CN106651542A - Goods recommendation method and apparatus - Google Patents
Goods recommendation method and apparatus Download PDFInfo
- Publication number
- CN106651542A CN106651542A CN201611261973.6A CN201611261973A CN106651542A CN 106651542 A CN106651542 A CN 106651542A CN 201611261973 A CN201611261973 A CN 201611261973A CN 106651542 A CN106651542 A CN 106651542A
- Authority
- CN
- China
- Prior art keywords
- article
- recommended
- behavior
- similarity
- user
- 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 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
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)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention discloses a goods recommendation method and apparatus. The method comprises the steps of calculating goods description similarity of each of to-be-recommended goods and a target goods; calculating behavior similarity of a behavior of each of users for each of the to-be-recommended goods and a behavior of the user for the target goods; calculating a quality indicator of each of the to-be-recommended goods; calculating similarity of each of the to-be-recommended goods and the target goods according to the goods description similarity, the behavior similarity and the quality indicator; and selecting a recommended goods from the to-be-recommended goods according to the calculated similarities. When the goods are recommended to the user, multiple factors are taken into account, the similarities of two goods are considered from multiple aspects, so that the calculated similarity of each of the to-be-recommended goods and the target goods is more accurate.
Description
Technical field
The application is related to Information Technology Agreement field, more particularly to the method and device that a kind of article is recommended.
Background technology
21 century is the epoch of an advanced IT application, and internet is had become indispensable one in people's life
Point, the application software of various functions is even more the life for enriching people.
At present, Internet user can online watch video, shopping, listen music, read etc..But, when user is want in product
It is inconvenient the information oneself liked to be found in various and enormous amount the network information of kind.
In prior art, to user recommend article when generally on the basis of the browsed article of the user.Specifically, if
Benchmark article is A, for article B, then obtains the character description information of two articles respectively, then calculates the word between two articles
The similarity of description information.If similarity is higher, B is recommended into user.But this recommendation method is based only on word content
Similitude, the way of recommendation is single, and the similarity between two articles for calculating is inaccurate, so, need a kind of new recommendation thing
The scheme of product.
The content of the invention
The embodiment of the present application provide a kind of article recommend method and device, to solve prior art in push away to user
Recommend due to being based only on word content similitude when recommending article, the way of recommendation for causing is single, two articles for calculating it
Between the inaccurate problem of similarity.
On the one hand, the embodiment of the present application provides a kind of method that article is recommended, including：
Calculate the article description similarity between each article and the target item in article to be recommended；
Calculate customer group in each user to the behavior of each article in the article to be recommended with the user to institute
State the behavior similarity of the behavior of target item；
Calculate the performance figure of each article in the article to be recommended；
The thing to be recommended is calculated according to the article description similarity, the behavior similarity and the performance figure
The similarity between each article and the target item in product；
Similarity according to acquisition is calculated selects to recommend article from the article to be recommended.
On the other hand, the embodiment of the present application provides the device that a kind of article is recommended, including：
Article description similarity computing module, for calculating article to be recommended in each article and target item between
Article description similarity；
Behavior similarity calculation module, for calculating customer group in each user to each thing in the article to be recommended
Behavior similarity of the behavior of product with the user to the behavior of the target item；
Performance figure computing module, for calculating the article to be recommended in each article performance figure；
Similarity calculation module, for according to the article description similarity, the behavior similarity and the quality
Index calculates the similarity between each article and the target item in the article to be recommended；
Selecting module, for being selected to recommend article from the article to be recommended according to the similarity for calculating acquisition.
The application has the beneficial effect that：In the technical scheme that the embodiment of the present application is provided, in calculating article to be recommended
Article description similarity between each article and target item；Each user is in the article to be recommended in calculating customer group
Each article behavior similarity of the behavior with the user to the behavior of the target item；Calculate the article to be recommended
In each article performance figure；According to the article description similarity, the behavior similarity and the performance figure
Calculate the similarity between each article and the target item in the article to be recommended；According to the similarity for calculating acquisition
Select to recommend article from the article to be recommended.So, when article is recommended to user, the thing between article has been considered
Product description similarity, user realize from many aspects to many factors such as behavior similarity, the performance figures of article itself of article
The similarity between two articles is considered, so so that each article in the article to be recommended for calculating and the target item
Between similarity it is more accurate.
Description of the drawings
Technical scheme in order to be illustrated more clearly that the embodiment of the present application, below will be to making needed for embodiment description
Accompanying drawing is briefly introduced, it should be apparent that, drawings in the following description are only some embodiments of the present application, for this
For the those of ordinary skill in field, on the premise of not paying creative work, can be obtaining other according to these accompanying drawings
Accompanying drawing.
Fig. 1 show the method flow schematic diagram that the article of the offer of the embodiment of the present application one is recommended；
Fig. 2 show the method flow schematic diagram that the article of the offer of the embodiment of the present application two is recommended；
Fig. 3 show the structure drawing of device that the article of the offer of the embodiment of the present application three is recommended；
Fig. 4 show the hardware configuration of the electronic equipment of the method that the article of the offer of the embodiment of the present application five is recommended and illustrates
Figure.
Specific embodiment
In order that the purpose of the application, technical scheme and advantage are clearer, below in conjunction with accompanying drawing the application is made into
One step ground is described in detail, it is clear that described embodiment is only some embodiments of the present application, rather than the enforcement of whole
Example.Based on the embodiment in the application, what those of ordinary skill in the art were obtained under the premise of creative work is not made
All other embodiment, belongs to the scope of the application protection.
Embodiment one：
As described in Figure 1, the schematic flow sheet of the method recommended for a kind of article that the embodiment of the present application is provided, the method bag
Include following steps：
Step 101：Calculate the article description similarity between each article and the target item in article to be recommended.
, wherein it is desired to illustrate, the article described in the embodiment of the present application refers to the displaying object on network, and the displaying is right
As being product, can also be service.The product is, for example, commodity, the APP (Application, application) of exploitation for making
Etc..
Step 102：Calculate in customer group each user to the behavior of each article in the article to be recommended with it is described
Behavior similarity of the user to the behavior of the target item.
Wherein, in one embodiment, the behavior include buying behavior, navigation patterns, collection behavior, marking behavior,
Download behavior etc..Wherein, marking behavior is, for example, thumb up, gives scoring etc..
Step 103：Calculate the performance figure of each article in the article to be recommended.
