracting attribute information of the purchasing commodity from the keyword of the purchasing commodity.

When a user inputs a keyword search in a search box according to the scheme in the prior art, an enterprise purchasing platform searches commodities in a commodity pool corresponding to a purchasing enterprise according to the keyword input by the user, and the commodities cannot be searched by the user due to the fact that the quantity of the commodities in the commodity pool corresponding to the purchasing enterprise is small or the keyword input by the user is inaccurate, or the quantity of the commodities searched by the user is small, the method of the embodiment can be adopted for commodity recommendation.

The keyword for purchasing the commodity is used for describing a purchasing requirement, can be a specific commodity name, and can also include description information of the commodity, for example, if the keyword for the commodity input by the user is 'Walungting black beer refined German original package import', a sequence labeling model of a cascade neural network can be adopted for extracting attribute information, and the attribute information is also called as attribute characteristics.

Sequence annotation, which is the most common task in natural language processing sequences, is a sequence-to-sequence process in which, given an input sequence, each position in the sequence is labeled with a corresponding label using a model. The sequence labeling model of the cascade neural network can perform sequence labeling through the following processes: word embedding, character embedding, value extraction, and attribute classification.

Illustratively, the attribute information of the commodity obtained after labeling the keyword "warringin black beer refined German original package import" through the sequence labeling model includes a class, a brand, an origin, an import, and the like, wherein the class is "beer", "the brand" is "warringin", "the origin" is "Germany", and the import "is" import ".

And S302, according to the attribute information of the purchased commodities and a knowledge graph of the comprehensive commodity pool, matching to obtain recommended commodities, wherein the knowledge graph comprises the attribute information of each commodity in the comprehensive commodity pool, and the commodity pool corresponding to the purchasing enterprise is a subset of the comprehensive commodity pool.

The comprehensive commodity pool comprises a large number of commodities, one or more attributes are defined for each commodity in the knowledge graph, the attributes of commodities of different categories or categories may be different, for example, the attributes of wine comprise a brand, a place of origin and an import, the attributes of a mobile phone comprise a brand, a model, a memory, a camera, a shop and the like, and the attributes of a computer comprise a brand, a CPU model, a memory and a hard disk.

Fig. 9 is a schematic diagram of attribute information of a commodity, as shown in fig. 9, the trade name is AAAV5, the attributes of the commodity include a category, a brand, a model, a memory, a camera, and a store, and each attribute corresponds to an attribute value, for example, the attribute value of the category is a mobile phone, the attribute value of the brand is a1, the attribute value of the model is a V5 memory is 256G, the attribute value of the camera is 5000 ten thousand pixels, and the attribute value of the store is B1 official.

Matching the attribute information of the purchased commodities with the attribute information of each commodity in the comprehensive commodity pool, determining the commodities with the similarity larger than a set threshold value as recommended commodities by calculating the similarity between the attributes of the purchased commodities and the attributes of the commodities in the comprehensive upper flat pool, and sequencing the recommended commodities by adopting a BM25 algorithm or other algorithms.

303. And displaying the recommended commodity to the user.

The commodities displayed to the user are sorted commodities, so that the user can be helped to quickly find the commodity required to be purchased.

In this embodiment, when no commodity is searched in a commodity pool corresponding to a purchasing enterprise according to a keyword of a purchased commodity input by a user or the number of searched commodities is less than a preset value, attribute information of the purchased commodity is extracted from the keyword of the purchased commodity, a recommended commodity is obtained by matching according to the attribute information of the purchased commodity and a knowledge graph of a comprehensive commodity pool, the knowledge graph includes attribute information of each commodity in the comprehensive commodity pool, and the recommended commodity is displayed to the user. The method can recommend the commodities to the user according to the attribute information, avoids the condition that the commodities cannot be recommended to the user because the commodities cannot be searched in the commodity pool corresponding to the purchasing enterprise in the prior art, can recommend the required commodities to the user, and improves the user experience.

Fig. 10 is a flowchart of a commodity recommendation method for an enterprise user according to a fifth embodiment of the present invention, and as shown in fig. 10, the method according to this embodiment further includes the following steps based on the fourth embodiment:

and S304, determining the first commodity which does not belong to the commodity pool corresponding to the purchasing enterprise from the recommended commodities, wherein the commodity pool corresponding to the purchasing enterprise is a subset of the comprehensive commodity pool.

