CN109087177B - Method, device and computer-readable storage medium for recommending commodities to target users - Google Patents
Method, device and computer-readable storage medium for recommending commodities to target users Download PDFInfo
- Publication number
- CN109087177B CN109087177B CN201810952009.0A CN201810952009A CN109087177B CN 109087177 B CN109087177 B CN 109087177B CN 201810952009 A CN201810952009 A CN 201810952009A CN 109087177 B CN109087177 B CN 109087177B
- Authority
- CN
- China
- Prior art keywords
- user
- target user
- attribute
- commodity
- commodities
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Recommending goods or services
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0623—Electronic shopping [e-shopping] by investigating goods or services
- G06Q30/0625—Electronic shopping [e-shopping] by investigating goods or services by formulating product or service queries, e.g. using keywords or predefined options
- G06Q30/0627—Electronic shopping [e-shopping] by investigating goods or services by formulating product or service queries, e.g. using keywords or predefined options by specifying product or service characteristics, e.g. product dimensions
Landscapes
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
本公开提供了一种向目标用户推荐商品的方法、装置及计算机可读存储介质，涉及大数据技术领域。其中向目标用户推荐商品的方法包括：获取各个用户对目标用户已购商品在各个商品属性上的评分；利用评分计算各个用户中的其他用户与目标用户的相似度；利用与目标用户相似度最高的其他用户对目标用户未购商品的综合评分，预测目标用户对目标用户未购商品的综合评分；向目标用户推荐综合评分高于第一阈值的目标用户未购商品。本公开能够更加准确的推荐用户喜欢且未购买过的商品。
The present disclosure provides a method, a device and a computer-readable storage medium for recommending commodities to a target user, and relates to the technical field of big data. The method for recommending products to the target user includes: obtaining the scores of each user on the attributes of the products purchased by the target user; using the scores to calculate the similarity between other users in each user and the target user; using the score with the highest similarity with the target user According to the comprehensive score of other users on the target user's unpurchased products, predict the target user's comprehensive score on the target user's unpurchased products; recommend the target user's unpurchased products with a comprehensive score higher than the first threshold to the target user. The present disclosure can more accurately recommend products that the user likes but has not purchased.
Description
Technical Field
The present disclosure relates to the field of big data technologies, and in particular, to a method and an apparatus for recommending a commodity to a target user, and a computer-readable storage medium.
Background
The purchase flow of the user purchasing the commodity on the e-commerce platform is roughly as follows: and the user selects commodities independently, adds the commodities into the shopping cart, screens and determines final commodities independently, and pays. The existing e-commerce platform generally recommends commodities for users, and the recommendation process is as follows: and recommending commodities by the platform, adding the commodities into a shopping cart, automatically screening and determining final commodities, and paying.
When the user independently selects commodities, the commodity categories are too many, so that the time and labor are wasted when the user independently selects commodities, and the shopping experience is influenced. When a large number of interesting goods are selected, further selection is required before payment, and when a large number of goods which do not accord with the preference of the user are purchased, a large amount of return goods are generated.
Disclosure of Invention
One technical problem solved by the present disclosure is how to more accurately recommend goods that a user likes and has not purchased.
According to an aspect of the embodiments of the present disclosure, there is provided a method for recommending a commodity to a target user, including: the method comprises the steps of obtaining scores of various users on various commodity attributes of commodities purchased by target users; calculating the similarity between other users in each user and the target user by using the scores; predicting the comprehensive score of the target user on the commodities which are not purchased by the target user by using the comprehensive score of other users with the highest similarity to the target user on the commodities which are not purchased by the target user; and recommending the target user unpurchased commodities with the comprehensive score higher than the first threshold value to the target user.
In some embodiments, calculating the similarity of the other users of the respective users to the target user using the scores includes: generating a scoring vector of each user on a single commodity attribute by utilizing the scoring of each user on the single commodity attribute of the commodity purchased by the target user; calculating the similarity of other users and a target user on a single commodity attribute by utilizing the scoring vectors of the users on the single commodity attribute; and calculating the similarity between other users and the target user in each user by utilizing the similarity between other users and the target user in each commodity attribute.
In some embodiments, the similarity between other users and the target user on the single commodity attribute is calculated by the following method:
wherein x represents a scoring vector of a target user on a single commodity attribute, y represents a scoring vector of some other user in each user on the single commodity attribute, sim (x, y) represents the similarity of some other user and the target user on the single commodity attribute, and d (x, y) represents the Euclidean distance between the vectors x and y.
In some embodiments, the respective merchandise attributes include at least one of: brand, category, price, weight, color, size, material, origin, packaging, logistics.
In some embodiments, recommending to the target user that the target user has not purchased goods with a composite score above a first threshold comprises: selecting a user attribute with the user attribute weight higher than a second threshold value from all user attributes of the target user; and selecting commodities associated with the user attribute weight higher than the second threshold value from the target user unpurchased commodities with the comprehensive score higher than the first threshold value, and recommending the commodities to the target user.
In some embodiments, the method further comprises: generating each user attribute of the target user by using the registration information and the ordering information of the target user, wherein the user attribute comprises at least one of the following: gender, age, birthday, height, weight, nationality, hobbies, place of residence, time to place an order, current weather, current temperature.
In some embodiments, the method further comprises: initializing the user attribute weight of each user attribute to a default value; if the target user purchases a commodity associated with a certain user attribute, the user attribute weight of the user attribute is increased; and if the target user purchases a commodity which is not related to a certain user attribute, reducing the user attribute weight of the user attribute.