Step 104：According to the article description similarity, the behavior similarity and the performance figure are calculated
The similarity between each article and the target item in article to be recommended.
Step 105：Similarity according to acquisition is calculated selects to recommend article from the article to be recommended.
For ease of understanding, the method that a kind of article that the embodiment of the present application is provided is recommended is described further below, can
Including herein below：
Wherein, in one embodiment, step 102 can be performed specifically as one of following two modes：
Mode one：
Step A1：The behavior of each article and target item during each user is to article to be recommended in acquisition customer group；And
Distribute behavior coefficient for each behavior.
When being embodied as, the corresponding behavior coefficient of each behavior can be set and configured according to the actual requirements.Of course, it is possible to
The rank of behavior is first set, for example, shopping website is corresponded to, behavior rank order from high to low can be successively：Purchase, receipts
Hide, browse；For the APP websites of service are provided, behavior rank order from high to low can be installed, collect, browsing.
Wherein, the rank of marking behavior can be less than collection behavior but higher than navigation patterns.It should be noted that when being embodied as, can
According to specific behavior and to need setting behavior rank, the embodiment of the present application to be not construed as limiting this.
During concrete setting behavior coefficient, the corresponding coefficient of the higher behavior of rank is higher, when specifically calculating, can adopt level
The corresponding behavior coefficient of other highest behavior.Such as buying behavior is identical with the corresponding behavior coefficient of the behavior of download, and is designated as 3；
Collection behavior is identical with the corresponding behavior coefficient of marking behavior, and is designated as 2；The corresponding behavior coefficient of navigation patterns is designated as 1；No
Behavior is then designated as 0.
Wherein, the behavior for calculating behavior similarity is highest behavior of the user to article.For example user downloads
Certain article, has just necessarily browsed the article, so the behavior of user then includes navigation patterns and downloads behavior, then adopt during calculating
With the behavior coefficient of the behavior of download.
Step A2：For each user in customer group and for each article to be recommended in article to be recommended, meter
The user is calculated to the behavior coefficient of the article to be recommended and the product of the behavior coefficient of target item, using the product for calculating as this
Behavior of the user to the article to be recommended and the similarity of the behavior to target item, we are referred to as individual behavior similarity.
Such as user A is 3 to the behavior coefficient of article A to be recommended, is 2 to the behavior coefficient of target item.So it is directed to
The behavior similarity of user A, article A to be recommended and target item is 6.By that analogy, the individuality of each user can be calculated
Behavior similarity.
Step A3：Each article in for article to be recommended, each user is to the article and object in calculating customer group
The mean value of the individual behavior similarity of product, using the mean value as customer group to the behavior of the article with to the target item
Behavior similarity, we are referred to as behavior similarity.
It is assumed that for the article A in article to be recommended, behavior similarity such as table 1 of each user to article A and target item
Shown, then, customer group is 1.5 to the behavior similarity of article A and target item.
Table 1
More preferably, when being embodied as, can also be by user's heap sort, such as gender-disaggregated or age-based classification etc..
When calculating customer group to the behavior similarity of the behavior of each article in article set and the behavior of target item, can also be concrete
The behavior of each article in calculating each class user to article to be recommended and the individual behavior similarity of the behavior of target item
Mean value, as behavior similarity of such user to the behavior of each article in article set and the behavior of target item.
Mode two：Can carry out data mining to determine behavior similarity using matrix decomposition, specifically：
Step B1：Article set is constituted by article to be recommended and target item, each user is to article collection in acquisition customer group
The behavior of each article in conjunction.
Step B2：Each user in for customer group, according to default behavior and the corresponding relation of behavior coefficient, determines
The user to article set in each article behavior coefficient, and using behavior coefficient as matrix element, determine the expression use
Behavioural matrix of the family to the behavior coefficient of article,
It is assumed that user is determined into behavioural matrix to the behavior coefficient of each article as row vector, then the row for finally determining
It is as shown in table 2 for matrix.Certainly, when being embodied as, it is also possible to determine user as column vector to the behavior coefficient of each article
Go out behavioural matrix.
Table 2
| Article 1 | Article 2 | …… | Article N | |
| User 1 | T11 | T12 | …… | T1N |
| User 2 | T21 | T22 | …… | T2N |
| …… | …… | …… | …… | …… |
| User n | Tn1 | Tn3 | …… | TnN |
Step B3：Matrix decomposition is carried out to user behavior matrix, pre-conditioned split-matrix is met；The default bar
Part includes：The split-matrix it is total n it is vectorial, each vector has d element, wherein, n represents the total of the article in article set
Quantity, d represents preset value.
Wherein, the value of d can be the integer between 40-200.When being embodied as, can determine based on experience value, it is also possible to
Recommendation effect according to subsequently doing when article is recommended determines.
Wherein, when being embodied as, can be using alternating least-squares, triangle decomposition method, QR value decomposition methods, singular value point
The methods such as solution carry out matrix decomposition to user behavior matrix.All methods that can carry out matrix decomposition are fitted in prior art
For the present embodiment, the embodiment of the present application is not limited this.
Step B4：Each article in for article to be recommended, by the article in split-matrix it is corresponding vector and mesh
The vector product of mark article corresponding vector in split-matrix is used as the article and the behavior similarity of target item.
For example, user behavior matrix is being recorded as Am*n, its i-th row element represent i-th user in article set own
The behavior coefficient of article, jth row represent behavior of each user to article j in customer group.To Am*nApproximate factorization is done, is decomposited
Um*d,Vn*d, and meet satisfactionThen article i and article j behavior similarities Sij=ViVj, wherein,
ViRepresent Vn*dI-th vectorial, VjRepresent Vn*dJ-th it is vectorial.If article j is the article in article to be recommended, article i
For target item.The S for then calculatingijAs in customer group each user to the behavior of article j with the user to the object
The behavior similarity of the behavior of product.
So, the principle of data mining is carried out based on matrix decomposition, each user in customer group can be calculated and treated to described
Behavior similarity of the behavior of each article in recommendation article with the user to the behavior of the target item so that calculate
Result can more represent behavior similarity of the public users to two articles.
Wherein, in one embodiment, when article is recommended for user to be recommended (such as first user), in order that recommending
Article more meet the individual demand of first user, improve the use for recommending first user the accuracy and first user of article
Family is experienced, and in the embodiment of the present application, can also carry out following steps：
Step C1：Receive browse request of the first user to the target item.
