CN109087177A - To the method, apparatus and computer readable storage medium of target user's Recommendations - Google Patents
To the method, apparatus and computer readable storage medium of target user's Recommendations Download PDFInfo
- Publication number
- CN109087177A CN109087177A CN201810952009.0A CN201810952009A CN109087177A CN 109087177 A CN109087177 A CN 109087177A CN 201810952009 A CN201810952009 A CN 201810952009A CN 109087177 A CN109087177 A CN 109087177A
- Authority
- CN
- China
- Prior art keywords
- user
- target user
- commodity
- property
- target
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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—Item recommendations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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—Item investigation
- G06Q30/0625—Directed, with specific intent or strategy
- G06Q30/0627—Directed, with specific intent or strategy using item specifications
Abstract
Present disclose provides a kind of method, apparatus and computer readable storage medium to target user's Recommendations, are related to big data technical field.It include: wherein to obtain each user to have purchased scoring of the commodity in each item property to target user to the method for target user's Recommendations;The other users in each user and the similarity of target user are calculated using scoring;Using the comprehensive score for not purchasing commodity to target user with the highest other users of target user's similarity, predict that target user does not purchase the comprehensive score of commodity to target user;The target user for recommending comprehensive score to be higher than first threshold to target user does not purchase commodity.The disclosure being capable of the more accurate recommended user commodity liking and do not bought.
Description
Technical field
This disclosure relates to big data technical field, in particular to a kind of method, apparatus to target user's Recommendations and
Computer readable storage medium.
Background technique
User buys the process of purchase of commodity substantially in electric business platform are as follows: and user independently selects commodity, and shopping cart is added, from
Main screening determines final goods, payment.Existing electric business platform generally can be user's Recommendations, recommended flowsheet are as follows: platform pushes away
Commodity are recommended, shopping cart is added, autonomous screen determines final goods, payment.
When user independently selects commodity, since commodity category is excessive, user independently select it is time-consuming and laborious, influence do shopping body
It tests.After having selected a large amount of interested commodity, need further to be selected again before payment, and work as and have purchased largely not
After the commodity for meeting oneself hobby, a large amount of return of goods can be generated.
Summary of the invention
The disclosure solve a technical problem be, how the quotient that more accurate recommended user likes and do not bought
Product.
According to the one aspect of the embodiment of the present disclosure, a kind of method to target user's Recommendations is provided, comprising: obtain
Each user is taken to purchase scoring of the commodity in each item property to target user;Its in each user is calculated using scoring
The similarity of his user and target user;Using with target user's similarity highest other users commodity are not purchased to target user
Comprehensive score, prediction target user do not purchase the comprehensive scores of commodity to target user;Recommend comprehensive score high to target user
Commodity are not purchased in the target user of first threshold.
In some embodiments, the other users in each user and the similarity packet of target user are calculated using scoring
It includes: scoring of the commodity on particular commodity attribute having been purchased to target user using each user, has generated each user in single quotient
Scoring vector on product attribute;Using scoring vector of each user on particular commodity attribute, other users and target are calculated
Similarity of the user on particular commodity attribute;Using in each user other users and target user in each item property
On similarity, calculate the other users in each user and the similarity of target user.
In some embodiments, phase of the other users with target user on particular commodity attribute is calculated with the following method
Like degree:
Wherein, x indicates scoring vector of the target user on particular commodity attribute, y indicate in each user some its
His scoring vector of the user on particular commodity attribute, sim (x, y) indicate some other users and target user in particular commodity
Similarity on attribute, d (x, y) indicate the Euclidean distance between vector x, y.
In some embodiments, each item property comprises at least one of the following: brand, classification, price, weight, color,
Size, material, the place of production, packaging, logistics.
In some embodiments, commodity bundle is not purchased to the target user that target user recommends comprehensive score to be higher than first threshold
It includes: from each user property of target user, user property weight being selected to be higher than the user property of second threshold;It is commented from synthesis
Divide the target user higher than first threshold not purchase in commodity, selects the user property phase for being higher than second threshold with user property weight
Associated commodity, recommend target user.
In some embodiments, this method further include: using the registration information and order placement information of target user, generate target
Each user property of user, user property comprise at least one of the following: gender, age, birthday, height, weight, nationality, love
Good, residence, lower single time, current weather, Current Temperatures.
In some embodiments, this method further include: by the user property weights initialisation of each user property be default
Value;If target user has purchased commodity associated with certain user property, the user property weight of the user property is improved;If
Target user has purchased the commodity unrelated to certain user property, then reduces the user property weight of the user property.
In some embodiments, commodity bundle is not purchased to the target user that target user recommends comprehensive score to be higher than first threshold
Include: the target user for being higher than first threshold from comprehensive score does not purchase in commodity, selects comprehensive score highest in each merchandise classification
Target user do not purchase commodity, recommend target user.
In some embodiments, this method further include: after target user buys commodity, initialized target user is to having purchased quotient
The hobby weight of product;If target user continues to have purchased commodity, target user is improved to the hobby weight for having purchased commodity;If
Target user, which no longer buys, has purchased commodity, then reduces target user to the hobby weight for having purchased commodity;Hobby weight is higher than the
The target user of three threshold values has purchased commercial product recommending to target user.