S305, adding the second commodity into a commodity pool corresponding to the purchasing enterprise, wherein the second commodity belongs to the first commodity.

Illustratively, an adding control of a first commodity is displayed, a first operation of a user on the adding control of a second commodity is received, and the second commodity is added into a commodity pool corresponding to the purchasing enterprise according to the first operation.

S306, receiving the purchasing operation of the user on the commodities in the commodity pool corresponding to the purchasing enterprise.

The specific implementation and beneficial effects of this embodiment refer to the description of the second embodiment, which is not described herein again.

Fig. 11 is a flowchart of a product recommendation method for enterprise users according to a sixth embodiment of the present invention, where the method in this embodiment is used to construct a knowledge graph of an integrated product pool, and steps in this embodiment may be executed before step S101 in the fourth embodiment, as shown in fig. 11, the method in this embodiment includes the following steps:

s401, extracting the attributes and attribute values of the commodities in the comprehensive commodity pool.

All candidate attributes existing in each commodity are extracted from the original commodity data, and the candidate attribute with higher occurrence frequency is taken as the attribute of the commodity, for example, 8 candidate attributes are shared by the commodity 003, and only 5 of the candidate attributes have higher occurrence frequency, so that the 5 candidate attributes can be taken as the attributes of the commodity.

S402, correcting the attributes of the commodities.

The attribute of the goods is irregular, and different manufacturers or operators may have different names or different expressions for the same attribute, for example, when describing the size of the refrigerating chamber of the refrigerator, the function of manufacturer 1, namely refrigerating chamber (liter), is used for representing, and then the attribute can be corrected to be the volume of the refrigerating chamber.

And S403, clustering and fusing the corrected attributes of the commodities to obtain target attributes of the commodities.

The attribute specifications of the commodities of the respective manufacturers are different, some similar or similar attributes exist, and clustering means that similar attributes are classified into one category, for example, the attributes "EXTRA-volume", "EXTRA-capacity", "function-total volume (liter)" and "specification-volume" are similar attributes.

The fusion is to define attribute values of a specification, and to fuse similar attributes obtained by the above-described aggregation to attribute values of the new specification, for example, to define an attribute "volume" of one specification, and to fuse the attributes "EXTRA-volume", "EXTRA-capacity", "function-total volume (liter)" and "specification-volume" to an attribute "volume".

S404, determining the form of the attribute value of the target attribute.

The form of the attribute value of the target attribute is determined, that is, the attribute value of the attribute is normalized, and for the same attribute, description may be performed through attribute values of a plurality of different forms. For example, there may be multiple attribute values for an attribute volume, e.g., liters (L), milliliters (ml), gallons, cubic meters, etc., then the attribute values for the attribute "volume" may be normalized to a uniform attribute value of "number + liters".

S405, constructing a knowledge graph according to the target attributes and attribute values of the commodities.

After the target attributes and the attribute values after the normalization processing of the commodities are determined, the attributes and the attribute values of the commodities are stored according to a preset format, and the knowledge graph is obtained.

In this embodiment, the extracted attributes and attribute values of the commodities are normalized to obtain uniform attributes and attribute values, which facilitates subsequent commodity matching accuracy.

Optionally, in the fourth to sixth embodiments, the knowledge graph further includes historical transaction total amount of the commodity and a repurchase rate of the commodity. The historical transaction total amount and the commodity repurchase rate refer to the description of the above embodiment, and are not described again here. Correspondingly, when commodity recommendation is carried out, the commodity recommendation can be carried out according to the historical transaction total amount and the commodity repurchase rate.

Illustratively, the attribute information of the purchased commodity is matched with the attribute information of each commodity in the knowledge graph to obtain a first candidate commodity, a second candidate commodity is obtained according to the historical transaction total amount of each commodity and the repurchase rate of the commodity in the knowledge graph, and a recommended commodity is obtained according to the first candidate commodity and the second candidate commodity. The recommended commodity may be obtained by taking a union of the first candidate commodity and the second candidate commodity.