Step C2：The personal information of the first user is obtained according to the browse request.
Wherein, the personal information of first user includes the row of the log-on message and/or first user of first user to article
For.Log-on message for example includes user's sex, age, the hobby etc..The hobby example that the hobby is input into when being, for example, user's registration
The hobby of such as shopping website is skirt, the video that sweater clothes, video website are registered is liked as suspense class video, terrible class
Video, game class hobby are the game of gunbattle class) etc..
Certainly, the item recommendation method of the embodiment of the present application is applicable not only to the above-mentioned application scenarios enumerated, and can recommend
The scene of article is useful in the protection domain of the application.
Step C3：The recommendation article of the first user to selection is calculated according to the personal information of the first user
In each article preference value.
Wherein, when personal information is log-on message.First customer group can be classified according to log-on message, for example, be pressed
Sex, character classification by age.Then the user preferences for being included according to log-on message, count preference of all types of user to all kinds of articles
Value.If hypothesis is configured to the peak of preference value to have 100 people in 10, one class user, wherein have 80 personal like's skirts, that
The preference value of skirt is 8；There are 60 personal like's sweaters, then the preference value of sweater is 6, by that analogy.Draw such
Preference value of the user to all kinds of articles.So, for first user, can first according to the log-on message of first user, it is determined that
Class of subscriber belonging to first user.For each recommendation article selected in step 105, determine belonging to the recommendation article
Goods categories.Then in the class of subscriber for first user being belonged to, the corresponding preference value of goods categories of the recommendation article is made
For preference value of the first user to the recommendation article.
When being embodied as certainly, the method classified to customer group, the method classified to article and obtaining is used
The method of family hobby can be realized using alternate manner, do not repeated here.
Step C4：The recommendation article of the selection is ranked up according to the preference value.
Wherein, the recommendation article of the selection is ranked up according to the preference value and is e.g. pushed away for each of selection
Article is recommended, the behavior similarity between the recommendation article and the target item is calculated with first user to the inclined of the recommendation article
Good being worth and value.Then it is ranked up according to calculate and value.It is of course also possible to the preference by first user to the recommendation article
Value calculates the product of the behavior similarity between the recommendation article and the target item and its weighted value, so as weighted value
Afterwards, it is ranked up according to product.
Certainly, it should be noted that when being embodied as, can according to the actual requirements set selection is treated using preference value
The scheme for recommending article to be ranked up, the application contrast is not limited.
Step C5：The recommendation article after according to sequence is recommended to the first user.
Wherein, in one embodiment, when the personal information be the first user to the article set in article
Behavior when, step B3 can specifically perform and be：
Each article in for recommending article, according to below equation preference of the first user to the article is determined
Value：
Puj=∑K ∈ (S, u)fukSij
Wherein, j is to recommend the article in article, PujFor preference value of the first user to the article, u is first user, and i is
Target item, it is in article set to implementing the set of the article of highest behavior rank, k in article set that S is first user
K-th article, fukTo represent behavior coefficient of the first user to k-th article, SijFor the behavior of article j and target item i
Similarity.
For example, N number of article is had in article set, first user was implemented the other article of highest behavioral scaling and is designated as respectively
Article 1, article 2 and article 3, behavior coefficient of the first user to each article, respectively are A1, A2, A3, each in customer group
User is A4 to the behavior similarity of article j and target item i.Then the first user of final calculating is to the preference value of article j
(A1+A2+A3)*A4.So, preference value of the user to article is obtained equivalent to the actual hobby according to user so that preference
The determination of value is more more accurate.
Wherein, in one embodiment, in order that the article recommended more meets the demand of user, Consumer's Experience is improved, this
In application embodiment, (article calculated between each article and target item in article to be recommended describes similar to step 101
Degree) can specifically perform and be：
Step D1：Obtain the word description of each article in the article to be recommended.
Step D2：Participle is carried out to the word description.
Step D3：Calculate the term vector of each word obtained after participle.
It is for instance possible to use TF-IDF (a kind of weighting technique prospected for information retrieval and information) is calculated obtaining after participle
The term vector of each word for obtaining.Certainly, all methods that can calculate term vector are applied to the present embodiment in prior art,
In the protection domain of the application, the embodiment of the present application is not limited this.
Step D4：The each article and target item in the article to be recommended is calculated according to the term vector of each word
Between article description similarity.
For example, the article description similarity in article set between two articles can pass through formula
Calculate.Wherein, XijRepresent article description similarity；| | | | square is sought in expression
The norm of battle array, wiRepresent the matrix of the term vector composition of article i to be recommended, wjRepresent the square of the term vector composition of target item j
Battle array.
Wherein, in one embodiment, because the quantity of article is very more, the height such as quality such as cost performance of article is not
One, if some low-quality articles are recommended user, meeting cause the user puzzlement causes recommendation effect poor.So, in order to
Recommendation effect is improved, in the embodiment of the present application, step 103 (calculates the performance figure of each article in the article to be recommended)
Can specifically perform and be：
The performance figure of each article in following at least one described article to be recommended of calculating：Article popularity,
Article grading system, goods review number etc..
Wherein it is possible to will be used to calculate article quality index in article popularity, article grading system, goods review number etc.
Information be normalized；Afterwards using information and as the article the performance figure after normalized.So,
Some high-quality articles can be recommended user, save the time of user, improve Consumer's Experience.
Wherein, in one embodiment, for the demand that the article for further making recommendation more meets user, user's body is improved
Test, step 104 (is waited to push away according to the article description similarity, the behavior similarity and the performance figure are calculated
Recommend the similarity between each article in article and the target item) can specifically perform and be：
Step E1：Determine that the article description similarity, the behavior similarity and the performance figure are each corresponded to
Weighted value.
Step E2：According to the weighted value and the article description similarity, the behavior similarity and described that determine
Performance figure, by way of weighted sum, calculates between each article and the target item in the article to be recommended
Similarity.
In the same manner, the recommendation article of the selection is ranked up according to the preference value, can performs and be：
Step F1：Determine the article description similarity, the behavior similarity, the performance figure and the preference
It is worth each self-corresponding weighted value.
Step F2：The quality according to the weighted value and the article description similarity, the behavior similarity that determine
Index and the preference value, by way of weighted sum, calculate the ranking value of the recommendation article of selection, then according to sequence
Value is ranked up.