In some embodiments, recommending to the target user that the target user has not purchased goods with a composite score above a first threshold comprises: and selecting the target user unpurchased commodities with the highest comprehensive score in each commodity category from the target user unpurchased commodities with the comprehensive score higher than the first threshold value, and recommending the target user unpurchased commodities to the target user.
In some embodiments, the method further comprises: after the target user purchases the commodity, initializing the favorite weight of the target user to the purchased commodity; if the target user continues to purchase the purchased commodity, the preference weight of the target user on the purchased commodity is improved; if the target user does not purchase the purchased commodity any more, the favorite weight of the target user on the purchased commodity is reduced; recommending the commodities purchased by the target user with the preference weight higher than the third threshold value to the target user.
According to another aspect of the embodiments of the present disclosure, there is provided an apparatus for recommending a commodity to a target user, including: the score acquisition module is configured to acquire scores of commodities purchased by the target user on various commodity attributes by various users; the similarity calculation module is configured to calculate the similarity between other users in the users and the target user by using the scores; the score prediction module is configured to predict the comprehensive score of the target user on the commodities which are not purchased by the target user by utilizing the comprehensive score of other users with the highest similarity to the target user on the commodities which are not purchased by the target user; and the commodity recommending module is configured to recommend the target user unpurchased commodities with the comprehensive score higher than a first threshold value to the target user.
In some embodiments, the similarity calculation module is configured to: generating a scoring vector of each user on a single commodity attribute by utilizing the scoring of each user on the single commodity attribute of the commodity purchased by the target user; calculating the similarity of other users and a target user on a single commodity attribute by utilizing the scoring vectors of the users on the single commodity attribute; and calculating the similarity between other users and the target user in each user by utilizing the similarity between other users and the target user in each commodity attribute.
In some embodiments, the similarity calculation module is configured to: the similarity of other users and the target user on the attribute of the single commodity is calculated by adopting the following method:
wherein x represents a scoring vector of a target user on a single commodity attribute, y represents a scoring vector of some other user in each user on the single commodity attribute, sim (x, y) represents the similarity of some other user and the target user on the single commodity attribute, and d (x, y) represents the Euclidean distance between the vectors x and y.
In some embodiments, the respective merchandise attributes include at least one of: brand, category, price, weight, color, size, material, origin, packaging, logistics.
In some embodiments, the item recommendation module is configured to: selecting a user attribute with the user attribute weight higher than a second threshold value from all user attributes of the target user; and selecting commodities associated with the user attribute weight higher than the second threshold value from the target user unpurchased commodities with the comprehensive score higher than the first threshold value, and recommending the commodities to the target user.
In some embodiments, the apparatus further comprises a user attribute generation module configured to: generating each user attribute of the target user by using the registration information and the ordering information of the target user, wherein the user attribute comprises at least one of the following: gender, age, birthday, height, weight, nationality, hobbies, place of residence, time to place an order, current weather, current temperature.
In some embodiments, the apparatus further comprises a user attribute weight setting module configured to: initializing the user attribute weight of each user attribute to a default value; if the target user purchases a commodity associated with a certain user attribute, the user attribute weight of the user attribute is increased; and if the target user purchases a commodity which is not related to a certain user attribute, reducing the user attribute weight of the user attribute.
In some embodiments, the item recommendation module is configured to: and selecting the target user unpurchased commodities with the highest comprehensive score in each commodity category from the target user unpurchased commodities with the comprehensive score higher than the first threshold value, and recommending the target user unpurchased commodities to the target user.
In some embodiments, the apparatus further comprises a purchased goods recommendation module configured to: after the target user purchases the commodity, initializing the favorite weight of the target user to the purchased commodity; if the target user continues to purchase the purchased commodity, the preference weight of the target user on the purchased commodity is improved; if the target user does not purchase the purchased commodity any more, the favorite weight of the target user on the purchased commodity is reduced; recommending the commodities purchased by the target user with the preference weight higher than the third threshold value to the target user.
According to another aspect of the embodiments of the present disclosure, there is provided an apparatus for recommending a product to a target user, including: a memory; and a processor coupled to the memory, the processor configured to perform the aforementioned method of recommending merchandise to a target user based on instructions stored in the memory.
According to still another aspect of the embodiments of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the instructions when executed by a processor implement the aforementioned method for recommending goods to a target user.
The method and the device can recommend the commodities which are liked by the user and not purchased by the user more accurately.
Other features of the present disclosure and advantages thereof will become apparent from the following detailed description of exemplary embodiments thereof, which proceeds with reference to the accompanying drawings.
Drawings
In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, it is obvious that the drawings in the following description are only some embodiments of the present disclosure, and for those skilled in the art, other drawings can be obtained according to the drawings without inventive exercise.
Fig. 1 illustrates a flow diagram of a method of recommending merchandise to a target user according to some embodiments of the present disclosure.
FIG. 2 is a flow chart illustrating a method for recommending merchandise to a target user according to further embodiments of the present disclosure.
FIG. 3 illustrates a flow diagram of a method of recommending merchandise to a target user in accordance with yet further embodiments of the present disclosure.
Fig. 4 is a schematic structural diagram of an apparatus for recommending a product to a target user according to some embodiments of the present disclosure.
Fig. 5 is a schematic structural diagram of an apparatus for recommending goods to a target user according to further embodiments of the present disclosure.