Or matching the attribute information of the purchased commodity with the attribute information of each commodity in the knowledge graph to obtain a first candidate commodity, then selecting a fourth candidate commodity meeting the conditions from the first candidate commodity according to the historical total volume of the first candidate commodity and the rate of repurchase of the commodity, wherein the fourth candidate commodity is a subset of the first candidate commodity, for example, selecting the commodity with larger historical total volume of transaction and larger rate of repurchase from the first candidate commodity as the fourth candidate commodity, and using the fourth candidate commodity as the recommended commodity of the user.

Optionally, the method may further include recommending the commodity for the enterprise by combining characteristics of the commodity purchased by the enterprise, exemplarily, performing characteristic extraction on the historical orders of the commodities in the comprehensive commodity pool by using a machine learning method to obtain characteristics of the commodity purchased by each enterprise, where the characteristics of the commodity purchased by the enterprise include at least one of the following data: category of the purchased goods, price of the purchased goods, combinability of the purchased goods, and repurchase rate of the purchased goods.

Correspondingly, when commodity recommendation is carried out, the attribute information of the purchased commodities is matched with the attribute information of each commodity in the knowledge graph to obtain first candidate commodities, then third candidate commodities are matched from the comprehensive commodity pool according to the characteristics of the purchased commodities of the purchasing enterprise, and recommended commodities are obtained according to the first candidate commodities and the third candidate commodities. Or after the first candidate commodities are obtained according to the attributes of the purchased commodities, fifth candidate commodities can be determined from the first candidate commodities according to the characteristics of the purchased commodities of the purchasing enterprise, the fifth candidate commodities are a subset of the first candidate commodities, and the fifth candidate commodities are taken as recommended commodities of the user.

Optionally, the characteristics of the purchased commodities of the purchasing enterprise, the attributes of the commodities, the historical transaction total amount of the commodities, and the rate of repurchase of the commodities may be combined together to recommend the commodities for the user, which is specifically described with reference to the third embodiment, and will not be described herein again.

Fig. 12 is a schematic structural diagram of a product recommendation device for an enterprise user according to a seventh embodiment of the present invention, and as shown in fig. 12, a device 100 according to this embodiment includes:

the receiving module 11 is configured to receive a purchasing requirement input by a user, where the purchasing requirement includes an industry to which a purchasing enterprise belongs and/or an application scenario of a purchasing commodity;

the matching module 12 is configured to match the purchasing demand with a knowledge graph of the comprehensive commodity pool to obtain recommended commodities, where the knowledge graph includes industries and/or application scenarios of the commodities in the comprehensive commodity pool;

and the display module 13 is used for displaying the recommended commodities to the user.

Optionally, the method further includes:

a determining module 14, configured to determine, from the recommended commodities, a first commodity that does not belong to a commodity pool corresponding to the purchasing enterprise, where the commodity pool corresponding to the purchasing enterprise is a subset of the comprehensive commodity pool;

the pool adding module 15 is configured to add a second commodity to a commodity pool corresponding to the purchasing enterprise, where the second commodity belongs to the first commodity;

and the purchasing module 16 is configured to receive a purchasing operation of the user on the commodities in the commodity pool corresponding to the purchasing enterprise.

Optionally, the pool adding module 15 is specifically configured to:

displaying an add control for the first item;

Optionally, the system further comprises a knowledge graph constructing module, configured to:

Optionally, the system further includes an update module, configured to:

Optionally, the matching module 12 is specifically configured to: and matching to obtain recommended commodities according to the characteristics of the purchased commodities of the purchasing enterprise, the purchasing demand and the knowledge graph of the comprehensive commodity pool.

Optionally, the matching module 12 is specifically configured to:

The apparatus of this embodiment may be configured to perform the method according to any one of the first to third embodiments, and the specific implementation manner and the technical effect are similar, which are not described herein again.

Fig. 13 is a schematic structural diagram of a product recommendation device for enterprise users according to an eighth embodiment of the present invention, and as shown in fig. 13, a device 200 according to this embodiment includes:

the attribute extraction module 21 is configured to extract attribute information of a purchased commodity from a keyword of the purchased commodity when no commodity is searched in a commodity pool corresponding to a purchasing enterprise according to the keyword of the purchased commodity input by a user or the number of searched commodities is less than a preset value;

the matching module 22 is configured to match the attribute information of the purchased commodity with a knowledge graph of the comprehensive commodity pool to obtain a recommended commodity, where the knowledge graph includes the attribute information of each commodity in the comprehensive commodity pool, and the commodity pool corresponding to the purchasing enterprise is a subset of the comprehensive commodity pool;

and the display module 23 is configured to display the recommended product to the user.