Then, according to sequence after the recommendation article carry out recommendation to the first user and can specifically perform be：The row of selection
The recommendation article of L recommends first user before sequence, and L is positive integer.
Wherein, in one embodiment, in order to obtain preferably weighted value, step F1 can specifically perform and be：
Step G1：Determine the article description similarity, the behavior similarity, the performance figure and the preference
It is worth each self-corresponding initial weight value.
Step G2：In each initial weight value that will be obtained, the corresponding initial weight value of the article description similarity is considered as
Parameter to be adjusted during check experiment, obtains first group of weighted value test object；And
In each initial weight value that will be obtained, when the corresponding initial weight value of the behavior similarity is considered as check experiment
Parameter to be adjusted, obtains second group of weighted value test object；And,
In each initial weight value that will be obtained, treating when the corresponding initial weight value of the performance figure is considered as check experiment
Adjusting parameter, obtains the 3rd group of weighted value test object；And,
In each initial weight value that will be obtained, treating when the corresponding initial weight value of the preference value is considered as check experiment is adjusted
Whole parameter, obtains the 4th group of weighted value test object.
Step G3：For each group weighted value test object, check experiment is carried out, and according to check experiment result, determine institute
State article description similarity, the behavior similarity, the performance figure and each self-corresponding weighted value of the preference value.
Wherein, in one embodiment, step G3 can specifically perform and be：
For each group weighted value test object, following operation is performed：
Step H1：Start using preset duration as a measurement period from the scheduled date；And, in first measurement period
Using the initial weight value of this group of weighted value test object as experiment sample, recommending for representing for first measurement period is counted
The value of feedback of effect, the value of feedback is any one information in following information：Clicking rate, click total amount, buying rate, purchase are total
Amount, download rate, download total amount.
Step H2：The parameter to be adjusted of this group of weighted value test object is adjusted by the first preset rules；Wherein, first preset
Rule is to increase or decrease.
Step H3：Using this group of weighted value test object after adjustment is as the experiment sample of next measurement period and counts
Value of feedback of the experiment sample in the next measurement period.
Step H4：For each measurement period outside first measurement period, if the value of feedback of the measurement period is relative
Rise in the value of feedback of a upper measurement period, then return and perform the step of adjusting parameter to be adjusted by the first preset rules；If should
The value of feedback of measurement period declines relative to the value of feedback of a upper measurement period, then adjust this group of weighted value by the second preset rules
The parameter to be adjusted of test object, and execution is returned using this group of weighted value test object after adjustment as next measurement period
Experiment sample and count the experiment sample the step of value of feedback of the next measurement period；Wherein, if the first default rule
Then to increase, then the second preset rules are reduction；If the first preset rules are to reduce, the second preset rules are increase.
Step H5：If the value of feedback in the continuous measurement period of predetermined number meets pre-conditioned, continuous from this
This group of weighted value test object after the corresponding adjustment of highest value of feedback is selected in measurement period as one group of preferred weighted value.
Wherein, the maximum difference of the value of feedback in the pre-conditioned continuous measurement period for referring to predetermined number is poor less than default
Value.That is, it is desirable to which the value of feedback in the continuous measurement period of predetermined number is basically unchanged.
It is for instance possible to use with Gaussian Profile N (0,1) be randomly provided initial weight value, can be with when being embodied as certainly
Check experiment is carried out using the method for other setting initial weight values, is repeated no more here.If the article description similarity,
The each self-corresponding initial weight value of the behavior similarity, the performance figure and the preference value for (2,3,3), initial power
Weight values are designated as W, such as following table：
Table 3
| Test group W | W1 | W2 | W3 | W3 |
| First group | 2+(0.5) | 3 | 3 | 1 |
| Second group | 2 | 3+(0.5) | 3 | 1 |
| 3rd group | 2 | 3 | 3+(0.5) | 1 |
| 4th group | 2 | 3 | 3 | 1+(0.5) |
N days are assumed for a measurement period, determine the daily feedback for feeding back to, then taking n days in a measurement period
The value of feedback of the mean value of the value measurement period the most.Then, the value of feedback of upper a measurement period is contrasted, if current statistic
The value of feedback in cycle rises, then increase corresponding parameter value to be adjusted, for example, improve 0.5, reduces if declining and treats accordingly
Adjusting parameter value, for example, reduce 0.5；1 can be again set to when being reduced to 0.
In sum：In the embodiment of the present application, due between each article and target item in calculating article to be recommended
Article description similarity；Calculate in customer group each user to the behavior of each article in the article to be recommended with it is described
Behavior similarity of the user to the behavior of the target item；The quality for calculating each article in the article to be recommended refers to
Number；Calculated in the article to be recommended according to the article description similarity, the behavior similarity and the performance figure
Each article and the target item between similarity；Similarity according to acquisition is calculated is selected from the article to be recommended
Select recommendation article.So, when article is recommended to user, article description similarity, the user couple between article has been considered
The many factors such as behavior similarity, the performance figure of article itself of article so that each in the article to be recommended for calculating
Similarity between article and the target item is more accurate, can more conform to the actual demand of user, improves user
Experience.
Embodiment two
For ease of the method that the article for further understanding the application offer is recommended, the embodiment of the present application is done into one to the method
Step explanation.As shown in Fig. 2 comprising the following steps：
Step 201：Calculate the article description similarity between each article and the target item in article to be recommended.
Step 202：Calculate in customer group each user to the behavior of each article in the article to be recommended with it is described
Behavior similarity of the user to the behavior of the target item.
Step 203：Calculate the performance figure of each article in the article to be recommended.
Step 204：Determine that the article description similarity, the behavior similarity and the performance figure are each corresponded to
Weighted value.
Step 205：According to the weighted value and the article description similarity, the behavior similarity and described that determine
Performance figure, by way of weighted sum, calculates between each article and the target item in the article to be recommended
Similarity.
Step 206：Similarity according to acquisition is calculated selects to recommend article from the article to be recommended.
Step 207：Receive browse request of the first user to the target item.
Step 208：The personal information of the first user is obtained according to the browse request.
Step 209：The recommendation thing of the first user to selection is calculated according to the personal information of the first user
The preference value of each article in product.
Step 210：Obtain the corresponding weighted value of preference value, and calculate first user to the inclined of each recommendation article for selecting
The product of the weighted value of good value and preference value.
For example, the weighted value of preference value is P, then for each recommendation article, calculate first user to the recommendation article
The product of preference value and P.