Detailed Description
The technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the drawings in the embodiments of the present disclosure, and it is obvious that the described embodiments are only a part of the embodiments of the present disclosure, and not all of the embodiments. The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the disclosure, its application, or uses. All other embodiments, which can be derived by a person skilled in the art from the embodiments disclosed herein without making any creative effort, shall fall within the protection scope of the present disclosure.
The method is a recommendation strategy for finding users with similar interests through the historical information of the users and recommending the commodities purchased by the users with higher interest similarity to the target user. Briefly, if A, B both users purchased three books, x, y, and z, and given a 5-star rating, then A and B are similar users. The book w viewed by a may also be recommended to user B. Therefore, the collaborative filtering algorithm is mainly divided into two steps, and a similar user set is searched; and finding the favorite and unpurchased commodities of the target user in the set for recommendation. If each user in the e-commerce platform has an overall score for the commodity, and for the single-dimensional recommendation strategy, the score of the commodity i by the user x is predicted according to the score of the similar user group of the target user x for the commodity i. Therefore, if a similar user group of x can be found more accurately, the more accurate the predicted score is, and the more accurate the goods recommended according to the score is.
Table 1 shows a scheme of scoring a commodity after a user purchases the commodity in the related art. As shown in table 1, assuming that three users x1, x2, x3 score four commodities i1, i2, i3, i4, users most similar to x3 are synergistically filtered according to the scores of the three users for the respective commodities, and the score of x3 for commodity i4 is predicted. It appears that x2, x3 are most similar because they score the merchandise consistently. Since user x2 scored commodity i4 by 4, prediction x3 also scored i4 by 4.
TABLE 1
However, the inventor studies to find that, although some goods can be recommended to the user to guide the user to purchase in the related art, the goods that the user likes and does not purchase cannot be accurately recommended. This causes that the user does not purchase the recommended goods after the platform recommends, or cancels the order or applies for a refund after purchasing the recommended goods, and the user experience is poor. Based on the above problems, the inventor proposes a method for recommending a commodity to a target user, which can recommend a commodity which is liked by the user and not purchased by the user more accurately.
Some embodiments of the method of the present disclosure for recommending merchandise to a target user are first described in conjunction with FIG. 1.
Fig. 1 illustrates a flow diagram of a method of recommending merchandise to a target user according to some embodiments of the present disclosure. As shown in fig. 1, the present embodiment includes steps S102 to S108.
In step S102, the scores of the respective users on the respective commodity attributes of the commodities purchased by the target user are acquired.
Wherein each commodity attribute may include: brand, category, price, weight, color, size, material, origin, packaging, logistics, and the like.
TABLE 2
As shown in table 2, three users x1, x2, x3 multi-dimensionally score four commodity attributes of four commodities i1, i2, i3, i 4. Although the composite score of the user on the goods is the same as that in table 1 (the composite score is an average value of the multidimensional scores), table 2 provides multidimensional scores on a plurality of goods attributes, such as brand names, prices, colors and whether to package mails, and the multidimensional scores can more accurately reflect the preference and opinion of the user. As can be seen from table 2, although the composite scores of x2 and x3 for commodities i1, i2, and i3 are the same, the scores of x2 and x3 for commodities i1, i2, and i3 are completely different in each commodity attribute. x2 likes the brand and price of the commodities i1, i2, i3, but x3 is just dissatisfied with both aspects. A composite score based on a single dimension obviously ignores such implicit information, resulting in a computational bias in user similarity. In Table 2, user x1 is more similar to user x3 in that they have more similar scores for various merchandise attributes. The scores of the user on the multi-dimensional commodity attributes can more clearly indicate which aspects of the purchased commodities the user likes, so that the scores on the multi-dimensional commodity attributes can more accurately estimate the similarity of the user.
In step S104, the degree of similarity between the other users and the target user among the users is calculated by using the scores of the respective users on the attributes of the commodities purchased by the target user.
First, scores of the commodities purchased by the target user on a single commodity attribute can be generated by the users, and a score vector of each user on the single commodity attribute can be generated. For example, the score vector of user x1 on the brand is (3, 3, 4), the score vector of user x2 on the brand is (9, 8, 8), and the score vector of user x3 on the brand is (5, 2, 2).
Then, the similarity of other users and the target user on the single commodity attribute is calculated by utilizing the scoring vectors of the users on the single commodity attribute.
For example, the following method may be adopted to calculate the similarity of other users and the target user on the single commodity attribute:
wherein x represents a scoring vector of a target user on a single commodity attribute, y represents a scoring vector of some other user in each user on the single commodity attribute, sim (x, y) represents the similarity of some other user and the target user on the single commodity attribute, and d (x, y) represents the Euclidean distance between the vectors x and y.
Here, the similarity between the target user x3 and the other user x1 in the brand may be calculated by setting x to (5, 2, 2) and y to (3, 3, 4), and the similarity between the target user x3 and the other user x2 in the brand may be calculated by setting x to (5, 2, 2) and y to (9, 8, 8).
And finally, calculating the similarity between other users and the target user in each user by utilizing the similarity between other users and the target user in each commodity attribute.
For example, the average of the similarity between the other user and the target user in each product attribute may be used as the similarity between the other user and the target user.
In step S106, the overall rating of the target user for the unpurchased goods of the target user is predicted by using the overall rating of the target user for the unpurchased goods of the other users with the highest similarity to the target user.
For example, if the user x1 is another user with the highest similarity to the user x4, the overall score of the target user for the unpurchased goods of the target user is predicted by using the overall score of 6 for the goods i4 of the user x 1. Wherein, the composite score of the user x1 on the commodity i4 may be an average of scores of the user x1 on the commodity i4 on each commodity attribute.