Optionally, the method further includes:

a determining module 24, configured to determine, from the recommended commodities, a first commodity that does not belong to a commodity pool corresponding to the purchasing enterprise;

a pool adding module 25, configured to add a second commodity to a commodity pool corresponding to the purchasing enterprise, where the second commodity belongs to the first commodity;

and the purchasing module 26 is used for receiving the purchasing operation of the user on the commodities in the commodity pool corresponding to the purchasing enterprise.

Optionally, the pool adding module 25 is specifically configured to:

displaying an add control for the first item;

correcting the attributes of the commodities;

determining a form of an attribute value of the target attribute;

Optionally, the matching module 22 is specifically configured to:

Optionally, further comprises a feature module for

correspondingly, the matching module 22 is specifically configured to:

The apparatus of this embodiment may be configured to perform the method according to any one of the fourth to sixth embodiments, and specific implementation and technical effects are similar and will not be described herein again.

Fig. 14 is a schematic structural diagram of an electronic device according to a ninth embodiment of the present application, and as shown in fig. 14, the electronic device 300 includes: the processor 31, the memory 32, and the transceiver 23, where the memory 32 is configured to store instructions, the transceiver 33 is configured to communicate with other devices, and the processor 31 is configured to execute the instructions stored in the memory, so that the electronic device 300 executes the method steps according to any one of the first to sixth embodiments, and specific implementation and technical effects are similar, and are not described herein again.

An embodiment tenth of the present application provides a computer-readable storage medium, where computer-executable instructions are stored in the computer-readable storage medium, and when the computer-executable instructions are executed by a processor, the computer-executable instructions are used to implement the method steps according to any one of the first to sixth embodiments, and specific implementation manners and technical effects are similar, and are not described herein again.

An eleventh embodiment of the present application provides a computer program product, including a computer program, where the computer program, when executed by a processor, implements the method steps described in any one of the first to sixth embodiments, and specific implementation and technical effects are similar, and are not described herein again.

Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

It will be understood that the present disclosure is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A commodity recommendation method for enterprise users is characterized by comprising the following steps:

and displaying the recommended commodity to a user.

2. The method of claim 1, further comprising:

3. The method of claim 2, wherein said adding a second item to a pool of items corresponding to said purchasing enterprise comprises:

displaying an add control for the first item;

4. The method according to any one of claims 1-3, further comprising:

5. The method of claim 4, further comprising:

6. The method of claim 5, wherein said matching recommended goods based on said procurement requirements and a knowledge-graph of a comprehensive pool of goods comprises:

7. The method of claim 6, wherein the knowledge-graph further comprises historical trades and historical repurchase rates of the goods.

8. The method of claim 7, wherein matching the recommended goods according to the characteristics of the purchased goods of the purchasing enterprise, the purchasing requirement and the knowledge graph of the comprehensive goods pool comprises:

9. A commodity recommendation method for enterprise users is characterized by comprising the following steps:

and displaying the recommended commodity to a user.

10. The method of claim 9, further comprising:

11. The method of claim 10, wherein said adding a second item to a pool of items corresponding to said purchasing enterprise comprises:

displaying an add control for the first item;

12. The method according to any one of claims 9-11, further comprising:

correcting the attributes of the commodities;

determining a form of an attribute value of the target attribute;

13. The method according to any one of claims 9 to 11, wherein the knowledge-graph further comprises historical trade total and repurchase rate of the commodity.

14. The method of claim 13, wherein the matching to obtain the recommended goods according to the attribute information of the purchased goods and the knowledge graph of the comprehensive goods pool comprises:

15. The method according to any one of claims 9-11, further comprising:

16. An enterprise user's merchandise recommendation device, comprising:

17. An enterprise user's merchandise recommendation device, comprising:

18. An electronic device, comprising: at least one processor and memory;

the memory stores computer-executable instructions;

the at least one processor executing the computer-executable instructions stored by the memory causes the at least one processor to perform the method of any of claims 1-15.

19. A computer-readable storage medium having computer-executable instructions stored thereon, which when executed by a processor, are configured to implement the method of any one of claims 1 to 15.

20. A computer program product comprising a computer program, characterized in that the computer program, when executed by a processor, implements the method of any one of claims 1 to 15.