According to the other side of the embodiment of the present disclosure, a kind of device to target user's Recommendations is provided, comprising:
Scoring obtains module, is configured as obtaining each user and has purchased scoring of the commodity in each item property to target user;Phase
Like degree computing module, it is configured as calculating the other users in each user and the similarity of target user using scoring;Scoring
Prediction module is configured as commenting using the synthesis for not purchasing target user commodity with the highest other users of target user's similarity
Point, prediction target user does not purchase the comprehensive score of commodity to target user;Commercial product recommending module is configured as pushing away to target user
The target user that comprehensive score is recommended higher than first threshold does not purchase commodity.
In some embodiments, similarity calculation module is configured as: having purchased commodity to target user using each user
Scoring on particular commodity attribute generates scoring vector of each user on particular commodity attribute;Existed using each user
Scoring vector on particular commodity attribute calculates the similarity of other users and target user on particular commodity attribute;It utilizes
The similarity of other users and target user in each item property in each user calculates other use in each user
The similarity at family and target user.
In some embodiments, similarity calculation module is configured as: calculating other users and target with the following method
Similarity of the user on particular commodity attribute:
Wherein, x indicates scoring vector of the target user on particular commodity attribute, y indicate in each user some its
His scoring vector of the user on particular commodity attribute, sim (x, y) indicate some other users and target user in particular commodity
Similarity on attribute, d (x, y) indicate the Euclidean distance between vector x, y.
In some embodiments, each item property comprises at least one of the following: brand, classification, price, weight, color,
Size, material, the place of production, packaging, logistics.
In some embodiments, commercial product recommending module is configured as: from each user property of target user, selection is used
Family attribute weight is higher than the user property of second threshold;The target user for being higher than first threshold from comprehensive score does not purchase in commodity,
Selection, higher than the associated commodity of the user property of second threshold, recommends target user with user property weight.
In some embodiments, which further includes user property generation module, is configured as: utilizing the note of target user
Volume information and order placement information, generate each user property of target user, user property comprises at least one of the following: gender, year
Age, birthday, height, weight, nationality, hobby, residence, lower single time, current weather, Current Temperatures.
In some embodiments, which further includes user property weight setting module, is configured as: each user is belonged to
Property user property weights initialisation be default value;If target user has purchased commodity associated with certain user property, mention
The user property weight of the high user property;If target user has purchased the commodity unrelated to certain user property, reduce
The user property weight of the user property.
In some embodiments, commercial product recommending module is configured as: being higher than the target user of first threshold from comprehensive score
It does not purchase in commodity, selects the highest target user of comprehensive score in each merchandise classification not purchase commodity, recommend target user.
In some embodiments, which further includes having purchased commercial product recommending module, is configured as: target user buys commodity
Afterwards, initialized target user is to the hobby weight for having purchased commodity;If target user continues to have purchased commodity, target use is improved
Family is to the hobby weight for having purchased commodity;If target user no longer buys and purchased commodity, target user is reduced to having purchased commodity
Like weight;Target user of the weight higher than third threshold value will be liked and purchased commercial product recommending to target user.
According to the another aspect of the embodiment of the present disclosure, another device to target user's Recommendations is provided, is wrapped
It includes: memory;And it is coupled to the processor of memory, processor is configured as holding based on instruction stored in memory
The row method above-mentioned to target user's Recommendations.
According to another aspect of the embodiment of the present disclosure, a kind of computer readable storage medium is provided, wherein computer
Readable storage medium storing program for executing is stored with computer instruction, and instruction is realized above-mentioned to target user's Recommendations when being executed by processor
Method.
The disclosure being capable of the more accurate recommended user commodity liking and do not bought.
By the detailed description referring to the drawings to the exemplary embodiment of the disclosure, the other feature of the disclosure and its
Advantage will become apparent.
Detailed description of the invention
In order to illustrate more clearly of the embodiment of the present disclosure or technical solution in the prior art, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Disclosed some embodiments without any creative labor, may be used also for those of ordinary skill in the art
To obtain other drawings based on these drawings.
Fig. 1 shows the flow diagram of the method to target user's Recommendations of some embodiments of the disclosure.
Fig. 2 shows the flow diagrams of the method to target user's Recommendations of the disclosure other embodiments.
Fig. 3 shows the flow diagram of the method to target user's Recommendations of the other embodiment of the disclosure.
Fig. 4 shows the structural schematic diagram of the device to target user's Recommendations of some embodiments of the disclosure.
Fig. 5 shows the structural schematic diagram of the device to target user's Recommendations of the disclosure other embodiments.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present disclosure, the technical solution in the embodiment of the present disclosure is carried out clear, complete
Site preparation description, it is clear that described embodiment is only disclosure a part of the embodiment, instead of all the embodiments.Below
Description only actually at least one exemplary embodiment be it is illustrative, never as to the disclosure and its application or making
Any restrictions.Based on the embodiment in the disclosure, those of ordinary skill in the art are not making creative work premise
Under all other embodiment obtained, belong to the disclosure protection range.