Step 211：Calculate first user to product and the article of each article in the recommendation article that selects with
Target item behavior similarity and value, and according to being ranked up to the recommendation article of selection with value.
If the recommendation article for selecting has Q, Q and value are obtained, be ranked up according to the order descending with value.
Step 212：The recommendation article of L before sequence is sent into the first user.
In the technical scheme that the embodiment of the present application is provided, due to each article and object in calculating article to be recommended
Article description similarity between product；Calculate behavior of each user to each article in the article to be recommended in customer group
With behavior similarity of the user to the behavior of the target item；Calculate the matter of each article in the article to be recommended
Volume index；The thing to be recommended is calculated according to the article description similarity, the behavior similarity and the performance figure
The similarity between each article and the target item in product；According to calculate obtain similarity from the article to be recommended
It is middle to select to recommend article.So, when article is recommended to user, the article description similarity between article considered, used
Family is considered from many aspects between two articles by many factors such as behavior similarity, the performance figure of article itself of article, realization
Similarity, so so that the similarity between each article and the target item in the article to be recommended for calculating is more
Plus accurately.
Embodiment three：
Based on identical inventive concept, the embodiment of the present application also provides the device that a kind of article is recommended, the article of the device
The principle that the principle of recommendation is recommended with the article of the method that above-mentioned article is recommended is similar.Specifically can be found in the interior of said method
Hold, do not repeat here.
As shown in figure 3, for the structural representation of the device, described device includes：
Article description similarity computing module 301：Each article in for calculating article to be recommended and target item it
Between article description similarity.
Behavior similarity calculation module 302：It is every during each user is to the article to be recommended in for calculating customer group
Behavior similarity of the behavior of one article with the user to the behavior of the target item.
Performance figure computing module 303：The performance figure of each article in for calculating the article to be recommended.
Similarity calculation module 304：For according to the article description similarity, the behavior similarity and the matter
Volume index calculates the similarity between each article and the target item in the article to be recommended.
Selecting module 305：For being selected to recommend article from the article to be recommended according to the similarity for calculating acquisition.
Wherein, in one embodiment, described device also includes：
Receiver module, for receiving browse request of the first user to the target item；
Personal information acquisition module, for obtaining the personal information of the first user according to the browse request；
Preference value computing module, for calculating the first user to selection according to the personal information of the first user
The preference value of each article in the recommendation article；
Order module, for being ranked up to the recommendation article of the selection according to the preference value；
Recommending module, is recommended for the recommendation article after according to sequence to the first user.
Wherein, in one embodiment, the behavior include buying behavior, navigation patterns, collection behavior, marking behavior,
Download behavior.
Wherein, in one embodiment, the article description similarity computing module, specifically includes：
Word description acquiring unit, for obtaining the article to be recommended in each article word description；
Participle unit, for carrying out participle to the word description；
Term vector computing unit, for calculating participle after obtain each word term vector；
Article description similarity computing unit, for being calculated in the article to be recommended according to the term vector of each word
Each article and target item between article description similarity.
Wherein, in one embodiment, the performance figure computing module, specifically for：
The performance figure of each article in following at least one described article to be recommended of calculating：Article popularity,
Article grading system, goods review number.
In sum, the device that the article that the embodiment of the present application is provided is recommended, article description similarity computing module is calculated
The article description similarity between each article and target item in article to be recommended；Behavior similarity calculation module is calculated to be used
Row of each user to the behavior of each article in the article to be recommended with the user to the target item in the group of family
For behavior similarity；Performance figure computing module calculates the performance figure of each article in the article to be recommended；It is similar
Degree computing module calculates described to be recommended according to the article description similarity, the behavior similarity and the performance figure
The similarity between each article and the target item in article；Selecting module is according to calculating the similarity for obtaining from described
Select to recommend article in article to be recommended.So, when article is recommended to user, the article description between article has been considered
, to many factors such as behavior similarity, the performance figures of article itself of article, realization considers from many aspects two for similarity, user
Similarity between article, so so that between each article and the target item in the article to be recommended for calculating
Similarity is more accurate.
Example IV
The embodiment of the present application four provides a kind of nonvolatile computer storage media, the computer-readable storage medium storage
There are computer executable instructions, the computer executable instructions can perform the side that the article in above-mentioned any means embodiment is recommended
Method.
Embodiment five
Fig. 4 is that the hardware configuration of the electronic equipment for performing the method that article is recommended that the embodiment of the present application five is provided is illustrated
Figure, as shown in figure 4, the electronic equipment includes：
One or more processors 410 and memory 420, in Fig. 4 by taking a processor 410 as an example.Perform article to push away
The electronic equipment of the method recommended can also include：Input unit 430 and output device 440.
Processor 410, memory 420, input unit 430 and output device 440 can be by bus or other modes
Connection, in Fig. 4 as a example by being connected by bus.
Memory 420 can be used to store non-volatile software journey as a kind of non-volatile computer readable storage medium storing program for executing
Sequence, non-volatile computer executable program and module, the corresponding journey of method that the article such as in the embodiment of the present application is recommended
Sequence instruction/module (article description similarity computing module 301, behavior similarity calculation module 302 for example, shown in accompanying drawing 3,
Performance figure computing module 303, similarity calculation module 304, selecting module 305).Processor 410 is stored in by operation
Non-volatile software program, instruction and module in reservoir 420, so as to the various function application and data of execute server
Process, that is, realize the method that said method embodiment article is recommended.
Memory 420 can include storing program area and storage data field, wherein, storing program area can store operation system
Application program required for system, at least one function；Storage data field can store the using for device recommended according to article and be created
Data built etc..Additionally, memory 420 can include high-speed random access memory, nonvolatile memory can also be included,
For example, at least one disk memory, flush memory device or other non-volatile solid state memory parts.In certain embodiments,
Memory 420 is optional including relative to the remotely located memory of processor 410, and these remote memories can be connected by network
It is connected to the device of article recommendation.The example of above-mentioned network includes but is not limited to internet, intranet, LAN, mobile logical
Letter net and combinations thereof.
Input unit 430 can receives input numeral or character information, and produce with article recommend device user
The relevant key signals input of setting and function control.Output device 440 may include the display devices such as display screen.
One or more of modules are stored in the memory 420, when by one or more of processors
During 410 execution, the method that the article in above-mentioned any means embodiment is recommended is performed.