In step S108, the target user who has a composite score higher than the first threshold value and does not purchase the commodity is recommended to the target user.
For example, the target user may be recommended the unpurchased goods of the target user whose composite score is higher than 6.
The embodiment optimizes the algorithm of recommending the commodities by the platform, and introduces the multi-dimensional commodity attributes and the scores of the multi-dimensional commodity attributes. The grading of the user on the multi-dimensional commodity attribute can reflect the preference degree of the user to different aspects, so that more information between the user and the commodity is utilized, commodities which are liked by the user and not purchased by the user are recommended more intelligently and accurately, and the user experience is improved.
In some embodiments, in step S108, from the target user unpurchased commodities with the composite score higher than the first threshold, the target user unpurchased commodity with the highest composite score in each commodity category may be selected and recommended to the target user.
For example, the target user unpurchased goods with the composite score higher than 6 points include three pieces of clothes (the composite scores are respectively 9 points, 8 points and 8 points), two pieces of trousers (the composite scores are respectively 9 points, 8 points and 7 points), and two pieces of shoes (the composite scores are respectively 8 points and 7 points). Then, clothes with a composite score of 9, trousers with a composite score of 8, and shoes with a composite score of 8 may be recommended to the target user as a final combination.
In this way, if the user purchases a large number of commodities of the repeated categories according to the recommendation, the purchasing combination of only one commodity at most in each category is screened out for the user to select according to the user preference. Therefore, the embodiment further reduces the time for selecting the commodities for the user, and reduces the probability of goods return after the user purchases the unneeded commodities, thereby further providing intelligent shopping experience for the user.
Further embodiments of the method of recommending merchandise to a target user of the present disclosure are described below in conjunction with FIG. 2.
FIG. 2 is a flow chart illustrating a method for recommending merchandise to a target user according to further embodiments of the present disclosure. As shown in fig. 2, the present embodiment includes steps S202 to S220.
In step S202, the scores of the respective users on the respective product attributes of the products purchased by the target user are acquired. The specific implementation process of step S202 may refer to step S102.
In step S204, the scores of the respective users on the attributes of the commodities purchased by the target user are used to calculate the similarity between the target user and the other users in the respective users. The specific implementation process of step S204 may refer to step S104.
In step S206, the overall rating of the target user for the unpurchased goods of the target user is predicted by using the overall rating of the target user for the unpurchased goods of the other users with the highest similarity to the target user. The specific implementation process of step S206 may refer to step S106.
In step S208, user attributes of the target user are generated using the registration information and the order placing information of the target user.
Wherein the user attributes include: gender, age, birthday, height, weight, nationality, hobbies, place of residence, time of placing an order, current weather, current temperature, etc. The user attributes may be specifically classified into user individual attributes and user integrated attributes. The personal attributes of the user may include, for example, sex, age, birthday, height, weight, nationality, hobbies, etc., and the comprehensive attributes of the user may include, for example, time to place an order, current weather, current temperature, etc.
In step S210, the user attribute weight of each user attribute is initialized to a default value.
For example, the user attribute weight for each user attribute may be initialized to a default value of 5.
In step S212, it is determined whether the target user purchases a product associated with a certain user attribute. If yes, go to step S214; if not, go to step S216.
For example, the registration information of the target user includes a height of 185cm and a taste of spicy food. When the target user purchases the goods, all the clothing sizes are 185cm, but the purchased food never has a hot taste, the weight of the height attribute in the user attribute needs to be increased, and the weight of the taste attribute that likes to eat hot needs to be decreased.
In step S214, the user attribute weight of the user attribute is increased.
For example, each time a user purchases a good associated with the user attribute, the user attribute weight for the user attribute may be multiplied by 1.1.
In step S216, the user attribute weight of the user attribute is reduced.
For example, each time a user purchases a product that is not associated with the user attribute, the user attribute weight for the user attribute may be divided by 1.1.
In step S218, a user attribute whose user attribute weight is higher than the second threshold value is selected from the respective user attributes of the target user.
For example, a user attribute with a user attribute weight greater than 7 may be selected from the user attributes of the target user.
In step S220, a product associated with the user attribute having the user attribute weight higher than the second threshold value is selected from the target user unpurchased products having the composite score higher than the first threshold value, and recommended to the target user.
When the commodities are recommended, the commodities needing to be recommended can be further screened according to the user attributes. For example, the residence place where the user is located has high humidity and high temperature, and summer-heat-avoiding and dehumidifying commodities (such as sun cream, cold drink and coix seed powder) can be further selected from the commodities not purchased by the target user with the comprehensive score higher than the first threshold value and recommended to the user.
In the above embodiment, when the data of the purchased commodities is less, the user's preference and shopping habit may not be accurately known, and at this time, the commodities recommended for the user can be further intelligently screened according to the user attributes to make intelligent associated recommendations, so that the problem of cold start caused by data coefficients in the conventional recommendation scheme is solved. As the number of commodities purchased by the user is increased, the purchasing habits and the preferences of the learning user are gradually improved, the user attributes and the commodity attributes can be gradually combined, and the unpurchased commodities which are liked by the user can be more accurately recommended to new users or old users. On the other hand, the embodiment can attract users, increase the user quantity, enhance the trust and the good sensitivity of the users to the E-commerce platform, further reduce the time for the users to select commodities, reduce the goods return rate of the users and further improve the user experience.
Still other embodiments of the method of recommending merchandise to a target user of the present disclosure are described below in conjunction with FIG. 3.