The similar user of interest, then the quotient that the higher user of Interest Similarity is bought are found by the historical information of user
It is a kind of Generalization bounds that product, which recommend target user,.In simple terms, exactly if two users of A, B have purchased x, y, z three
Books, and the favorable comment of 5 stars is given, then A and B are with regard to similar users.The books w that A has been seen can also be recommended into user B.
So collaborative filtering is broadly divided into two steps, similar user's set is found;Find what target user in set liked
And non-purchased goods are recommended.If each user has an entirety scoring to commodity in electric business platform, for single dimension
Generalization bounds then need the scoring according to the similar users group of target user x to commodity i, to predict that user x comments commodity i
Point.Therefore, if the similar users group of x can more accurately be found, the scoring of prediction is more accurate, the commodity recommended according to scoring
Also more accurate.
Table 1 shows user in the related technology and buys the scheme to score after commodity commodity.As shown in table 1, it is assumed that
Three users x1, x2, x3 to four commodity i1, i2, i3, i4 scorings, according to three users to the scoring situations of each commodity come
Collaborative filtering goes out the user most like with x3, and predicts scoring of the x3 to commodity i4.Seem that x2, x3 are most like, because they
It is consistent to the scoring of commodity.Because user x2 makes 4 scores to commodity i4, prediction x3 is also 4 points to the scoring of i4.
Table 1
However, although inventor is the study found that have been able to recommend some commodity guiding users for user in the related technology
Purchase, but still can not accurately recommended user's commodity for liking and not buying.Which results in users after platform recommendation not
It cancels an order or applies for a refund after having purchase Recommendations, or purchase Recommendations, user experience is poor.Based on asking above
Topic, inventors herein propose a kind of method to target user's Recommendations, can more accurate recommended user like and do not purchase
The commodity bought.
First combine Fig. 1 describe method from the disclosure to target user's Recommendations some embodiments.
Fig. 1 shows the flow diagram of the method to target user's Recommendations of some embodiments of the disclosure.Such as Fig. 1
Shown, the present embodiment includes step S102~step S108.
In step s 102, it obtains each user and scoring of the commodity in each item property has been purchased to target user.
Wherein, each item property may include: brand, classification, price, weight, color, size, material, the place of production, packet
Dress, logistics etc..
Table 2
As shown in table 2, three users x1, x2, x3 carry out multidimensional to four item properties of four commodity i1, i2, i3, i4
Degree scoring.Although user (average value that comprehensive score is various dimensions scoring), table as table 1 to the comprehensive score of commodity
2 provide in multiple item properties various dimensions scoring, as brand, price, color, whether packet postal, various dimensions scoring can be more
Add the preference and opinion for accurately embodying user.Although from table 2 it can be seen that comprehensive score of the x2 and x3 to commodity i1, i2, i3
Equally, but x2 and x3 is entirely different to the scoring of commodity i1, i2, i3 in each item property.X2 like commodity i1, i2,
The brand and price of i3, but x3 is just unsatisfied with the two aspects.Comprehensive score based on single dimension obviously has ignored in this way
Implicit information, so as to cause the calculating deviation of user's similarity.In table 2, user x1 is more similar to x3 because they
Scoring in each item property is increasingly similar.The explanation that the scoring in various dimensions item property of user can be more clear is used
Which aspect of commodity is liked having purchased in family, thus the scoring in various dimensions item property for user's similarity estimation more
Accurately.
In step S104, scoring of the commodity in each item property has been purchased to target user using each user, has been counted
Calculate the other users in each user and the similarity of target user.
It is possible, firstly, to purchase scoring of the commodity on particular commodity attribute to target user using each user, generate each
Scoring vector of a user on particular commodity attribute.For example, scoring vector of the user x1 in brand is (3,3,4), user
Scoring vector of the x2 in brand is (9,8,8), and scoring vector of the user x3 in brand is (5,2,2).
Then, the scoring vector using each user on particular commodity attribute, calculates other users and target user exists
Similarity on particular commodity attribute.
For example, the similarity of other users and target user on particular commodity attribute can be calculated with the following method:
Wherein, x indicates scoring vector of the target user on particular commodity attribute, y indicate in each user some its
His scoring vector of the user on particular commodity attribute, sim (x, y) indicate some other users and target user in particular commodity
Similarity on attribute, d (x, y) indicate the Euclidean distance between vector x, y.
Here x=(5,2,2), y=(3,3,4) can be enabled to calculate target user x3 and other users x1 in brand
Similarity enables x=(5,2,2), y=(9,8,8) calculate target user x3 and similarity of the other users x2 in brand.
Finally, being calculated using the similarity of other users and target user in each item property in each user
The similarity of other users and target user in each user.
For example, can be by the average value of the similarity of other users and target user in each item property, as it
The similarity of his user and target user.
In step s 106, the comprehensive of commodity is not purchased to target user using with the highest other users of target user's similarity
Scoring is closed, prediction target user does not purchase the comprehensive score of commodity to target user.
For example, user x1 be with the highest other users of user's x4 similarity, then using user x1 to the synthesis of commodity i4
6 points of scoring, prediction target user do not purchase the comprehensive score of commodity to target user.Wherein, user x1 comments the synthesis of commodity i4
Dividing can be average value of the user x1 to the scoring in each item property of commodity i4.
In step S108, the target user for recommending comprehensive score to be higher than first threshold to target user does not purchase commodity.