The method that the executable the embodiment of the present application of the said goods is provided, possesses the corresponding functional module of execution method and has
Beneficial effect.Ins and outs of detailed description in the present embodiment, not can be found in the method that the embodiment of the present application is provided.
It will be understood by those skilled in the art that embodiments herein can be provided as method, device (equipment) or computer journey
Sequence product.Therefore, the application can using complete hardware embodiment, complete software embodiment or with reference to software and hardware in terms of
The form of embodiment.And, the application can be adopted and wherein include the calculating of computer usable program code at one or more
The computer program implemented in machine usable storage medium (including but not limited to magnetic disc store, CD-ROM, optical memory etc.)
The form of product.
The application is the flow chart with reference to method, device (equipment) and computer program according to the embodiment of the present application
And/or block diagram is describing.It should be understood that can be by each flow process in computer program instructions flowchart and/or block diagram
And/or the combination of square frame and flow chart and/or the flow process in block diagram and/or square frame.These computer programs can be provided to refer to
The processor of all-purpose computer, special-purpose computer, Embedded Processor or other programmable data processing devices is made to produce
One machine so that produced for realizing by the instruction of computer or the computing device of other programmable data processing devices
The device of the function of specifying in one flow process of flow chart or one square frame of multiple flow processs and/or block diagram or multiple square frames.
These computer program instructions may be alternatively stored in can guide computer or other programmable data processing devices with spy
In determining the computer-readable memory that mode works so that the instruction being stored in the computer-readable memory is produced to be included referring to
Make the manufacture of device, the command device realize in one flow process of flow chart or one square frame of multiple flow processs and/or block diagram or
The function of specifying in multiple square frames.
These computer program instructions also can be loaded into computer or other programmable data processing devices so that in meter
Series of operation steps is performed on calculation machine or other programmable devices to produce computer implemented process, so as in computer or
The instruction performed on other programmable devices is provided for realizing in one flow process of flow chart or multiple flow processs and/or block diagram one
The step of function of specifying in individual square frame or multiple square frames.
Although having been described for the preferred embodiment of the application, those skilled in the art once know basic creation
Property concept, then can make other change and modification to these embodiments.So, claims are intended to be construed to include excellent
Select embodiment and fall into having altered and changing for the application scope.
Obviously, those skilled in the art can carry out the essence of various changes and modification without deviating from the application to the application
God and scope.So, if these modifications of the application and modification belong to the scope of the application claim and its equivalent technologies
Within, then the application is also intended to comprising these changes and modification.
Claims (10)
1. a kind of method that article is recommended, it is characterised in that methods described includes：
Calculate the article description similarity between each article and the target item in article to be recommended；
Calculate customer group in each user to the behavior of each article in the article to be recommended with the user to the mesh
The behavior similarity of the behavior of mark article；
Calculate the performance figure of each article in the article to be recommended；
Calculated in the article to be recommended according to the article description similarity, the behavior similarity and the performance figure
Each article and the target item between similarity；
Similarity according to acquisition is calculated selects to recommend article from the article to be recommended.
2. method according to claim 1, it is characterised in that methods described also includes：
Receive browse request of the first user to the target item；
The personal information of the first user is obtained according to the browse request；
Each thing in the recommendation article of the first user to selection is calculated according to the personal information of the first user
The preference value of product；
The recommendation article of the selection is ranked up according to the preference value；
The recommendation article after according to sequence is recommended to the first user.
3. method according to claim 1 and 2, it is characterised in that the behavior includes buying behavior, navigation patterns, receipts
Tibetan behavior, marking behavior, download behavior.
4. method according to claim 1 and 2, it is characterised in that each article in calculating article to be recommended with
Article description similarity between target item, including：
Obtain the word description of each article in the article to be recommended；
Participle is carried out to the word description；
Calculate the term vector of each word obtained after participle；
The article between each article and target item in the article to be recommended is calculated according to the term vector of each word
Description similarity.
5. method according to claim 1 and 2, it is characterised in that each thing in the calculating article to be recommended
The performance figure of product, including：
The performance figure of each article in following at least one described article to be recommended of calculating：Article popularity, article
Grading system, goods review number.
6. the device that a kind of article is recommended, it is characterised in that described device includes：
Article description similarity computing module, for calculating article to be recommended in each article and target item between article
Description similarity；
Behavior similarity calculation module, for calculating customer group in each user to each article in the article to be recommended
Behavior similarity of the behavior with the user to the behavior of the target item；
Performance figure computing module, for calculating the article to be recommended in each article performance figure；
Similarity calculation module, for according to the article description similarity, the behavior similarity and the performance figure
Calculate the similarity between each article and the target item in the article to be recommended；
Selecting module, for being selected to recommend article from the article to be recommended according to the similarity for calculating acquisition.
7. device according to claim 6, it is characterised in that described device also includes：
Receiver module, for receiving browse request of the first user to the target item；
Personal information acquisition module, for obtaining the personal information of the first user according to the browse request；
Preference value computing module, for calculating the first user to described in selection according to the personal information of the first user
Recommend the preference value of each article in article；
Order module, for being ranked up to the recommendation article of the selection according to the preference value；
Recommending module, is recommended for the recommendation article after according to sequence to the first user.
8. the device according to claim 6 or 7, it is characterised in that the behavior includes buying behavior, navigation patterns, receipts
Tibetan behavior, marking behavior, download behavior.
9. the device according to claim 6 or 7, it is characterised in that the article description similarity computing module, concrete bag
Include：
Word description acquiring unit, for obtaining the article to be recommended in each article word description；
Participle unit, for carrying out participle to the word description；
Term vector computing unit, for calculating participle after obtain each word term vector；
Article description similarity computing unit, for calculating every in the article to be recommended according to the term vector of each word
Article description similarity between one article and target item.