FIG. 3 illustrates a flow diagram of a method of recommending merchandise to a target user in accordance with yet further embodiments of the present disclosure. As shown in fig. 3, in addition to the embodiment shown in fig. 2, the present embodiment further includes steps S322 to S330.
In step S322, after the target user purchases the product, the favorite weight of the target user for the purchased product is initialized.
For example, the favorite weight of the target user for the purchased product v may be initialized to 5.
In step S324, it is determined whether the target user continues to purchase the purchased commodity v.
If the target user continues to purchase the purchased commodity, go to step S326; if the target user no longer purchases the purchased product, step S328 is executed.
In step S326, the preference weight of the target user for the purchased product is increased.
For example, each time the target user buys the purchased product v again, the target user may multiply by 1.2 based on the preference weight of the purchased product v.
In step S328, the preference weight of the target user for the purchased goods is reduced.
For example, the target user may divide the value by 1.2 based on the preference weight of the purchased product v, where the purchased product v is not included in every five purchased products.
In step S330, the target user is recommended the purchased goods whose preference weight is higher than the third threshold.
For example, the target user may be recommended the purchased goods with the preference weight higher than 8.
The embodiment can effectively avoid the situation that the user purchases the commodity and continuously recommends the purchased commodity. Therefore, the embodiment enables the e-commerce platform to understand the purchasing intention of the user more according to the multidimensional data, and intelligent purchasing guidance is achieved.
An apparatus for recommending merchandise to a target user according to some embodiments of the present disclosure is described below with reference to fig. 4.
Fig. 4 is a schematic structural diagram of an apparatus for recommending a product to a target user according to some embodiments of the present disclosure. As shown in fig. 4, the apparatus 40 for recommending a product to a target user in this embodiment includes a score obtaining module 402, a similarity calculating module 404, a score predicting module 406, and a product recommending module 412.
The score obtaining module 402 is configured to obtain scores of the commodities purchased by the target user on various commodity attributes by the respective users; the similarity calculation module 404 is configured to calculate similarities of other users among the users with the target user using the scores; the score prediction module 406 is configured to predict the comprehensive score of the target user on the unpurchased goods of the target user by using the comprehensive score of the other users with the highest similarity to the target user on the unpurchased goods of the target user; the goods recommendation module 408 is configured to recommend the target user unpurchased goods with the composite score higher than the first threshold value to the target user.
The embodiment optimizes the algorithm of recommending the commodities by the platform, and introduces the multi-dimensional commodity attributes and the scores of the multi-dimensional commodity attributes. The grading of the user on the multi-dimensional commodity attribute can reflect the preference degree of the user to different aspects, so that more information between the user and the commodity is utilized, commodities which are liked by the user and not purchased by the user are recommended more intelligently and accurately, and the user experience is improved.
In some embodiments, the similarity calculation module 404 is configured to: generating a scoring vector of each user on a single commodity attribute by utilizing the scoring of each user on the single commodity attribute of the commodity purchased by the target user; calculating the similarity of other users and a target user on a single commodity attribute by utilizing the scoring vectors of the users on the single commodity attribute; and calculating the similarity between other users and the target user in each user by utilizing the similarity between other users and the target user in each commodity attribute.
In some embodiments, the similarity calculation module 404 is configured to: the similarity of other users and the target user on the attribute of the single commodity is calculated by adopting the following method:
wherein x represents a scoring vector of a target user on a single commodity attribute, y represents a scoring vector of some other user in each user on the single commodity attribute, sim (x, y) represents the similarity of some other user and the target user on the single commodity attribute, and d (x, y) represents the Euclidean distance between the vectors x and y.
In some embodiments, the respective merchandise attributes include at least one of: brand, category, price, weight, color, size, material, origin, packaging, logistics.
In some embodiments, the item recommendation module 412 is configured to: selecting a user attribute with the user attribute weight higher than a second threshold value from all user attributes of the target user; and selecting commodities associated with the user attribute weight higher than the second threshold value from the target user unpurchased commodities with the comprehensive score higher than the first threshold value, and recommending the commodities to the target user.
In some embodiments, the apparatus 40 further comprises a user attribute generation module 408 configured to: generating each user attribute of the target user by using the registration information and the ordering information of the target user, wherein the user attribute comprises at least one of the following: gender, age, birthday, height, weight, nationality, hobbies, place of residence, time to place an order, current weather, current temperature.
In some embodiments, the apparatus 40 further comprises a user attribute weight setting module 410 configured to: initializing the user attribute weight of each user attribute to a default value; if the target user purchases a commodity associated with a certain user attribute, the user attribute weight of the user attribute is increased; and if the target user purchases a commodity which is not related to a certain user attribute, reducing the user attribute weight of the user attribute.
In some embodiments, the item recommendation module 412 is configured to: and selecting the target user unpurchased commodities with the highest comprehensive score in each commodity category from the target user unpurchased commodities with the comprehensive score higher than the first threshold value, and recommending the target user unpurchased commodities to the target user.
Therefore, if the user purchases a large number of commodities of repeated categories according to the recommendation, a purchase combination of at most one commodity of each category is screened out for the user to select according to the preference of the user, so that the time for the user to select the commodities is further shortened, the probability of goods return after the user purchases the commodities which are not needed is reduced, and the intelligent shopping experience is further provided for the user.