For example, target user of the comprehensive score higher than 6 points can be recommended not purchase commodity to target user.
The algorithm of platform Recommendations is optimized in above-described embodiment, introduces the item property of various dimensions and more
The scoring of dimension item property.Scoring of the user in various dimensions item property can embody user to the preference journey of different aspect
Degree, so that more information between user and commodity be utilized, more intelligently, accurate recommended user likes and do not bought
Commodity, to promote user experience.
In some embodiments, it in step S108, can not purchased from the target user that comprehensive score is higher than first threshold
In commodity, selects the highest target user of comprehensive score in each merchandise classification not purchase commodity, recommend target user.
For example, comprehensive score higher than 6 points of target user do not purchase commodity include three clothes (comprehensive score is respectively 9 points,
8 points, 8 points), two pieces trousers (comprehensive score is respectively 9 points, 8 points, 7 points), two pairs of shoes (comprehensive score is respectively 8 points, 7 points).That
, the shoes that trousers that clothes that comprehensive score 9 is divided, comprehensive score 8 are divided, comprehensive score 8 are divided can be pushed away as final combination
It recommends to target user.
So, if user is according to recommendation plus has purchased a large amount of repetition category commodity, further according to user preferences, sieve
Selecting each category at most only has a kind of purchase combination of commodity to select for user.Thus above-described embodiment further reduces use
Family selects the time of commodity, reduces user and buys the probability returned goods after this unwanted commodity, to further mention for user
For intelligentized shopping experience.
Below with reference to Fig. 2 describe method from the disclosure to target user's Recommendations other embodiments.
Fig. 2 shows the flow diagrams of the method to target user's Recommendations of the disclosure other embodiments.Such as
Shown in Fig. 2, the present embodiment includes step S202~step S220.
In step S202, obtains each user and scoring of the commodity in each item property has been purchased to target user.Step
The specific implementation process of rapid S202 is referred to step S102.
In step S204, scoring of the commodity in each item property has been purchased to target user using each user, has been counted
Calculate the other users in each user and the similarity of target user.The specific implementation process of step S204 is referred to step
S104。
In step S206, the comprehensive of commodity is not purchased to target user using with the highest other users of target user's similarity
Scoring is closed, prediction target user does not purchase the comprehensive score of commodity to target user.The specific implementation process of step S206 can join
According to step S106.
In step S208, using the registration information and order placement information of target user, each user of target user is generated
Attribute.
Wherein, user property include: gender, the age, the birthday, height, weight, nationality, hobby, residence, lower single time,
Current weather, Current Temperatures etc..User property can be specifically divided into individual subscriber attribute and the synthesized attribute of user.Wherein, it uses
Family personal attribute for example may include gender, age, birthday, height, weight, nationality, hobby etc., and the synthesized attribute of user is for example
It may include lower single time, current weather, Current Temperatures etc..
It is default value by the user property weights initialisation of each user property in step S210.
For example, can be default value 5 by the user property weights initialisation of each user property.
In step S212, judge whether target user has purchased commodity associated with certain user property.If so, holding
Row step S214;If it is not, thening follow the steps S216.
For example, in the registration information of target user comprising height 185cm, hobby eat it is peppery.When target user buys commodity,
All garment dimensions are all 185cm, but the taste that the food bought is never peppery, then the height attribute in user property
Weight needs to improve, and the weight for liking eating peppery hobby attribute weight needs to reduce.
In step S214, the user property weight of the user property is improved.
For example, user's one commodity associated with the user property of every purchase, the user of the user property can be belonged to
Property weight is multiplied by 1.1.
In step S216, the user property weight of the user property is reduced.
For example, user's one commodity unrelated to the user property of every purchase, it can be by the user of the user property
Attribute weight is divided by 1.1.
In step S218, from each user property of target user, user property weight is selected to be higher than second threshold
User property.
For example, user property weight can be selected to be greater than 7 user property from each user property of target user.
In step S220, the target user for being higher than first threshold from comprehensive score is not purchased in commodity, and selection belongs to user
Property weight be higher than second threshold the associated commodity of user property, recommend target user.
When Recommendations, the commodity for needing to recommend can be further screened according to user property.Such as the residence where user
Residence humidity is larger, temperature is higher, can not purchase in commodity from the target user that comprehensive score is higher than first threshold, further select
Select avoid heat, the commodity (such as suncream, cold drink, lotus-seed-heart powder) that dehumidify recommend user.
In above-described embodiment, when the data of user's purchased item are less, the happiness of user possibly can not be accurately known
Good and purchasing habits can intelligent screening be further at this time commodity that user recommends by user property, make intelligence
Correlation recommendation, to solve the problems, such as traditional suggested design because being cold-started caused by data coefficient.Increase as user buys commodity
More, buying habit and the hobby for learning user are also gradually perfect, then can gradually combine user property, item property, no
Pipe to new user or old user, can more accurately recommended user like do not purchase commodity.On the other hand, above-described embodiment can
To attract user, increase user volume, enhancing user to the degree of belief and likability of electric business platform, it is further reduced user and selects quotient
The time of product, the return of goods rate for reducing user, further raising user experience.