10. the device according to claim 6 or 7, it is characterised in that the performance figure computing module, specifically for：
The performance figure of each article in following at least one described article to be recommended of calculating：Article popularity, article
Grading system, goods review number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611261973.6A CN106651542B (en) | 2016-12-31 | 2016-12-31 | Article recommendation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611261973.6A CN106651542B (en) | 2016-12-31 | 2016-12-31 | Article recommendation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106651542A true CN106651542A (en) | 2017-05-10 |
| CN106651542B CN106651542B (en) | 2021-06-25 |
Family
ID=58838876
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611261973.6A Active CN106651542B (en) | 2016-12-31 | 2016-12-31 | Article recommendation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106651542B (en) |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107169842A (en) * | 2017-05-31 | 2017-09-15 | 合肥亿迈杰软件有限公司 | A kind of electronic commerce data screening system based on commodity data |
| CN107230136A (en) * | 2017-05-31 | 2017-10-03 | 合肥亿迈杰软件有限公司 | A kind of shopping sequence method for pushing based on big data |
| CN107391692A (en) * | 2017-07-26 | 2017-11-24 | 腾讯科技（北京）有限公司 | The appraisal procedure and device of a kind of recommendation effect |
| CN107679943A (en) * | 2017-09-27 | 2018-02-09 | 广州市万表科技股份有限公司 | Intelligent recommendation method and system |
| CN108109052A (en) * | 2017-12-29 | 2018-06-01 | 广州优视网络科技有限公司 | Article method for pushing, device and server |
| CN108171569A (en) * | 2017-12-11 | 2018-06-15 | 武汉纺织大学 | A kind of reaction type intelligent clothing recommends method and system |
| CN108600792A (en) * | 2018-04-02 | 2018-09-28 | 武汉斗鱼网络科技有限公司 | A kind of method for measuring similarity, device, equipment and storage medium |
| CN108765051A (en) * | 2018-04-16 | 2018-11-06 | 达而观信息科技（上海）有限公司 | item recommendation method and device |
| CN108763318A (en) * | 2018-04-27 | 2018-11-06 | 达而观信息科技（上海）有限公司 | item recommendation method and device |
| CN109033233A (en) * | 2018-06-29 | 2018-12-18 | 武汉斗鱼网络科技有限公司 | A kind of direct broadcasting room recommended method, storage medium, electronic equipment and system |
| CN109101563A (en) * | 2018-07-13 | 2018-12-28 | 东软集团股份有限公司 | A kind of object recommendation method, apparatus, medium and equipment |
| CN109102127A (en) * | 2018-08-31 | 2018-12-28 | 杭州贝购科技有限公司 | Method of Commodity Recommendation and device |
| CN109840825A (en) * | 2017-11-29 | 2019-06-04 | 迪特技术公司 | The recommender system of physical features based on user |
| CN110110206A (en) * | 2018-01-17 | 2019-08-09 | 阿里巴巴集团控股有限公司 | Relation excavation and recommended method, device, calculating equipment, storage medium between article |
| CN110473043A (en) * | 2018-05-11 | 2019-11-19 | 北京京东尚科信息技术有限公司 | A method and device for item recommendation based on user behavior |
| CN110738538A (en) * | 2018-07-18 | 2020-01-31 | 北京京东尚科信息技术有限公司 | Method and apparatus for identifying similar items |
| WO2020135189A1 (en) * | 2018-12-29 | 2020-07-02 | 深圳Tcl新技术有限公司 | Product recommendation method, product recommendation system and storage medium |
| CN111666481A (en) * | 2019-03-05 | 2020-09-15 | 北京沃东天骏信息技术有限公司 | Data mining method, apparatus, computer readable medium and electronic device |
| CN112182397A (en) * | 2020-10-12 | 2021-01-05 | 北京自如信息科技有限公司 | Item recommendation method, device, electronic device and readable storage medium |
| CN112907312A (en) * | 2019-12-04 | 2021-06-04 | 深圳云天励飞技术有限公司 | Article recommendation method and system for offline user and storage medium |
| CN113781147A (en) * | 2020-11-17 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Item recommendation method and device |
| CN115249164A (en) * | 2021-04-27 | 2022-10-28 | 中国移动通信集团有限公司 | Commodity recommendation method and device, electronic equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103514239A (en) * | 2012-11-26 | 2014-01-15 | Tcl美国研究所 | Recommendation method and system integrating user behaviors and object content |
| CN103514304A (en) * | 2013-10-29 | 2014-01-15 | 海南大学 | Project recommendation method and device |
| CN103679489A (en) * | 2012-09-10 | 2014-03-26 | 腾讯科技（深圳）有限公司 | Item-based collaborative recommendation method and system |
| US20150149295A1 (en) * | 2013-11-25 | 2015-05-28 | Kobo Incorporated | Method and system for generating recommendations based on media usage and purchase behavior |
| CN104765751A (en) * | 2014-01-07 | 2015-07-08 | 腾讯科技（深圳）有限公司 | Application recommendation method and device |
| CN105608180A (en) * | 2015-12-22 | 2016-05-25 | 北京奇虎科技有限公司 | Application recommendation method and system |
-
2016
- 2016-12-31 CN CN201611261973.6A patent/CN106651542B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103679489A (en) * | 2012-09-10 | 2014-03-26 | 腾讯科技（深圳）有限公司 | Item-based collaborative recommendation method and system |
| CN103514239A (en) * | 2012-11-26 | 2014-01-15 | Tcl美国研究所 | Recommendation method and system integrating user behaviors and object content |
| CN103514304A (en) * | 2013-10-29 | 2014-01-15 | 海南大学 | Project recommendation method and device |
| US20150149295A1 (en) * | 2013-11-25 | 2015-05-28 | Kobo Incorporated | Method and system for generating recommendations based on media usage and purchase behavior |
| CN104765751A (en) * | 2014-01-07 | 2015-07-08 | 腾讯科技（深圳）有限公司 | Application recommendation method and device |
| CN105608180A (en) * | 2015-12-22 | 2016-05-25 | 北京奇虎科技有限公司 | Application recommendation method and system |
Cited By (33)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107169842A (en) * | 2017-05-31 | 2017-09-15 | 合肥亿迈杰软件有限公司 | A kind of electronic commerce data screening system based on commodity data |
| CN107230136A (en) * | 2017-05-31 | 2017-10-03 | 合肥亿迈杰软件有限公司 | A kind of shopping sequence method for pushing based on big data |
| CN107391692A (en) * | 2017-07-26 | 2017-11-24 | 腾讯科技（北京）有限公司 | The appraisal procedure and device of a kind of recommendation effect |
| CN107391692B (en) * | 2017-07-26 | 2023-04-07 | 腾讯科技（北京）有限公司 | Recommendation effect evaluation method and device |
| CN107679943A (en) * | 2017-09-27 | 2018-02-09 | 广州市万表科技股份有限公司 | Intelligent recommendation method and system |
| US12118602B2 (en) | 2017-11-29 | 2024-10-15 | Ditto Technologies, Inc. | Recommendation system, method and computer program product based on a user's physical features |
| CN109840825A (en) * | 2017-11-29 | 2019-06-04 | 迪特技术公司 | The recommender system of physical features based on user |
| US11157985B2 (en) | 2017-11-29 | 2021-10-26 | Ditto Technologies, Inc. | Recommendation system, method and computer program product based on a user's physical features |
| CN108171569B (en) * | 2017-12-11 | 2021-12-21 | 武汉纺织大学 | Feedback type intelligent clothing recommendation method and system |
| CN108171569A (en) * | 2017-12-11 | 2018-06-15 | 武汉纺织大学 | A kind of reaction type intelligent clothing recommends method and system |
| CN108109052A (en) * | 2017-12-29 | 2018-06-01 | 广州优视网络科技有限公司 | Article method for pushing, device and server |
| CN110110206B (en) * | 2018-01-17 | 2023-10-31 | 斑马智行网络(香港)有限公司 | Method, device, computing equipment and storage medium for mining and recommending relationships among articles |
| CN110110206A (en) * | 2018-01-17 | 2019-08-09 | 阿里巴巴集团控股有限公司 | Relation excavation and recommended method, device, calculating equipment, storage medium between article |
| CN108600792B (en) * | 2018-04-02 | 2020-08-04 | 武汉斗鱼网络科技有限公司 | Similarity measurement method, device, equipment and storage medium |
| CN108600792A (en) * | 2018-04-02 | 2018-09-28 | 武汉斗鱼网络科技有限公司 | A kind of method for measuring similarity, device, equipment and storage medium |
| CN108765051B (en) * | 2018-04-16 | 2022-03-22 | 达而观信息科技（上海）有限公司 | Item recommendation method and device |
| CN108765051A (en) * | 2018-04-16 | 2018-11-06 | 达而观信息科技（上海）有限公司 | item recommendation method and device |
| CN108763318A (en) * | 2018-04-27 | 2018-11-06 | 达而观信息科技（上海）有限公司 | item recommendation method and device |
| CN110473043A (en) * | 2018-05-11 | 2019-11-19 | 北京京东尚科信息技术有限公司 | A method and device for item recommendation based on user behavior |
| CN109033233B (en) * | 2018-06-29 | 2021-05-28 | 武汉斗鱼网络科技有限公司 | A method, storage medium, electronic device and system for recommending a live broadcast room |
| CN109033233A (en) * | 2018-06-29 | 2018-12-18 | 武汉斗鱼网络科技有限公司 | A kind of direct broadcasting room recommended method, storage medium, electronic equipment and system |
| CN109101563A (en) * | 2018-07-13 | 2018-12-28 | 东软集团股份有限公司 | A kind of object recommendation method, apparatus, medium and equipment |
| CN110738538A (en) * | 2018-07-18 | 2020-01-31 | 北京京东尚科信息技术有限公司 | Method and apparatus for identifying similar items |
| CN110738538B (en) * | 2018-07-18 | 2024-05-24 | 北京京东尚科信息技术有限公司 | Method and device for identifying similar objects |
| CN109102127A (en) * | 2018-08-31 | 2018-12-28 | 杭州贝购科技有限公司 | Method of Commodity Recommendation and device |
| CN109102127B (en) * | 2018-08-31 | 2021-10-26 | 杭州贝购科技有限公司 | Commodity recommendation method and device |
| WO2020135189A1 (en) * | 2018-12-29 | 2020-07-02 | 深圳Tcl新技术有限公司 | Product recommendation method, product recommendation system and storage medium |
| CN111666481B (en) * | 2019-03-05 | 2024-04-09 | 北京沃东天骏信息技术有限公司 | Data mining method, device, computer readable medium and electronic device |
| CN111666481A (en) * | 2019-03-05 | 2020-09-15 | 北京沃东天骏信息技术有限公司 | Data mining method, apparatus, computer readable medium and electronic device |
| CN112907312A (en) * | 2019-12-04 | 2021-06-04 | 深圳云天励飞技术有限公司 | Article recommendation method and system for offline user and storage medium |
| CN112182397A (en) * | 2020-10-12 | 2021-01-05 | 北京自如信息科技有限公司 | Item recommendation method, device, electronic device and readable storage medium |
| CN113781147A (en) * | 2020-11-17 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Item recommendation method and device |
| CN115249164A (en) * | 2021-04-27 | 2022-10-28 | 中国移动通信集团有限公司 | Commodity recommendation method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106651542B (en) | 2021-06-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106651542A (en) | Goods recommendation method and apparatus | |
| CN104090919B (en) | Advertisement recommending method and advertisement recommending server | |
| Johnson | Logistic matrix factorization for implicit feedback data | |
| Li et al. | Using multidimensional clustering based collaborative filtering approach improving recommendation diversity | |
| CN107451894B (en) | Data processing method, device and computer readable storage medium | |
| Teo et al. | Adaptive, personalized diversity for visual discovery | |
| US8442984B1 (en) | Website quality signal generation | |
| Chang et al. | Using groups of items for preference elicitation in recommender systems | |
| CN108629665A (en) | A kind of individual commodity recommendation method and system | |
| CN104331459B (en) | A kind of network resource recommended method and device based on on-line study | |
| CN109492180A (en) | Resource recommendation method, device, computer equipment and computer readable storage medium | |
| CN107562818A (en) | Information recommendation system and method | |
| CN107944986A (en) | A kind of O2O Method of Commodity Recommendation, system and equipment | |
| Wang et al. | Utilizing marginal net utility for recommendation in e-commerce | |
| CN107169052A (en) | Recommend method and device | |
| WO2015034850A2 (en) | Feature selection for recommender systems | |
| CN103246672A (en) | Method and device for performing personalized recommendation on users | |
| CN109064293A (en) | Method of Commodity Recommendation, device, computer equipment and storage medium | |
| CN109711931A (en) | Method of Commodity Recommendation, device, equipment and storage medium based on user's portrait | |
| CN112488863A (en) | Dangerous seed recommendation method and related equipment in user cold start scene | |
| CN104182392A (en) | Method and device for processing service recommendation | |
| CN103309894B (en) | Based on search implementation method and the system of user property | |
| CN113407854A (en) | Application recommendation method, device and equipment and computer readable storage medium | |
| CN105868254A (en) | Information recommendation method and apparatus | |
| CN105468628B (en) | A kind of sort method and device |
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 |