In the above embodiment, when the data of the purchased commodities is less, the user's preference and shopping habit may not be accurately known, and at this time, the commodities recommended for the user can be further intelligently screened according to the user attributes to make intelligent associated recommendations, so that the problem of cold start caused by data coefficients in the conventional recommendation scheme is solved. As the number of commodities purchased by the user is increased, the purchasing habits and the preferences of the learning user are gradually improved, the user attributes and the commodity attributes can be gradually combined, and the unpurchased commodities which are liked by the user can be more accurately recommended to new users or old users. On the other hand, the embodiment can attract users, increase the user quantity, enhance the trust and the good sensitivity of the users to the E-commerce platform, further reduce the time for the users to select commodities, reduce the goods return rate of the users and further improve the user experience.
In some embodiments, the apparatus 40 further includes a purchased goods recommendation module 414 configured to: after the target user purchases the commodity, initializing the favorite weight of the target user to the purchased commodity; if the target user continues to purchase the purchased commodity, the preference weight of the target user on the purchased commodity is improved; if the target user does not purchase the purchased commodity any more, the favorite weight of the target user on the purchased commodity is reduced; recommending the commodities purchased by the target user with the preference weight higher than the third threshold value to the target user.
The embodiment can effectively avoid the situation that the user purchases the commodity and continuously recommends the purchased commodity. Therefore, the embodiment enables the e-commerce platform to understand the purchasing intention of the user more according to the multidimensional data, and intelligent purchasing guidance is achieved.
Fig. 5 is a schematic structural diagram of an apparatus for recommending goods to a target user according to further embodiments of the present disclosure. As shown in fig. 5, the apparatus 50 for recommending a product to a target user according to this embodiment includes: a memory 510 and a processor 520 coupled to the memory 510, the processor 520 configured to perform a method of recommending merchandise to a target user in any of the foregoing embodiments based on instructions stored in the memory 510.
The apparatus 40 for recommending goods to a target user may further include an input-output interface 530, a network interface 4540, a storage interface 550, and the like. These interfaces 530, 540, 550 and the connections between the memory 510 and the processor 520 may be, for example, via a bus 560. The input/output interface 530 provides a connection interface for input/output devices such as a display, a mouse, a keyboard, and a touch screen. The network interface 540 provides a connection interface for various networking devices. The storage interface 550 provides a connection interface for external storage devices such as an SD card and a usb disk.
The present disclosure also includes a computer-readable storage medium having stored thereon computer instructions that, when executed by a processor, implement a method of recommending items to a target user in any of the foregoing embodiments.
As will be appreciated by one skilled in the art, embodiments of the present disclosure may be provided as a method, system, or computer program product. Accordingly, the present disclosure may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable non-transitory storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present disclosure is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the disclosure. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only exemplary of the present disclosure and is not intended to limit the present disclosure, so that any modification, equivalent replacement, or improvement made within the spirit and principle of the present disclosure should be included in the scope of the present disclosure.
Claims (12)
1. A method of recommending merchandise to a target user, comprising:
the method comprises the steps of obtaining scores of various users on various commodity attributes of commodities purchased by target users;
calculating the similarity between other users in each user and the target user by using the scores, wherein the similarity comprises the following steps: generating a scoring vector of each user on a single commodity attribute by utilizing the scoring of each user on the single commodity attribute of the commodity purchased by the target user; calculating the similarity of other users and target users on the single commodity attribute by utilizing the scoring vectors of the users on the single commodity attribute
Wherein x represents a scoring vector of a target user on a single commodity attribute, y represents a scoring vector of some other user in each user on the single commodity attribute, sim (x, y) represents the similarity of some other user and the target user on the single commodity attribute, and d (x, y) represents the Euclidean distance between the vectors x and y; calculating the similarity between other users and the target user in each user by utilizing the similarity between other users and the target user in each commodity attribute;
predicting the comprehensive score of the target user on the commodities which are not purchased by the target user by using the comprehensive score of other users with the highest similarity to the target user on the commodities which are not purchased by the target user;
initializing the user attribute weight of each user attribute to a default value; if the target user purchases a commodity associated with a certain user attribute, the user attribute weight of the user attribute is increased; if the target user purchases a commodity which is not associated with a certain user attribute, reducing the user attribute weight of the user attribute;
recommending the target user unpurchased commodities with the composite score higher than a first threshold value to the target user, wherein the recommending comprises the following steps: selecting a user attribute with the user attribute weight higher than a second threshold value from all user attributes of the target user; and selecting commodities associated with the user attribute weight higher than the second threshold value from the target user unpurchased commodities with the comprehensive score higher than the first threshold value, and recommending the commodities to the target user.
2. The method of claim 1, wherein the respective merchandise attributes comprise at least one of: brand, category, price, weight, color, size, material, origin, packaging, logistics.
3. The method of claim 1, further comprising:
generating each user attribute of the target user by using the registration information and the ordering information of the target user, wherein the user attribute comprises at least one of the following: gender, age, birthday, height, weight, nationality, hobbies, place of residence, time to place an order, current weather, current temperature.
4. The method of claim 1, wherein the recommending the target user of the unpopulated goods with the composite score higher than the first threshold value to the target user comprises:
and selecting the target user unpurchased commodities with the highest comprehensive score in each commodity category from the target user unpurchased commodities with the comprehensive score higher than the first threshold value, and recommending the target user unpurchased commodities to the target user.
5. The method of claim 1, further comprising:
after the target user purchases the commodity, initializing the favorite weight of the target user to the purchased commodity;
if the target user continues to purchase the purchased commodity, the preference weight of the target user on the purchased commodity is improved;
if the target user does not purchase the purchased commodity any more, reducing the favorite weight of the target user on the purchased commodity;
recommending the commodities purchased by the target user with the preference weight higher than the third threshold value to the target user.