Below with reference to Fig. 3 describe method from the disclosure to target user's Recommendations other embodiment.
Fig. 3 shows the flow diagram of the method to target user's Recommendations of the other embodiment of the disclosure.Such as
Shown in Fig. 3, on the basis of embodiment shown in Fig. 2, the present embodiment further includes step S322~step S330.
In step S322, after target user buys commodity, initialized target user is to the hobby weight for having purchased commodity.
For example, can with initialized target user to purchased commodity v hobby weight be 5.
In step S324, judge whether target user continues to have purchased commodity v.
If target user continues to have purchased commodity, S326 is thened follow the steps;If target user no longer buys and has purchased commodity,
Then follow the steps S328.
In step S326, target user is improved to the hobby weight for having purchased commodity.
For example, target user has once purchased commodity v per purchase again, it can be on the hobby weighted basis for having purchased commodity v
Multiplied by 1.2.
In step S328, target user is reduced to the hobby weight for having purchased commodity.
Commodity v has been purchased for example, not including in target user's five commodity of every purchase, it can be in the hobby power for having purchased commodity v
Divided by 1.2 on the basis of weight.
In step S330, target user of the weight higher than third threshold value will be liked and purchased commercial product recommending to target user.
For example, can will like target user of the weight higher than 8 has purchased commercial product recommending to target user.
Above-described embodiment can effectively avoid user from having purchased commodity, still constantly recommend the feelings of purchased item
Condition.Therefore, above-described embodiment can make electric business platform according to multidimensional data, more solve the buying intention of user, realize intelligence purchase
Buy guidance.
The device to target user's Recommendations of some embodiments of the disclosure is described below with reference to Fig. 4.
Fig. 4 shows the structural schematic diagram of the device to target user's Recommendations of some embodiments of the disclosure.Such as Fig. 4
Shown, in the present embodiment includes that scoring obtains module 402, similarity calculation module to the device 40 of target user's Recommendations
404, score in predicting module 406 and commercial product recommending module 412.
Scoring, which obtains module 402 and is configured as obtaining each user, has purchased commodity in each item property to target user
Scoring;Similarity calculation module 404 is configured as calculating the other users and target user in each user using scoring
Similarity;Score in predicting module 406 be configured as using with the highest other users of target user's similarity to target user not
The comprehensive score of commodity is purchased, prediction target user does not purchase the comprehensive score of commodity to target user;Commercial product recommending module 408 is matched
It is set to the target user that target user recommends comprehensive score to be higher than first threshold and does not purchase commodity.
The algorithm of platform Recommendations is optimized in above-described embodiment, introduces the item property of various dimensions and more
The scoring of dimension item property.Scoring of the user in various dimensions item property can embody user to the preference journey of different aspect
Degree, so that more information between user and commodity be utilized, more intelligently, accurate recommended user likes and do not bought
Commodity, to promote user experience.
In some embodiments, similarity calculation module 404 is configured as: having purchased quotient to target user using each user
Scoring of the product on particular commodity attribute generates scoring vector of each user on particular commodity attribute;Utilize each user
Scoring vector on particular commodity attribute calculates the similarity of other users and target user on particular commodity attribute;Benefit
With the similarity of other users and target user in each item property in each user, other in each user are calculated
The similarity of user and target user.
In some embodiments, similarity calculation module 404 is configured as: calculating other users and mesh with the following method
Mark similarity of the user on particular commodity attribute:
Wherein, x indicates scoring vector of the target user on particular commodity attribute, y indicate in each user some its
His scoring vector of the user on particular commodity attribute, sim (x, y) indicate some other users and target user in particular commodity
Similarity on attribute, d (x, y) indicate the Euclidean distance between vector x, y.
In some embodiments, each item property comprises at least one of the following: brand, classification, price, weight, color,
Size, material, the place of production, packaging, logistics.
In some embodiments, commercial product recommending module 412 is configured as: from each user property of target user, choosing
Select the user property that user property weight is higher than second threshold;The target user for being higher than first threshold from comprehensive score does not purchase commodity
In, selection, higher than the associated commodity of the user property of second threshold, recommends target user with user property weight.
In some embodiments, which further includes user property generation module 408, is configured as: being used using target
The registration information and order placement information at family generate each user property of target user, and user property comprises at least one of the following: property
Not, age, birthday, height, weight, nationality, hobby, residence, lower single time, current weather, Current Temperatures.
In some embodiments, which further includes user property weight setting module 410, is configured as: will be each
The user property weights initialisation of user property is default value;If target user has purchased quotient associated with certain user property
Product then improve the user property weight of the user property;If target user has purchased the commodity unrelated to certain user property,
Then reduce the user property weight of the user property.
In some embodiments, commercial product recommending module 412 is configured as: the target for being higher than first threshold from comprehensive score is used
Family is not purchased in commodity, is selected the highest target user of comprehensive score in each merchandise classification not purchase commodity, is recommended target user.
So, if user is according to recommendation plus has purchased a large amount of repetition category commodity, further according to user preferences, sieve
Selecting each category at most only has a kind of purchase combination of commodity to select for user, so that further reducing user selects commodity
Time, reduce user and buy the probability returned goods after this unwanted commodity, to further be provided for user intelligentized
Shopping experience.
In above-described embodiment, when the data of user's purchased item are less, the happiness of user possibly can not be accurately known
Good and purchasing habits can intelligent screening be further at this time commodity that user recommends by user property, make intelligence
Correlation recommendation, to solve the problems, such as traditional suggested design because being cold-started caused by data coefficient.Increase as user buys commodity
More, buying habit and the hobby for learning user are also gradually perfect, then can gradually combine user property, item property, no
Pipe to new user or old user, can more accurately recommended user like do not purchase commodity.On the other hand, above-described embodiment can
To attract user, increase user volume, enhancing user to the degree of belief and likability of electric business platform, it is further reduced user and selects quotient
The time of product, the return of goods rate for reducing user, further raising user experience.
In some embodiments, which further includes having purchased commercial product recommending module 414, is configured as: target user's purchase
After buying commodity, initialized target user is to the hobby weight for having purchased commodity;If target user continues to have purchased commodity, improve
Target user is to the hobby weight for having purchased commodity;If target user no longer buys and purchased commodity, target user is reduced to having purchased
The hobby weight of commodity;Target user of the weight higher than third threshold value will be liked and purchased commercial product recommending to target user.
Above-described embodiment can effectively avoid user from having purchased commodity, still constantly recommend the feelings of purchased item
Condition.Therefore, above-described embodiment can make electric business platform according to multidimensional data, more solve the buying intention of user, realize intelligence purchase
Buy guidance.
Fig. 5 shows the structural schematic diagram of the device to target user's Recommendations of the disclosure other embodiments.Such as
Shown in Fig. 5, which includes: memory 510 to the device 50 of target user's Recommendations and is coupled to the memory
510 processor 520, processor 520 are configured as executing aforementioned any based on the instruction being stored in memory 510
The method to target user's Recommendations in embodiment.
Wherein, memory 510 is such as may include system storage, fixed non-volatile memory medium.System storage
Device is for example stored with operating system, application program, Boot loader (Boot Loader) and other programs etc..
It can also include input/output interface 530, network interface 4540, storage to the device 40 of target user's Recommendations
Interface 550 etc..Bus 560 can for example be passed through between these interfaces 530,540,550 and memory 510 and processor 520
Connection.Wherein, input/output interface 530 is display, the input-output equipment such as mouse, keyboard, touch screen provide connecting interface.
Network interface 540 provides connecting interface for various networked devices.The external storages such as memory interface 550 is SD card, USB flash disk provide
Connecting interface.
The disclosure further includes a kind of computer readable storage medium, is stored thereon with computer instruction, and the instruction is processed
Device realizes the method to target user's Recommendations in aforementioned any some embodiments when executing.
It should be understood by those skilled in the art that, embodiment of the disclosure can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the disclosure
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the disclosure, which can be used in one or more,
The calculating implemented in non-transient storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) can be used
The form of machine program product.
The disclosure is referring to method, the process of equipment (system) and computer program product according to the embodiment of the present disclosure
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
The foregoing is merely the preferred embodiments of the disclosure, not to limit the disclosure, all spirit in the disclosure and
Within principle, any modification, equivalent replacement, improvement and so on be should be included within the protection scope of the disclosure.
Claims (20)
1. a kind of method to target user's Recommendations, comprising:
It obtains each user and scoring of the commodity in each item property has been purchased to target user;
The other users in each user and the similarity of target user are calculated using the scoring;
Using the comprehensive score for not purchasing commodity to target user with the highest other users of target user's similarity, predict that target is used
The comprehensive score of commodity is not purchased to target user in family;
The target user for recommending comprehensive score to be higher than first threshold to target user does not purchase commodity.
2. the method for claim 1, wherein other users and mesh calculated using the scoring in each user
Mark user similarity include:
Scoring of the commodity on particular commodity attribute has been purchased to target user using each user, has generated each user in single quotient
Scoring vector on product attribute;
Using scoring vector of each user on particular commodity attribute, other users and target user are calculated in particular commodity category
Similarity in property;
Using the similarity of other users and target user in each item property in each user, calculate in each user
Other users and target user similarity.
3. method according to claim 2, wherein calculate other users and target user with the following method in particular commodity
Similarity on attribute:
Wherein, x indicates that scoring vector of the target user on particular commodity attribute, y indicate some other use in each user
Scoring vector of the family on particular commodity attribute, sim (x, y) indicate some other users and target user in particular commodity attribute
On similarity, d (x, y) indicates vector x, the Euclidean distance between y.
4. the method for claim 1, wherein each item property comprises at least one of the following: brand, classification,
Price, weight, color, size, material, the place of production, packaging, logistics.
5. the method for claim 1, wherein target for recommending comprehensive score to be higher than first threshold to target user
User does not purchase commodity
From each user property of target user, user property weight is selected to be higher than the user property of second threshold;
The target user for being higher than first threshold from comprehensive score does not purchase in commodity, and selection is higher than second threshold with user property weight
The associated commodity of user property, recommend target user.
6. method as claimed in claim 5, further includes:
Using the registration information and order placement information of target user, each user property of target user, the user property are generated
Comprise at least one of the following: gender, the age, the birthday, height, weight, nationality, hobby, residence, lower single time, current weather,
Current Temperatures.
7. method as claimed in claim 5, further includes:
It is default value by the user property weights initialisation of each user property;
If target user has purchased commodity associated with certain user property, the user property weight of the user property is improved;
If target user has purchased the commodity unrelated to certain user property, the user property power of the user property is reduced
Weight.
8. the method for claim 1, wherein target for recommending comprehensive score to be higher than first threshold to target user
User does not purchase commodity
The target user for being higher than first threshold from comprehensive score does not purchase in commodity, selects comprehensive score highest in each merchandise classification
Target user do not purchase commodity, recommend target user.
9. method as described in claim 1, further includes:
After target user buys commodity, initialized target user is to the hobby weight for having purchased commodity;
If target user continues to have purchased commodity, target user is improved to the hobby weight for having purchased commodity;
If target user no longer buys and purchased commodity, target user is reduced to the hobby weight for having purchased commodity;
Target user of the weight higher than third threshold value will be liked and purchased commercial product recommending to target user.
10. a kind of device to target user's Recommendations, comprising:
Scoring obtains module, is configured as obtaining each user and has purchased commodity commenting in each item property to target user
Point;
Similarity calculation module is configured as calculating the other users and the phase of target user in each user using the scoring
Like degree;
Score in predicting module is configured as not purchasing commodity to target user using with the highest other users of target user's similarity
Comprehensive score, prediction target user do not purchase the comprehensive scores of commodity to target user;
Commercial product recommending module, the target user for being configured as that comprehensive score is recommended to be higher than first threshold to target user do not purchase quotient
Product.
11. device as claimed in claim 10, wherein the similarity calculation module is configured as:
Scoring of the commodity on particular commodity attribute has been purchased to target user using each user, has generated each user in single quotient
Scoring vector on product attribute;
Using scoring vector of each user on particular commodity attribute, other users and target user are calculated in particular commodity category
Similarity in property;
Using the similarity of other users and target user in each item property in each user, calculate in each user
Other users and target user similarity.
12. device as claimed in claim 11, wherein the similarity calculation module is configured as:
The similarity of other users and target user on particular commodity attribute is calculated with the following method:
Wherein, x indicates that scoring vector of the target user on particular commodity attribute, y indicate some other use in each user
Scoring vector of the family on particular commodity attribute, sim (x, y) indicate some other users and target user in particular commodity attribute
On similarity, d (x, y) indicates vector x, the Euclidean distance between y.
13. device as claimed in claim 10, wherein each item property comprises at least one of the following: brand, class
Not, price, weight, color, size, material, the place of production, packaging, logistics.
14. device as claimed in claim 10, wherein the commercial product recommending module is configured as:
From each user property of target user, user property weight is selected to be higher than the user property of second threshold;
The target user for being higher than first threshold from comprehensive score does not purchase in commodity, and selection is higher than second threshold with user property weight
The associated commodity of user property, recommend target user.
15. device as claimed in claim 14 further includes user property generation module, is configured as:
Using the registration information and order placement information of target user, each user property of target user, the user property are generated
Comprise at least one of the following: gender, the age, the birthday, height, weight, nationality, hobby, residence, lower single time, current weather,
Current Temperatures.
16. device as claimed in claim 14 further includes user property weight setting module, is configured as:
It is default value by the user property weights initialisation of each user property;
If target user has purchased commodity associated with certain user property, the user property weight of the user property is improved;
If target user has purchased the commodity unrelated to certain user property, the user property power of the user property is reduced
Weight.
17. device as claimed in claim 10, wherein the commercial product recommending module is configured as:
The target user for being higher than first threshold from comprehensive score does not purchase in commodity, selects comprehensive score highest in each merchandise classification
Target user do not purchase commodity, recommend target user.
18. device as claimed in claim 10 further includes having purchased commercial product recommending module, is configured as:
After target user buys commodity, initialized target user is to the hobby weight for having purchased commodity;
If target user continues to have purchased commodity, target user is improved to the hobby weight for having purchased commodity;
If target user no longer buys and purchased commodity, target user is reduced to the hobby weight for having purchased commodity;
Target user of the weight higher than third threshold value will be liked and purchased commercial product recommending to target user.
19. a kind of device to target user's Recommendations, comprising:
Memory;And
It is coupled to the processor of the memory, the processor is configured to the instruction based on storage in the memory,
It executes as claimed in any one of claims 1-9 wherein to the method for target user's Recommendations.
20. a kind of computer readable storage medium, wherein the computer-readable recording medium storage has computer instruction, institute
It states when instruction is executed by processor and realizes as claimed in any one of claims 1-9 wherein to the method for target user's Recommendations.
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 user |
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 user |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109087177A true CN109087177A (en) | 2018-12-25 |
CN109087177B 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 user |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109087177B (en) |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
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 |
CN110033351A (en) * | 2019-04-15 | 2019-07-19 | 秒针信息技术有限公司 | A kind of determination method and device of similarity |
CN110807691A (en) * | 2019-10-31 | 2020-02-18 | 深圳市云积分科技有限公司 | Cross-commodity-class commodity recommendation method and device |
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 |
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 |
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 |
CN112950304A (en) * | 2019-12-11 | 2021-06-11 | 北京沃东天骏信息技术有限公司 | Information pushing method, device, equipment and storage medium |
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 |
CN113592588A (en) * | 2021-07-25 | 2021-11-02 | 北京慧橙信息科技有限公司 | E-commerce platform commodity recommendation system and method based on big data collaborative filtering technology |
CN113763067A (en) * | 2020-06-19 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Article information pushing method, device, equipment and storage medium |
CN113763065A (en) * | 2020-06-17 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Method, device, equipment and computer readable medium for recommending commodities |
CN117333203A (en) * | 2023-12-01 | 2024-01-02 | 广东付惠吧数据服务有限公司 | Member marketing platform combined with business marketing solution |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103345517A (en) * | 2013-07-10 | 2013-10-09 | 北京邮电大学 | Collaborative filtering recommendation algorithm simulating TF-IDF similarity calculation |
CN103617540A (en) * | 2013-10-17 | 2014-03-05 | 浙江大学 | E-commerce recommendation method of tracking user interest changes |
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 |
CN105069072A (en) * | 2015-07-30 | 2015-11-18 | 天津大学 | Emotional analysis based mixed user scoring information recommendation method and apparatus |
US20160294961A1 (en) * | 2015-03-31 | 2016-10-06 | 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 |
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 |
-
2018
- 2018-08-21 CN CN201810952009.0A patent/CN109087177B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103345517A (en) * | 2013-07-10 | 2013-10-09 | 北京邮电大学 | Collaborative filtering recommendation algorithm simulating TF-IDF similarity calculation |
CN103617540A (en) * | 2013-10-17 | 2014-03-05 | 浙江大学 | E-commerce recommendation method of tracking user interest changes |
US20160294961A1 (en) * | 2015-03-31 | 2016-10-06 | International Business Machines Corporation | Generation of content recommendations |
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 |
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 |
CN105069072A (en) * | 2015-07-30 | 2015-11-18 | 天津大学 | Emotional analysis based mixed user scoring information recommendation method and apparatus |
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 |
Cited By (17)
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 |
CN110033351A (en) * | 2019-04-15 | 2019-07-19 | 秒针信息技术有限公司 | A kind of determination method and device of 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 |
CN110807691A (en) * | 2019-10-31 | 2020-02-18 | 深圳市云积分科技有限公司 | Cross-commodity-class commodity recommendation method and device |
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 |
CN112950304A (en) * | 2019-12-11 | 2021-06-11 | 北京沃东天骏信息技术有限公司 | Information pushing 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 |
CN113763065A (en) * | 2020-06-17 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Method, device, equipment and computer readable medium for recommending commodities |
CN113763067A (en) * | 2020-06-19 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Article information pushing method, device, equipment and storage medium |
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 |
CN113592588A (en) * | 2021-07-25 | 2021-11-02 | 北京慧橙信息科技有限公司 | E-commerce platform commodity recommendation system and method based on big data collaborative filtering technology |
CN113592588B (en) * | 2021-07-25 | 2023-10-03 | 深圳市瀚力科技有限公司 | E-commerce platform commodity recommendation system and method based on big data collaborative filtering technology |
CN117333203A (en) * | 2023-12-01 | 2024-01-02 | 广东付惠吧数据服务有限公司 | Member marketing platform combined with business marketing solution |
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 |
Also Published As
Publication number | Publication date |
---|---|
CN109087177B (en) | 2021-05-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109087177A (en) | To the method, apparatus and computer readable storage medium of target user's Recommendations | |
CN107451894B (en) | Data processing method, device and computer readable storage medium | |
KR101726520B1 (en) | Personalized recommendation method, system and computer-readable record medium | |
CN108960992A (en) | A kind of information recommendation method and relevant device | |
WO2016150354A1 (en) | Method and system for classifying users of e-commerce platform | |
CN108665333A (en) | Method of Commodity Recommendation, device, electronic equipment and storage medium | |
CN108153791B (en) | Resource recommendation method and related device | |
CN103886001A (en) | Personalized commodity recommendation system | |
US9183510B1 (en) | Method and system for personalized recommendation of lifestyle items | |
CN102063433A (en) | Method and device for recommending related items | |
CN103646341B (en) | A kind of website provides the recommendation method and apparatus of object | |
CN108713212A (en) | Method, system and medium for providing information based on grouping information | |
WO2019174549A1 (en) | Information recommendation method and apparatus | |
CN105468628B (en) | A kind of sort method and device | |
CN109064293A (en) | Method of Commodity Recommendation, device, computer equipment and storage medium | |
CN106326318B (en) | Searching method and device | |
CN107423430B (en) | Data processing method, device and computer readable storage medium | |
US20150161674A1 (en) | Generation of fashion ensembles based on anchor article | |
CN104615721B (en) | For the method and system based on return of goods related information Recommendations | |
CN104809637A (en) | Commodity recommending method and system realized by computer | |
CN105654307A (en) | Commodity recommendation method based on user feedback and commodity recommendation system | |
CN106934680A (en) | A kind of method and device for business processing | |
CN106776697A (en) | Content recommendation method and device | |
CN105809275A (en) | Item scoring prediction method and apparatus | |
CN106682923A (en) | Commodity adjustment method and commodity adjustment system |
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 |