6. An apparatus for recommending items to a target user, comprising:
the score acquisition module is configured to acquire scores of commodities purchased by the target user on various commodity attributes by various users;
a similarity calculation module configured to calculate similarities of other users of the respective users with a target user using the scores, including: generating a scoring vector of each user on a single commodity attribute by utilizing the scoring of each user on the single commodity attribute of the commodity purchased by the target user; calculating the similarity of other users and target users on the single commodity attribute by utilizing the scoring vectors of the users on the single commodity attribute
Wherein x represents a scoring vector of a target user on a single commodity attribute, y represents a scoring vector of some other user in each user on the single commodity attribute, sim (x, y) represents the similarity of some other user and the target user on the single commodity attribute, and d (x, y) represents the Euclidean distance between the vectors x and y; calculating the similarity between other users and the target user in each user by utilizing the similarity between other users and the target user in each commodity attribute;
the score prediction module is configured to predict the comprehensive score of the target user on the commodities which are not purchased by the target user by utilizing the comprehensive score of other users with the highest similarity to the target user on the commodities which are not purchased by the target user;
a user attribute weight setting module configured to: initializing the user attribute weight of each user attribute to a default value; if the target user purchases a commodity associated with a certain user attribute, the user attribute weight of the user attribute is increased; if the target user purchases a commodity which is not associated with a certain user attribute, the user attribute weight of the user attribute is reduced
The commodity recommending module is configured to recommend the target user unpurchased commodities with the comprehensive score higher than a first threshold value to the target user, and comprises the following steps: selecting a user attribute with the user attribute weight higher than a second threshold value from all user attributes of the target user; and selecting commodities associated with the user attribute weight higher than the second threshold value from the target user unpurchased commodities with the comprehensive score higher than the first threshold value, and recommending the commodities to the target user.
7. The apparatus of claim 6, wherein the respective merchandise attributes comprise at least one of: brand, category, price, weight, color, size, material, origin, packaging, logistics.
8. The apparatus of claim 6, further comprising a user attribute generation module configured to:
generating each user attribute of the target user by using the registration information and the ordering information of the target user, wherein the user attribute comprises at least one of the following: gender, age, birthday, height, weight, nationality, hobbies, place of residence, time to place an order, current weather, current temperature.
9. The apparatus of claim 6, wherein the item recommendation module is configured to:
and selecting the target user unpurchased commodities with the highest comprehensive score in each commodity category from the target user unpurchased commodities with the comprehensive score higher than the first threshold value, and recommending the target user unpurchased commodities to the target user.
10. The apparatus of claim 6, further comprising a purchased goods recommendation module configured to:
after the target user purchases the commodity, initializing the favorite weight of the target user to the purchased commodity;
if the target user continues to purchase the purchased commodity, the preference weight of the target user on the purchased commodity is improved;
if the target user does not purchase the purchased commodity any more, reducing the favorite weight of the target user on the purchased commodity;
recommending the commodities purchased by the target user with the preference weight higher than the third threshold value to the target user.
11. An apparatus for recommending items to a target user, comprising:
a memory; and
a processor coupled to the memory, the processor configured to perform the method of recommending merchandise to a target user of any of claims 1-5 based on instructions stored in the memory.
12. A computer readable storage medium, wherein the computer readable storage medium stores computer instructions which, when executed by a processor, implement a method of recommending items to a target user as recited in any of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810952009.0A CN109087177B (en) | 2018-08-21 | 2018-08-21 | Method, device and computer-readable storage medium for recommending commodities to target users |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810952009.0A CN109087177B (en) | 2018-08-21 | 2018-08-21 | Method, device and computer-readable storage medium for recommending commodities to target users |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109087177A CN109087177A (en) | 2018-12-25 |
| CN109087177B true CN109087177B (en) | 2021-05-25 |
Family
ID=64793931
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810952009.0A Active CN109087177B (en) | 2018-08-21 | 2018-08-21 | Method, device and computer-readable storage medium for recommending commodities to target users |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109087177B (en) |
Families Citing this family (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111695023A (en) * | 2019-03-11 | 2020-09-22 | 北京京东尚科信息技术有限公司 | Information recommendation method and device, storage medium and equipment |
| CN110033351B (en) * | 2019-04-15 | 2021-09-10 | 秒针信息技术有限公司 | Method and device for determining similarity |
| CN110008410A (en) * | 2019-04-16 | 2019-07-12 | 上饶市中科院云计算中心大数据研究院 | A kind of personalization of product recommended method |
| CN110807691B (en) * | 2019-10-31 | 2022-03-04 | 深圳市云积分科技有限公司 | Cross-commodity-class commodity recommendation method and device |
| CN112950304B (en) * | 2019-12-11 | 2024-06-18 | 北京沃东天骏信息技术有限公司 | Information push method, device, equipment and storage medium |
| CN111292170A (en) * | 2020-02-18 | 2020-06-16 | 重庆锐云科技有限公司 | Method, device and storage medium for recommending intention customers for appointed building |
| CN111681086A (en) * | 2020-06-16 | 2020-09-18 | 上海风秩科技有限公司 | Commodity recommendation method and device, computer equipment and readable storage medium |
| CN113763065B (en) * | 2020-06-17 | 2024-07-19 | 北京沃东天骏信息技术有限公司 | Method, device, apparatus and computer-readable medium for recommending products |
| CN113763067B (en) * | 2020-06-19 | 2024-08-20 | 北京沃东天骏信息技术有限公司 | Method, device, equipment and storage medium for pushing item information |
| CN113763070B (en) * | 2020-06-29 | 2025-02-21 | 北京沃东天骏信息技术有限公司 | Information recommendation method and device |
| CN111985994A (en) * | 2020-08-06 | 2020-11-24 | 上海博泰悦臻电子设备制造有限公司 | Commodity recommendation method and related equipment |
| CN113034224A (en) * | 2021-03-16 | 2021-06-25 | 重庆锐云科技有限公司 | Similarity-based house source recommendation method, system, equipment and storage medium |
| CN113222687A (en) * | 2021-04-22 | 2021-08-06 | 杭州腾纵科技有限公司 | Deep learning-based recommendation method and device |
| US12033180B2 (en) * | 2021-06-25 | 2024-07-09 | Z2 Cool Comics Llc | Semi-autonomous advertising systems and methods |
| CN113592588B (en) * | 2021-07-25 | 2023-10-03 | 深圳市瀚力科技有限公司 | E-commerce platform product recommendation system and method based on big data collaborative filtering technology |
| CN116258541A (en) * | 2021-12-09 | 2023-06-13 | 沈阳精一智驾科技有限公司 | A push method, computer-readable storage medium and electronic device |
| CN115048585B (en) * | 2022-06-28 | 2025-09-05 | 中国工商银行股份有限公司 | Product recommendation method, product recommendation model training method, device and equipment |
| CN117333203B (en) * | 2023-12-01 | 2024-04-16 | 广东付惠吧数据服务有限公司 | Member marketing platform combined with business marketing solution |
| CN120278797A (en) * | 2025-06-10 | 2025-07-08 | 南京亿猫信息技术有限公司 | Commodity recommendation method and device for intelligent shopping cart and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104899246A (en) * | 2015-04-12 | 2015-09-09 | 西安电子科技大学 | Collaborative filtering recommendation method of user rating neighborhood information based on fuzzy mechanism |
| CN106776479A (en) * | 2016-12-16 | 2017-05-31 | 北京理工大学 | A kind of score in predicting method towards many attribute ratings systems |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103345517B (en) * | 2013-07-10 | 2019-03-26 | 北京邮电大学 | Simulate the Collaborative Filtering Recommendation Algorithm of TF-IDF Similarity measures |
| CN103617540B (en) * | 2013-10-17 | 2017-01-11 | 浙江大学 | E-commerce recommendation method of tracking user interest changes |
| US9936031B2 (en) * | 2015-03-31 | 2018-04-03 | International Business Machines Corporation | Generation of content recommendations |
| WO2016191959A1 (en) * | 2015-05-29 | 2016-12-08 | 深圳市汇游智慧旅游网络有限公司 | Time-varying collaborative filtering recommendation method |
| CN105069072B (en) * | 2015-07-30 | 2018-08-21 | 天津大学 | Hybrid subscriber score information based on sentiment analysis recommends method and its recommendation apparatus |
-
2018
- 2018-08-21 CN CN201810952009.0A patent/CN109087177B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104899246A (en) * | 2015-04-12 | 2015-09-09 | 西安电子科技大学 | Collaborative filtering recommendation method of user rating neighborhood information based on fuzzy mechanism |
| CN106776479A (en) * | 2016-12-16 | 2017-05-31 | 北京理工大学 | A kind of score in predicting method towards many attribute ratings systems |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109087177A (en) | 2018-12-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109087177B (en) | Method, device and computer-readable storage medium for recommending commodities to target users | |
| US11816721B2 (en) | Recommendation apparatus and method | |
| US20220138831A1 (en) | Method of Providing Fashion Item Recommendation Service Using User's Body Type and Purchase History | |
| US8165924B2 (en) | Virtual shopping center | |
| CN108648059B (en) | Order recommendation method and device | |
| US9460464B2 (en) | Systems and methods for displaying items | |
| US9361640B1 (en) | Method and system for efficient order placement | |
| US20150379609A1 (en) | Generating recommendations for unfamiliar users by utilizing social side information | |
| US20120284145A1 (en) | Item discovery tools and methods for shopping in an electronic commerce environment | |
| US9183510B1 (en) | Method and system for personalized recommendation of lifestyle items | |
| KR20130047808A (en) | System for ubiquitous smart shopping | |
| US20190108571A1 (en) | System and method for efficiently determining relevant products for display in an online marketplace | |
| US11393013B2 (en) | Method, non-transitory computer-readable device, and system for intelligent listing creation | |
| KR20130047799A (en) | System for ubiquitous smart shopping | |
| CN107679898A (en) | A kind of Method of Commodity Recommendation and device | |
| US20230245174A1 (en) | Systems and methods for determining user personas for a user | |
| CN108153792A (en) | A kind of data processing method and relevant apparatus | |
| KR101567684B1 (en) | Method for selecting recommendation type in product recommeding system | |
| JP7140588B2 (en) | Decision device, decision method and decision program | |
| KR102285942B1 (en) | A method of providing a fashion item recommendation service to a user | |
| JP6945518B2 (en) | Information processing equipment, information processing methods and information processing programs | |
| KR102378072B1 (en) | Method, apparatus and computer program for style recommendation | |
| KR20160059583A (en) | Apparatuts, method, and compuer program managing deal information | |
| KR20160020654A (en) | Goods Recommending System, Method and Readable Recoding Medium Using Purchasing Information | |
| CN112711706A (en) | Information interaction method and device, readable storage medium and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |