CN108205775A - The recommendation method, apparatus and client of a kind of business object - Google Patents
The recommendation method, apparatus and client of a kind of business object Download PDFInfo
- Publication number
- CN108205775A CN108205775A CN201611185242.8A CN201611185242A CN108205775A CN 108205775 A CN108205775 A CN 108205775A CN 201611185242 A CN201611185242 A CN 201611185242A CN 108205775 A CN108205775 A CN 108205775A
- Authority
- CN
- China
- Prior art keywords
- seed
- user
- preference
- specified services
- similarity
- 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.)
- Pending
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the present application provides the recommendation method, apparatus and client of a kind of business object, including：Determine the seed object there are interbehavior data with user；Calculate seed preference of the user to the seed object；Calculate the seed similarity between specified services object and the seed object；Using the seed preference and the seed similarity, Target Preference degree of the user to specified services object is determined；Recommend the specified services object for the user according to the Target Preference degree.Multiple information dimensions of user behavior data are merged in the embodiment of the present application so that the accuracy rate for calculating preference and similarity is high, so as to preferably be user service.
Description
Technical field
The invention relates to technical field of data processing, a kind of recommendation method more particularly to business object, one
Recommendation apparatus, a kind of determining method of business object similarity, a kind of determining dress of business object similarity of kind business object
Put, a kind of client, a kind of device and, one or more computer-readable mediums.
Background technology
In Internet technology, website is frequently necessary to recommend various products information to user, such as electric business platform is in webpage
On may interested commodity or shop etc. to user recommended user.By way of this recommendation, institute is found to shorten user
The path in commodity or shop is needed, promotes user experience.
The behavior of electric business platform consumer has the characteristics that fragment, contingency, and businessman safeguards the difficulty of frequent customer very
Greatly, therefore latent visitor's marketing is emphasis in businessman's operation always.For businessman, the visitor that dives includes two types, Yi Zhongshi
That is, with it relationship occurred for directly latent visitor, but and be not converted into the latent visitor directly bought, this certain customers is relative to the whole network businessman
For it is very limited.It is also another, it is the user for having demand but not establishing direct links.Since there is no direct users
Behavioral data is available, therefore the diffusion technique excavated usually by some similar to collaborative filtering class for the visitor that dives to this part, i.e.,
Using synergetic establish commodity to commodity, businessman to businessman, user to user similarity data, then based on some triggering sides
Method markets to target customer.Measurement of the collaborative filtering to user preference and the accuracy to Similarity measures, directly
It connects and is related to the ROI (Return On Investment, rate of return on investment) of marketing and the core of algorithm optimization.
Invention content
In view of the above problems, it is proposed that the embodiment of the present application overcomes the above problem or at least partly in order to provide one kind
A kind of recommendation method of the business object to solve the above problems, a kind of recommendation apparatus of business object, a kind of business object are similar
The determining method of degree, a kind of determining device of business object similarity, a kind of client, a kind of device and, it is one or more
Computer-readable medium.
To solve the above-mentioned problems, this application discloses a kind of recommendation method of business object, including：
Determine the seed object there are interbehavior data with user；
Calculate seed preference of the user to the seed object；
Calculate the seed similarity between specified services object and the seed object；
Using the seed preference and the seed similarity, determine that the user is inclined to the target of specified services object
Good degree；
Recommend the specified services object for the user according to the Target Preference degree.
Optionally, it is described to calculate user the step of seed preference of the seed object is included：
Obtain the interbehavior data between user and the seed object；
Using interbehavior data generation behavioural characteristic vector；
The behavioural characteristic vector is inputted into preset training pattern, it is inclined to the seed of the seed object to obtain user
Good degree.
Optionally, it is described calculate between specified services object and the seed object seed similarity the step of wrap
It includes：
Obtain the preference matrix between user and specified services object；
Comprehensive preference point is predicted using preset training pattern；
Using the comprehensive preference point update preference matrix；
It is similar to the seed between the seed object that specified services object is calculated using updated preference matrix
Degree.
Optionally, the preset training pattern is trained in the following way：
Obtain within a specified time user to the interbehavior data of business object；
Business scenario is recommended to determine optimization aim according to current；
According to the optimization aim characteristic is extracted from the interbehavior data；
The training pattern is trained using the characteristic.
Optionally, it is described calculate between specified services object and the seed object seed similarity the step of wrap
It includes：
Obtain the time difference factor；
It is similar to the seed between the seed object that the time difference factor is introduced into calculating specified services object
Degree.
Optionally, the step of acquisition time difference factor includes：
There are the first interaction times of interbehavior data between acquisition user and seed object；
There are the second interaction times of interbehavior data between acquisition user and specified services object；
The time difference factor is calculated using first interaction time and second interaction time.
Optionally, it is described to use the seed preference and the seed similarity, determine the user to specified services
The step of Target Preference of object is spent includes：
Using the total seed similarity of the seed similarity calculation；
It is updated using total seed similarity and the seed similarity, obtains updated seed similarity,
Using the seed preference and updated seed similarity, mesh of the user to specified services object is determined
Mark preference.
Optionally, the business object includes commodity or shop.
The embodiment of the present application also discloses a kind of recommendation apparatus of business object, including：
Seed object determining module, for determining the seed object with user there are interbehavior data；
Seed preference computing module, for calculating seed preference of the user to the seed object；
Seed similarity calculation module, for calculating the seed phase between specified services object and the seed object
Like degree；
Target Preference degree determining module for using the seed preference and the seed similarity, determines the use
Family is to the Target Preference degree of specified services object；
Specified services object recommendation module, for recommending the specified services for the user according to the Target Preference degree
Object.
The embodiment of the present application also discloses a kind of determining method of business object preference, including：
Determine the seed object there are interbehavior data with user；
Calculate seed preference of the user to the seed object；
Calculate the seed similarity between specified services object and the seed object；
Using the seed preference and the seed similarity, determine that the user is inclined to the target of specified services object
Good degree.
The embodiment of the present application also discloses a kind of determining device of business object preference, including：
Seed object determining module, for determining the seed object with user there are interbehavior data；
Seed preference computing module, for calculating seed preference of the user to the seed object；
Seed similarity calculation module, for calculating the seed phase between specified services object and the seed object
Like degree；
Target Preference degree determining module for using the seed preference and the seed similarity, determines the use
Family is to the Target Preference degree of specified services object.
The embodiment of the present application also discloses a kind of recommendation method of business object, including：
Business object is sent to server and obtains request, and described obtain asks to include user information；
The server is received for the specified services object for obtaining request feedback, wherein, the specified services pair
As determining seed object according to user information for the server and determining specified industry according to the seed object
It is engaged in after the Target Preference degree of object, the specified services object recommended according to the Target Preference degree.
The embodiment of the present application also discloses a kind of client, including：
Request sending module is obtained, obtains request for sending business object to server, the acquisition request includes using
Family information；
Business object receiving module, for receiving the server for the specified services pair for obtaining request feedback
As, wherein, the specified services object determines seed object and according to described for the server according to user information
After seed object determines the Target Preference degree of specified services object, according to the specified services pair of Target Preference degree recommendation
As.
The embodiment of the present application also discloses a kind of device, including：
One or more processors；With
Instruction in the one or more computer-readable mediums stored thereon, is performed by one or more of processors
When, described device is caused to perform above-mentioned method.
The embodiment of the present application also discloses one or more computer-readable mediums, is stored thereon with instruction, when by one
Or multiple processors are when performing so that terminal device performs above-mentioned method.
The embodiment of the present application includes advantages below：
The embodiment of the present application can determine the seed object there are interbehavior data, Ran Houji with user first
Calculate seed preference that user is to seed and, the seed similarity of specified services object and seed object,
It is finally based on seed preference and seed similarity determines Target Preference degree of the user to specified services object, it is inclined when obtaining target
After good degree, it is possible to recommend specified services object for user according to Target Preference degree, merge user in the embodiment of the present application
Multiple information dimensions of behavioral data so that the accuracy rate for calculating preference and similarity is high, so as to preferably be user
Service.
Description of the drawings
Fig. 1 is a kind of step flow chart of the recommendation embodiment of the method for business object of the application；
Fig. 2 is a kind of step flow chart of model training of the application；
Fig. 3 is a kind of commercial product recommending algorithm implementing procedure figure of the application；
Fig. 4 is a kind of step flow chart of the determining embodiment of the method for business object preference of the application；
Fig. 5 is the step flow chart that a kind of client-side of the application realizes the recommendation embodiment of the method for business object；
Fig. 6 is a kind of structure diagram of the recommendation apparatus embodiment of business object of the application；
Fig. 7 is a kind of structure diagram of the determining device embodiment of business object preference of the application；
Fig. 8 is a kind of structure diagram of client embodiment of the application.
Specific embodiment
Above-mentioned purpose, feature and advantage to enable the application are more obvious understandable, below in conjunction with the accompanying drawings and specific real
Mode is applied to be described in further detail the application.
Internet has the inherent advantage for collecting data, can monitor the full link behavioral data of user, these behaviors one
As in the case of all than sparse, and many users do not only have a kind of behavioral data, but occur in the form of behavioral data cluster.This
Outside, the information content that each behavioral data is covered is different, that is, reflects that the preference of client is different.For example scoring is the straight of user
The hobby of user is directly reacted in reversed feedback, and click is a kind of implicit preferences of user.How each behavioral data is measured, with
And each behavioral data how is comprehensively utilized, it is the key that determination data utilization ratio.
In previous research, relatively common is the following two kinds behavioral data Land use systems：
1st, it separately utilizes：
The different behavioral data of user, such as navigation patterns data, buying behavior data are used separately, based on each behavior
Data individually calculate similarity, and output is similar to " bought also buy, seen see also ".
2nd, simple weighted：
The behavioral data different to user simply assigns a weight, and weighting generates the preference measurement of a fusion, then
Preference measurement based on this synthesis carries out similarity calculation.It is this significantly to show user preferences such as to the comment of user
Larger weight coefficient is given in behavior, and smaller weight coefficient is then assigned to the detour behaviors such as clicking, browsing.
However, aforementioned two methods, cannot all utilize behavioral data well.First method is not to behavioral data
It is comprehensively utilized, second method employs subjective linear weighted function when behavioral data is merged, to various actions data
Measurement and fusion be comparison it is subjective roughly.
In fact, it is can be found that through analyzing the relationship between different behavioral datas：
1st, the unit of each behavioral data and scale difference, the value of different behavioral datas may differ by very greatly, reflect
Preference it is also different
2nd, it is not independence between each behavioral data, a such as user is after purchase behavior is added, then browses a quotient
Product, representing this user has had this commodity very strong purchase intention, purchase or browsing is added to show partially than simple
Good degree is stronger.
3rd, each behavioral data reflects that preference is not linear relationship, as number of visits reflects preference more 10 times
Degree is not necessarily 10 times of 1 time of browsing.
4th, for also for each behavioral data, simple counting can not extract behavioral data completely and believe accordingly
Breath such as navigation patterns data, in addition to number of visits, can also extract the number of days of browsing, the accounting browsed with classification etc. letter
Breath, than individual number of visits, more reflects the preference information of user.
5th, the behavioral data of user has very strong contingency, therefore matrix is inherently sparse, such as a user
To a commodity feedback-less, user is not represented and does not like this commodity, but since user has no chance to see this commodity, such as
It is 0 that fruit, which is directly considered as preference, is inappropriate, it is necessary to be smoothed here.
6th, in addition, in electric business field business model complexity, user behavior data link is longer, and the type of behavioral data is richer
Richness, in addition to clicking, browsing, collecting plus the information such as purchase, purchase, comment, while when can also obtain stop of the user to commodity
Between, click the information of the more details such as picture number.
7th, other than joint act data of the user on commodity, user is also very heavy to the time of the act difference of two commodity
It will.
Based on Such analysis, the embodiment of the present application proposes to introduce more user behavior datas, more becomes more meticulous to behavior
Information extraction merges more extensive information, if user likes commodity the extensive hobby to user's affiliated classification to commodity, with
Nonlinear model merges the information of extraction, is preferably fitted the non-linear relation of behavior and preference, shows simultaneously
The intersection information gone on a journey between variable.
Specifically, can non-linear partial can preferably be handled by training mode in the embodiment of the present application, below it is right
It is introduced in the background context and particular content of the training pattern of the application.
Collaborative filtering is a kind of typical big data thinking algorithm, utilizes the behavior number of the mass users group of historical accumulation
According to the incidence relation between searching things and things in statistical significance is diffused to realize commodity by these incidence relations
The applications such as recommendation.
In the realization and application of collaborative filtering, most crucial work includes three parts：
1st, behavioral data of the user to entity (such as shop or commodity) is collected as far as possible；
2nd, the similarity between entity is accurately measured；
3rd, rational diffusion is touched up to mechanism.
It is diffused final using collaborative filtering in application, can consider the similarity sim (i, i ') between entity simultaneously,
And user is to the preference r of kind of fructificationu,i′, the precision of the two measure of criterions, the final effect for determining diffusion.
Wherein, the factor of similarity calculation precision includes two parts：
1st, preference information；
2nd, the reasonability of similarity calculation.
In specific research, the researcher about collaborative filtering focuses on the metric algorithm of optimization similitude mostly
On, such as Cosine cosine similarities：
Wherein, i, j refer to i-th of entity and j-th of entity, and x represents some entity, and r represents preference.
Pearson correlation coefficient：
Wherein, m, n refer to m-th of entity and n-th of entity, and i represents some entity, and r represents preference.
Certainly, similarity algorithm can also include other algorithms, such as the more increasingly complex heat based on bigraph (bipartite graph) passes
Lead algorithm, LogOddsRatio etc..
In contrast, for how preferably less using the research for inputting message context.In fact, each similarity
Basic data all corresponds to behavioural matrix of the user to entity：
Element in above-mentioned behavioural matrix may due to different business scopes and scene difference, such as have that user's is straight
The behavioral data of comment marking, the behavioral data for also having purchase, the collection of user etc. obvious are connect, these behavioral datas are exactly
The information that collaborative filtering needs utilize, the validity of these information input, itself just plays decisive work to the measurement of similarity
With.
In order to which preferably service user, the Utilizing question proposition that the embodiment of the present application is directed to behavioral data in collaborative filtering are several
A view of innovation：
1st, by the way of information extraction, user behavior data is further refined from multiple angles, improves user behavior number
According to utilization rate.
2nd, in a manner that modeling is merged, nonlinear fitting is carried out to the preference that various actions data reflect, is provided
More accurately comprehensive measurement, preference during preference matrix and diffusion applied to collaboration.
3rd, more extensive information are incorporated in a model, are smoothly filled to part of not giving a mark to user.
4th, time difference coefficient is introduced in similarity algorithm, preferably embodies the contact between user behavior data.
Wherein, information extraction technique is extended by Feature Engineering technology on the user behaviors log Information base of user
Processing, from more perspective description information, to promote the utilization rate of information.Extensive technology refers to put forward information to more high-dimensional
Refining, obtains more high-dimensional generality information, can solve the sparse sex chromosome mosaicism of information.
In the embodiment of the present application non-linear fusion part using GBDT (Gradient Boosting Decision Tree,
Gradient promotes decision tree) algorithm, which is a Boosting algorithm, and loss function, the application instruction are represented with L (F (X), Y)
The target for practicing model learning is to find an optimal function to make total loss function minimum：
F*=argminFL(Y,F(x))
Wherein, Y represents training objective actual value, and F (x) represents to predict the prediction of training objective according to input feature value x
Function.Loss function L (Y, F (x)) is the function of difference between actual value and predicted value, smaller to represent to predict more accurate, mould
The process of type training is exactly to find optimal prediction model function F*, make loss function minimum.
GBDT algorithms carry out model solution in function space, can be expressed as loss function：
L (Y, F (x))=φ (F (X))
The renewal process of training pattern in the embodiment of the present application is：
Fm(X)=Fm-1(X)+ρ*f(x)
According to function additive property principle, optimal function is a series of the cumulative of increment functions, can be by increment iterative more
New mode acquires optimal models F*.Wherein Fm(X) pattern function of m steps, F are representedm-1(X) mould of single-step iteration thereon is represented
Type function, ρ represent the speed that study rate coefficient control loss function declines.F (x) represents increment function.Declined according to gradient former
Reason, increment function f (x) can reduce loss function most fast using the negative gradient of pattern function：
So as to walk iteration by M, the optimal solution for acquiring training pattern is：
That is, optimal function is a series of the cumulative of increment function f (x), due to increment function can not direct solution, algorithm is first
Values of the f (x) on each sample point is first calculated, uses giIt represents：
Wherein giIt is that f (x) is worth on i-th of sample point, i.e., pattern function F (X) is in the negative gradient of sample i.
Then the g being fitted on each sample is removed using decision treei, the increment function g that is fittedm(X), then training is updated
Model：
Fm(X)=Fm-1(X)+ρ*gm(X)
Wherein gm(X) it is to be fitted negative gradient gs of the F (X) on each sample using decision treeiObtained function, ρ represent to learn
Practise rate coefficient.Optimum prediction function F is being obtained by GBDT model learnings*(x) after, based on input feature value x
Obtain predicted value F*(x).
Prediction target is preference point in the embodiment of the present application, is represented with purchase probability.It is according to input behavior data
(feature) the feature vector x of composition, passes through F*(x) purchase probability of the user to commodity is obtained, represents its preference to commodity
Degree.
It contributes each time in this way and goes the negative gradient of fitting "current" model, continue to optimize model.Since each step is with tree mould
Type is fitted, and is exported and codetermined by more trees, so be a nonlinear training pattern, it is non-thread with fitting precision height
The advantages that property, anti-over-fitting.
It is extracted again by behavioral data and behavioral data cluster carries out modeling fusion, training pattern can be based on to user
The preference of entity is predicted, obtains synthesis preference of the user to entity, and user-entity preference square is filled with this
Battle array, smooth sky element, then carry out the calculating of similarity.Improve the calculating of similarity and the quantified precision of user preference degree simultaneously.
It should be noted that aforementioned training pattern is only a kind of example of training pattern of the application, implementing this
When applying for embodiment, modeling processing can be carried out using other modes, the embodiment of the present application does not limit this.
With reference to Fig. 1, show a kind of step flow chart of the recommendation embodiment of the method for business object of the application, specifically may be used
To include the following steps：
Step 101, the seed object with user there are interbehavior data is determined；
It should be noted that the business object in the embodiment of the present application can include the specific things in different business field,
Such as commodity or shop etc..Wherein, seed object refers to that there are the business of interbehavior between user and business object
Object, for example, the commodity that user clicked can be confirmed to be seed commodity.It is appreciated that there are interbehavior user and
Interbehavior data are corresponding between business object.
For those skilled in the art is made to more fully understand the embodiment of the present application, in the present specification, mainly using commodity and
Retail shop illustrates as business object.
Step 102, seed preference of the user to the seed object is calculated；
The seed preference of user can express fancy grade of the user to seed object.
In a preferred embodiment of the present application, the step 102 can include following sub-step：
Sub-step S11 obtains the interbehavior data between user and the seed object；
Sub-step S12, using interbehavior data generation behavioural characteristic vector；
The behavioural characteristic vector is inputted preset training pattern, obtains user to the seed pair by sub-step S13
The seed preference of elephant.
User, can be by the interaction row between user and seed object for the seed preference of seed object
It is behavioural characteristic vector for data quantization, then behavioural characteristic vector is input to the preset training pattern of the embodiment of the present application
In, you can obtain seed preference of the user for the seed object.
Step 103, the seed similarity between specified services object and the seed object is calculated；
In a preferred embodiment of the present application, the step 103 can include following sub-step：
Sub-step S21 obtains the preference matrix between user and specified services object；
Sub-step S22 predicts comprehensive preference point using preset training pattern；
Sub-step S23, using the comprehensive preference point update preference matrix；
Sub-step S24 is calculated using updated preference matrix between specified services object and the seed object
Seed similarity.
In the embodiment of the present application, the preference square between user and specified services object is updated using preset training pattern
Battle array, updated preference matrix has merged more user behavior datas, therefore is calculated on the basis of newer training pattern
Similarity, the similarity precision higher of gained.
In a preferred embodiment of the present application, the preset training pattern is trained in the following way：
Step a1 obtains training pattern；
Step a2, obtain within a specified time user to the interbehavior data of business object；
Step a3 recommends business scenario to determine optimization aim according to current；
Step a4 extracts characteristic according to the optimization aim from the interbehavior data；
Step a5 trains the training pattern using the characteristic.
In the embodiment of the present application, specifically can training be optimized for training pattern in the following way：Base first
User's various types of behavioral datas to commodity interior for a period of time are extracted in the daily record behavioral data of user, it is specific as follows：
In electric business field, the type of behavioral data can include search, browsing, collection plus purchase, purchase, comment, consulting,
Residence time, details page picture check, check comment etc., during the embodiment of the present application is trained model,
Further information extraction slot can be carried out to part behavioral data, by taking user is to the browsing in shop as an example, can further taken out
Take out following data：
User-shop number of visits
User-shop browsing commodity number
User-shop browsing number of days
Across the day number of visits in user-shop
…
After completing to the extension of behavioral data, it is possible to it is further extensive to feature progress, it realizes for behavior square
The battle array smooth filling of (preference matrix) without behavior, for example, user can be incorporated to the affiliated classification of shop commodity, user to shop quotient
The affiliated brand of product and shop itself, the affiliated classification of commodity, affiliated brand feature (feature) etc., in user couple
Shop does not have to provide smoothing processing in the case of direct action.
The training data such as lower structure is obtained after information extraction and feature extension is completed：
| user_id | item_id | feature1 | feature2 | feature3 | … | featuren |
With reference to shown in Fig. 2, after training data is obtained, the training and prediction of training pattern need to undergo following steps：
1, training dataset；
2, specify optimization aim；
3, data cleansing, sampling；
4, test set, model training, prediction marking.
For different application scenarios, different optimization aims can be selected, such as to buy as target, training pattern is given
What is gone out is transition probability of the behavior prediction user based on user to purchase, to represent synthesis preference journey of the user to shop
Degree.Prediction marking based on model, can obtain following newer user-shop preference matrix, pay attention to here for each use
Family-shop provides a unified prediction probability point in pairs, wherein, it is the portion that user has shop direct action to blacken part
Point, it is then that training pattern provides the part smoothly filled without blackening part：
| seller_id1 | seller_id2 | seller_id3 | seller_id4 | seller_id5 | seller_id6 | |
| user_id1 | Pre_score | Pre_score | Pre_score | Pre_score | Pre_score | Pre_score |
| user_id2 | Pre_score | Pre_score | Pre_score | Pre_score | Pre_score | Pre_score |
| user_id3 | Pre_score | Pre_score | Pre_score | Pre_score | Pre_score | Pre_score |
| user_id4 | Pre_score | Pre_score | Pre_score | Pre_score | Pre_score | Pre_score |
| user_id5 | Pre_score | Pre_score | Pre_score | Pre_score | Pre_score | Pre_score |
| user_id6 | Pre_score | Pre_score | Pre_score | Pre_score | Pre_score | Pre_score |
To sum up, the embodiment of the present application is obtaining Behavior-based control aggregate of data, flat by unified non-linear fusion and process
After the sliding updated user-shop preference matrix of processing, it is possible to carry out the calculating of similarity.Use cosine similarity public affairs
Formula, moleculeFor table to the inner product of the user preference vector of shop i and shop j, inner product is bigger, represents and uses
Family is higher to the common preference in two shops.
In a preferred embodiment of the present application, the step 103 calculates specified services object and the seed industry
Business object between seed similarity the step of may include steps of：
Sub-step S31 obtains the time difference factor；
The time difference factor is introduced and calculated between specified services object and the seed object by sub-step S32
Seed similarity.
Cosine similarity formula can be used to calculate similarity in the embodiment of the present application, examined, it can be seen that cosine is public
It is poor to the time of the act in two shops that user is not accounted in formula, and the time difference of two shop user concerns gets in practice
It is small, it should to show higher common interest-degree.Therefore the embodiment of the present application introduces a time difference factor delta in the molecule
(Tx,i-Tx,j) to get to：
Time difference saturation can be designed as：
δ(Tj-Ti)=exp (- θ * (Tj-Ti))
Wherein, θ is adjustment factor, and to the adjustment time rate of decay, which shows as, poor increasing at any time during beginning
It is long, decline it is especially fast, behind gradually ease up.
Step 104, using the seed preference and the seed similarity, determine the user to specified services object
Target Preference degree；
In the concrete realization, if necessary to prediction user u to the Target Preference degree p of shop iu,i, finding out user first has directly
The seed shop i ' of behavior is met, preference r of the measure user to each seed shop i ' is distinguished based on training patternu,i′, multiplied by
With seed shop i ' and the similarity sim (i, ' i) of shop i, add up on all seed commodity, you can obtain required
Target Preference degree pu,i。
In a preferred embodiment of the present application, the step 104 can include following sub-step：
Sub-step S41, using the total seed similarity of the seed similarity calculation；
Sub-step S42 is updated using total seed similarity and the seed similarity, obtains updated kind
Sub- similarity,
Sub-step S43 using the seed preference and updated seed similarity, determines the user to specifying industry
The Target Preference degree of business object.
In the embodiment of the present application, in order to prevent popular shop always by preferential recommendation, denominator divided by shop i and all kinds
The sum of the similarity in sub- shop, wherein ru,i′Unified Model can be based on to predict：
By this way, reduce popular shop always by the situation of preferential recommendation.
Step 105, recommend the specified services object for the user according to the Target Preference degree.
After the Target Preference degree that user is determined, it is possible to the business pair for recommending to specify for user according to Target Preference degree
As.
The embodiment of the present application can determine the seed object there are interbehavior data, Ran Houji with user first
Calculate seed preference that user is to seed and, the seed similarity of specified services object and seed object,
It is finally based on seed preference and seed similarity determines Target Preference degree of the user to specified services object, it is inclined when obtaining target
After good degree, it is possible to recommend specified services object for user according to Target Preference degree, merge user in the embodiment of the present application
Multiple information dimensions of behavioral data so that the accuracy rate for calculating preference and similarity is high, so as to preferably be user
Service.
In order to which those skilled in the art is made to more fully understand the application, pushing away for commodity is realized below for the embodiment of the present application
The algorithmic procedure recommended is introduced, and the application algorithm implementing procedure figure is specifically as shown in Figure 3：
1, collect the behavioral data of user-commodity；
2, behavioral data is extracted to carry out feature extension；
3, specify optimization aim；
4, training pattern is trained；
5, preference of the user to commodity projection is obtained based on training pattern；
6, the preference based on the prediction refills user-commodity preference matrix；
7, calculate the time difference factor between user-commodity；
8, similarity calculation is carried out, and finally obtain Target Preference based on the time difference factor and updated preference matrix
Degree；
9, it is user's Recommendations based on the Target Preference degree.
The embodiment of the present application uses information extraction and Modeling on Optimal, and user shop behavioral data is expanded, right
Different types of behavioral data provides unified accurate measurement, while sparse behavioural matrix is smoothly filled, in similarity
The time difference factor is introduced in calculating, while optimizes similarity and preference in collaborative filtering diffusion and calculates, final promotion is entire
The data precision of system.
With reference to Fig. 4, a kind of step flow chart of the determining embodiment of the method for business object preference of the application is shown,
Specifically it may include steps of：
Step 201, the seed object with user there are interbehavior data is determined；
Step 202, seed preference of the user to the seed object is calculated；
Step 203, the seed similarity between specified services object and the seed object is calculated；
Step 204, using the seed preference and the seed similarity, determine the user to specified services object
Target Preference degree.
The embodiment of the present application can determine the seed object there are interbehavior data, Ran Houji with user first
Calculate seed preference that user is to seed and, the seed similarity of specified services object and seed object,
It is finally based on seed preference and seed similarity determines Target Preference degree of the user to specified services object.
In the concrete realization, specified services object can be recommended for user according to Target Preference degree according to mesh, due to target
Preference is calculated to have merged multiple information dimensions of user behavior data, it is therefore provided that the accuracy rate recommended improves,
So as to preferably be user service.
With reference to Fig. 5, show that a kind of business object of the application recommends the step flow chart of embodiment of the method, it specifically can be with
Include the following steps：
Step 301, it sends business object to server and obtains request, described obtain asks to include user information；
In the embodiment of the present application, for the user at client, client can be logged at it or logged on specified
It during the page (such as logging on the page of commodity, shop, classification), is sent to server and obtains request, pushed away with request for the user
Recommend business object.
Step 302, the server is received for the specified services object for obtaining request feedback, wherein, the finger
Determine business object for the server according to user information determine seed object and according to the seed object it is true
After the Target Preference degree for determining specified services object, according to the specified services object of Target Preference degree recommendation.
After server receives acquisition request, it is possible to according to the user information for obtaining request, go to determine seed pair
After determining the Target Preference degree of specified services object as and according to seed object, according to the finger of Target Preference degree recommendation
Business object is determined, finally to user feedback specified services object.
Using the embodiment of the present application, electric business platform can be enabled preferably to recommend business object for client user,
Achieve the purpose that two-win.
It should be noted that for embodiment of the method, in order to be briefly described, therefore it is all expressed as to a series of action group
It closes, but those skilled in the art should know, the embodiment of the present application is not limited by described sequence of movement, because according to
According to the embodiment of the present application, certain steps may be used other sequences or be carried out at the same time.Secondly, those skilled in the art also should
Know, embodiment described in this description belongs to preferred embodiment, and involved action not necessarily the application is implemented
Necessary to example.
With reference to Fig. 6, a kind of structure diagram of the recommendation apparatus embodiment of business object of the application is shown, it specifically can be with
Including following module：
Seed object determining module 401, for determining the seed object with user there are interbehavior data；
Seed preference computing module 402, for calculating seed preference of the user to the seed object；
Seed similarity calculation module 403, for calculating the kind between specified services object and the seed object
Sub- similarity；
Target Preference degree determining module 404 for using the seed preference and the seed similarity, determines described
User is to the Target Preference degree of specified services object；
Specified services object recommendation module 405, it is described specified for recommending according to the Target Preference degree for the user
Business object.
In a preferred embodiment of the present application, the business object can include commodity or shop.
In a preferred embodiment of the present application, the seed preference computing module 402 includes：
Interbehavior data acquisition submodule, for obtaining the interbehavior number between user and the seed object
According to；
Feature vector generates submodule, for using interbehavior data generation behavioural characteristic vector；
Seed preference obtains submodule, for the behavioural characteristic vector to be inputted preset training pattern, obtains user
To the seed preference of the seed object.
In a preferred embodiment of the present application, the seed similarity calculation module 203 includes：
Preference matrix acquisition submodule, for obtaining the preference matrix between user and specified services object；
Preference matrix updates submodule, for updating the preference matrix using preset training pattern；
Seed similarity calculation submodule, for calculating specified services object and described kind using updated preference matrix
Seed similarity between subservice object.
In a preferred embodiment of the present application, described device further includes：
Training pattern generation module, for generating training pattern；
Interbehavior data acquisition module, for obtain within a specified time user to the interbehavior number of business object
According to；
Optimization aim determining module, for business scenario being recommended to determine optimization aim according to current；
Characteristic abstraction module, for extracting characteristic from the interbehavior data according to the optimization aim
According to；
Model training module, for training the training pattern using the characteristic.
In a preferred embodiment of the present application, the seed similarity calculation module 403 includes：
Time difference factor acquisition submodule, for obtaining the time difference factor；
Seed similarity calculation submodule calculates specified services object and described kind for the time difference factor to be introduced
Seed similarity between subservice object.
In a preferred embodiment of the present application, the time difference factor acquisition submodule includes：
First interaction time acquiring unit, for obtaining, there are interbehavior data between user and seed object
First interaction time；
Second interaction time acquiring unit, for obtaining, there are interbehavior data between user and specified services object
Second interaction time；
Time difference factor calculating unit, for calculating the time using first interaction time and second interaction time
The poor factor.
In a preferred embodiment of the present application, the Target Preference degree determining module 405 includes：
Total seed similarity calculation submodule, for using the total seed similarity of the seed similarity calculation；
Similarity updates submodule, for being updated using total seed similarity and the seed similarity, obtains
To updated seed similarity,
Target Preference degree computational submodule for using the seed preference and updated seed similarity, determines
The user is to the Target Preference degree of specified services object.
With reference to Fig. 7, a kind of structure diagram of the determining device embodiment of business object preference of the application, tool are shown
Body can include following module：
Seed object determining module 501, for determining the seed object with user there are interbehavior data；
Seed preference computing module 502, for calculating seed preference of the user to the seed object；
Seed similarity calculation module 503, for calculating the kind between specified services object and the seed object
Sub- similarity；
Target Preference degree determining module 504 for using the seed preference and the seed similarity, determines described
User is to the Target Preference degree of specified services object.
With reference to Fig. 8, a kind of structure diagram of client embodiment of the application is shown, can specifically include such as lower die
Block：
Request sending module 601 is obtained, obtains request for sending business object to server, the acquisition request includes
User information；
Business object receiving module 602, for receiving the server for the specified services for obtaining request feedback
Object, wherein, the specified services object determines seed object and according to institute for the server according to user information
After stating the Target Preference degree that seed object determines specified services object, according to the specified services of Target Preference degree recommendation
Object.
For device and client embodiment, since it is basicly similar to embodiment of the method, so the comparison of description
Simply, the relevent part can refer to the partial explaination of embodiments of method.
The present embodiment additionally provides a kind of non-volatile readable storage medium, and one or more is stored in the storage medium
Module (programs) when the one or more module is used in the equipment for touching screen, can cause the equipment to hold
The instruction (instructions) of row following steps：
Determine the seed object there are interbehavior data with user；
Calculate seed preference of the user to the seed object；
Calculate the seed similarity between specified services object and the seed object；
Using the seed preference and the seed similarity, determine that the user is inclined to the target of specified services object
Good degree；
Recommend the specified services object for the user according to the Target Preference degree.
Optionally, it is described to calculate user the step of seed preference of the seed object is included：
Obtain the interbehavior data between user and the seed object；
Using interbehavior data generation behavioural characteristic vector；
The behavioural characteristic vector is inputted into preset training pattern, it is inclined to the seed of the seed object to obtain user
Good degree.
Optionally, it is described calculate between specified services object and the seed object seed similarity the step of wrap
It includes：
Obtain the preference matrix between user and specified services object；
Comprehensive preference point is predicted using preset training pattern；
Using the comprehensive preference point update preference matrix；
It is similar to the seed between the seed object that specified services object is calculated using updated preference matrix
Degree.
Optionally, the preset training pattern is trained in the following way：
Obtain within a specified time user to the interbehavior data of business object；
Business scenario is recommended to determine optimization aim according to current；
According to the optimization aim characteristic is extracted from the interbehavior data；
The training pattern is trained using the characteristic.
Optionally, it is described calculate between specified services object and the seed object seed similarity the step of wrap
It includes：
Obtain the time difference factor；
It is similar to the seed between the seed object that the time difference factor is introduced into calculating specified services object
Degree.
Optionally, the step of acquisition time difference factor includes：
There are the first interaction times of interbehavior data between acquisition user and seed object；
There are the second interaction times of interbehavior data between acquisition user and specified services object；
The time difference factor is calculated using first interaction time and second interaction time.
Optionally, it is described to use the seed preference and the seed similarity, determine the user to specified services
The step of Target Preference of object is spent includes：
Using the total seed similarity of the seed similarity calculation；
It is updated using total seed similarity and the seed similarity, obtains updated seed similarity,
Using the seed preference and updated seed similarity, mesh of the user to specified services object is determined
Mark preference.
Optionally, the business object includes commodity or shop.
Each embodiment in this specification is described by the way of progressive, the highlights of each of the examples are with
The difference of other embodiment, just to refer each other for identical similar part between each embodiment.
It should be understood by those skilled in the art that, the embodiment of the embodiment of the present application can be provided as method, apparatus or calculate
Machine program product.Therefore, the embodiment of the present application can be used complete hardware embodiment, complete software embodiment or combine software and
The form of the embodiment of hardware aspect.Moreover, the embodiment of the present application can be used one or more wherein include computer can
With in the computer-usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) of program code
The form of the computer program product of implementation.
In a typical configuration, the computer equipment includes one or more processors (CPU), input/output
Interface, network interface and memory.Memory may include the volatile memory in computer-readable medium, random access memory
The forms such as device (RAM) and/or Nonvolatile memory, such as read-only memory (ROM) or flash memory (flash RAM).Memory is to calculate
The example of machine readable medium.Computer-readable medium includes permanent and non-permanent, removable and non-removable media can be with
Realize that information stores by any method or technique.Information can be computer-readable instruction, data structure, the module of program or
Other data.The example of the storage medium of computer includes, but are not limited to phase transition internal memory (PRAM), static RAM
(SRAM), dynamic random access memory (DRAM), other kinds of random access memory (RAM), read-only memory
(ROM), electrically erasable programmable read-only memory (EEPROM), fast flash memory bank or other memory techniques, CD-ROM are read-only
Memory (CD-ROM), digital versatile disc (DVD) or other optical storages, magnetic tape cassette, tape magnetic rigid disk storage or
Other magnetic storage apparatus or any other non-transmission medium, available for storing the information that can be accessed by a computing device.According to
Herein defines, and computer-readable medium does not include the computer readable media (transitory media) of non-standing, such as
The data-signal and carrier wave of modulation.
The embodiment of the present application is with reference to according to the method for the embodiment of the present application, terminal device (system) and computer program
The flowchart and/or the block diagram of product describes.It should be understood that flowchart and/or the block diagram can be realized by computer program instructions
In each flow and/or block and flowchart and/or the block diagram in flow and/or box combination.These can be provided
Computer program instructions are set to all-purpose computer, special purpose computer, Embedded Processor or other programmable data processing terminals
Standby processor is to generate a machine so that is held by the processor of computer or other programmable data processing terminal equipments
Capable instruction generation is used to implement in one flow of flow chart or multiple flows and/or one box of block diagram or multiple boxes
The device for the function of specifying.
These computer program instructions, which may also be stored in, can guide computer or other programmable data processing terminal equipments
In the computer-readable memory to work in a specific way so that the instruction being stored in the computer-readable memory generates packet
The manufacture of command device is included, which realizes in one flow of flow chart or multiple flows and/or one side of block diagram
The function of being specified in frame or multiple boxes.
These computer program instructions can be also loaded into computer or other programmable data processing terminal equipments so that
Series of operation steps are performed on computer or other programmable terminal equipments to generate computer implemented processing, thus
The instruction offer performed on computer or other programmable terminal equipments is used to implement in one flow of flow chart or multiple flows
And/or specified in one box of block diagram or multiple boxes function the step of.
Although the preferred embodiment of the embodiment of the present application has been described, those skilled in the art once know base
This creative concept can then make these embodiments other change and modification.So appended claims are intended to be construed to
Including preferred embodiment and fall into all change and modification of the embodiment of the present application range.
Finally, it is to be noted that, herein, relational terms such as first and second and the like be used merely to by
One entity or operation are distinguished with another entity or operation, without necessarily requiring or implying these entities or operation
Between there are any actual relationship or orders.Moreover, term " comprising ", "comprising" or its any other variant meaning
Covering non-exclusive inclusion, so that process, method, article or terminal device including a series of elements are not only wrapped
Those elements are included, but also including other elements that are not explicitly listed or are further included as this process, method, article
Or the element that terminal device is intrinsic.In the absence of more restrictions, it is wanted by what sentence "including a ..." limited
Element, it is not excluded that also there are other identical elements in the process including the element, method, article or terminal device.
Recommendation method to a kind of business object provided herein and a kind of recommendation apparatus of business object above, into
It has gone and has been discussed in detail, the principle and implementation of this application are described for specific case used herein, implements above
The explanation of example is merely used to help understand the present processes and its core concept；Meanwhile for the general technology people of this field
Member, according to the thought of the application, there will be changes in specific embodiments and applications, in conclusion this explanation
Book content should not be construed as the limitation to the application.
Claims (15)
1. a kind of recommendation method of business object, which is characterized in that including：
Determine the seed object there are interbehavior data with user；
Calculate seed preference of the user to the seed object；
Calculate the seed similarity between specified services object and the seed object；
Using the seed preference and the seed similarity, Target Preference of the user to specified services object is determined
Degree；
Recommend the specified services object for the user according to the Target Preference degree.
2. according to the method described in claim 1, it is characterized in that, described calculate seed of the user to the seed object
The step of preference, includes：
Obtain the interbehavior data between user and the seed object；
Using interbehavior data generation behavioural characteristic vector；
The behavioural characteristic vector is inputted into preset training pattern, obtains seed preference of the user to the seed object
Degree.
3. according to the method described in claim 1, it is characterized in that, the calculating specified services object and the seed pair
The step of seed similarity as between, includes：
Obtain the preference matrix between user and specified services object；
Comprehensive preference point is predicted using preset training pattern；
Using the comprehensive preference point update preference matrix；
Seed similarity between specified services object and the seed object is calculated using updated preference matrix.
4. according to the method in claim 2 or 3, which is characterized in that the preset training pattern carries out in the following way
Training：
Obtain within a specified time user to the interbehavior data of business object；
Business scenario is recommended to determine optimization aim according to current；
According to the optimization aim characteristic is extracted from the interbehavior data；
The training pattern is trained using the characteristic.
5. according to the method described in claim 1, it is characterized in that, the calculating specified services object and the seed pair
The step of seed similarity as between, includes：
Obtain the time difference factor；
The time difference factor is introduced to the seed similarity calculated between specified services object and the seed object.
6. according to the method described in claim 5, it is characterized in that, it is described acquisition the time difference factor the step of include：
There are the first interaction times of interbehavior data between acquisition user and seed object；
There are the second interaction times of interbehavior data between acquisition user and specified services object；
The time difference factor is calculated using first interaction time and second interaction time.
It is 7. according to the method described in claim 1, it is characterized in that, described similar with the seed using the seed preference
Degree, determines that the step of user spends the Target Preference of specified services object includes：
Using the total seed similarity of the seed similarity calculation；
It is updated using total seed similarity and the seed similarity, obtains updated seed similarity,
Using the seed preference and updated seed similarity, determine that the user is inclined to the target of specified services object
Good degree.
8. according to the method described in claim 1, it is characterized in that, the business object includes commodity or shop.
9. a kind of recommendation apparatus of business object, which is characterized in that including：
Seed object determining module, for determining the seed object with user there are interbehavior data；
Seed preference computing module, for calculating seed preference of the user to the seed object；
Seed similarity calculation module, it is similar for calculating the seed between specified services object and the seed object
Degree；
Target Preference degree determining module for using the seed preference and the seed similarity, determines the user couple
The Target Preference degree of specified services object；
Specified services object recommendation module, for recommending the specified services pair for the user according to the Target Preference degree
As.
A kind of 10. determining method of business object preference, which is characterized in that including：
Determine the seed object there are interbehavior data with user；
Calculate seed preference of the user to the seed object；
Calculate the seed similarity between specified services object and the seed object；
Using the seed preference and the seed similarity, Target Preference of the user to specified services object is determined
Degree.
11. a kind of determining device of business object preference, which is characterized in that including：
Seed object determining module, for determining the seed object with user there are interbehavior data；
Seed preference computing module, for calculating seed preference of the user to the seed object；
Seed similarity calculation module, it is similar for calculating the seed between specified services object and the seed object
Degree；
Target Preference degree determining module for using the seed preference and the seed similarity, determines the user couple
The Target Preference degree of specified services object.
12. a kind of recommendation method of business object, which is characterized in that including：
Business object is sent to server and obtains request, and described obtain asks to include user information；
The server is received for the specified services object for obtaining request feedback, wherein, the specified services object is
The server determines seed object according to user information and determines specified services pair according to the seed object
After the Target Preference degree of elephant, according to the specified services object of Target Preference degree recommendation.
13. a kind of client, which is characterized in that including：
Request sending module is obtained, obtains request for sending business object to server, the acquisition request includes user's letter
Breath；
Business object receiving module obtains the specified services object for asking to feed back for receiving the server for described,
In, the specified services object determines seed object and according to the seed for the server according to user information
After business object determines the Target Preference degree of specified services object, according to the specified services object of Target Preference degree recommendation.
14. a kind of device, which is characterized in that including：
One or more processors；With
Instruction in the one or more computer-readable mediums stored thereon, when being performed by one or more of processors,
Described device is caused to perform the method such as claim 1-9.
15. one or more computer-readable mediums, are stored thereon with instruction, when executed by one or more processors, make
The method for obtaining the claim 1-9 that terminal device performs.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611185242.8A CN108205775A (en) | 2016-12-20 | 2016-12-20 | The recommendation method, apparatus and client of a kind of business object |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611185242.8A CN108205775A (en) | 2016-12-20 | 2016-12-20 | The recommendation method, apparatus and client of a kind of business object |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108205775A true CN108205775A (en) | 2018-06-26 |
Family
ID=62603353
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611185242.8A Pending CN108205775A (en) | 2016-12-20 | 2016-12-20 | The recommendation method, apparatus and client of a kind of business object |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108205775A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109710739A (en) * | 2018-12-24 | 2019-05-03 | 联想(北京)有限公司 | A kind of information processing method and device, storage medium |
| CN109726664A (en) * | 2018-12-24 | 2019-05-07 | 出门问问信息科技有限公司 | A kind of intelligence dial plate recommended method, system, equipment and storage medium |
| CN109767264A (en) * | 2018-12-20 | 2019-05-17 | 深圳壹账通智能科技有限公司 | Product data method for pushing, device, computer equipment and storage medium |
| CN110097394A (en) * | 2019-03-27 | 2019-08-06 | 青岛高校信息产业股份有限公司 | The latent objective recommended method of product and device |
| CN110415039A (en) * | 2019-07-31 | 2019-11-05 | 北京三快在线科技有限公司 | The method and device of business processing |
| CN110796505A (en) * | 2018-08-03 | 2020-02-14 | 阿里巴巴集团控股有限公司 | Service object recommendation method and device |
| CN111292154A (en) * | 2018-12-10 | 2020-06-16 | 阿里巴巴集团控股有限公司 | Information obtaining method and device |
| CN111639268A (en) * | 2020-06-01 | 2020-09-08 | 上海大学 | User similarity calculation method |
| CN112818377A (en) * | 2019-11-18 | 2021-05-18 | 广东美云智数科技有限公司 | Authority data recommendation method, authority setting method, authority data recommendation system, authority setting system, electronic device and medium |
| CN113486249A (en) * | 2021-07-27 | 2021-10-08 | 中国银行股份有限公司 | Mobile banking business recommendation method and device |
| CN113674010A (en) * | 2020-05-15 | 2021-11-19 | 北京沃东天骏信息技术有限公司 | Information processing method, device and storage medium |
| CN113888291A (en) * | 2021-12-08 | 2022-01-04 | 北京达佳互联信息技术有限公司 | Virtual space processing method and device, electronic equipment and storage medium |
| CN115408601A (en) * | 2022-05-18 | 2022-11-29 | 安胜旗信息咨询(深圳)有限公司 | Labeling method, device, medium and equipment based on user operation behaviors |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1588400A (en) * | 2004-10-21 | 2005-03-02 | 上海交通大学 | Method for recommending goods in electronic business |
| CN103455522A (en) * | 2012-06-04 | 2013-12-18 | 北京搜狗科技发展有限公司 | Recommendation method and system of application extension tools |
| CN103678709A (en) * | 2013-12-30 | 2014-03-26 | 中国科学院自动化研究所 | Recommendation system attack detection method based on time series data |
| CN104424247A (en) * | 2013-08-28 | 2015-03-18 | 北京闹米科技有限公司 | Product information filtering recommendation method and device |
| CN105260458A (en) * | 2015-10-15 | 2016-01-20 | 海信集团有限公司 | Video recommendation method for display apparatus and display apparatus |
| CN106202328A (en) * | 2016-07-01 | 2016-12-07 | 中国传媒大学 | A kind of recommendation method for new projects' cold start-up |
-
2016
- 2016-12-20 CN CN201611185242.8A patent/CN108205775A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1588400A (en) * | 2004-10-21 | 2005-03-02 | 上海交通大学 | Method for recommending goods in electronic business |
| CN103455522A (en) * | 2012-06-04 | 2013-12-18 | 北京搜狗科技发展有限公司 | Recommendation method and system of application extension tools |
| CN104424247A (en) * | 2013-08-28 | 2015-03-18 | 北京闹米科技有限公司 | Product information filtering recommendation method and device |
| CN103678709A (en) * | 2013-12-30 | 2014-03-26 | 中国科学院自动化研究所 | Recommendation system attack detection method based on time series data |
| CN105260458A (en) * | 2015-10-15 | 2016-01-20 | 海信集团有限公司 | Video recommendation method for display apparatus and display apparatus |
| CN106202328A (en) * | 2016-07-01 | 2016-12-07 | 中国传媒大学 | A kind of recommendation method for new projects' cold start-up |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110796505A (en) * | 2018-08-03 | 2020-02-14 | 阿里巴巴集团控股有限公司 | Service object recommendation method and device |
| CN110796505B (en) * | 2018-08-03 | 2023-07-04 | 淘宝（中国）软件有限公司 | Business object recommendation method and device |
| CN111292154A (en) * | 2018-12-10 | 2020-06-16 | 阿里巴巴集团控股有限公司 | Information obtaining method and device |
| CN109767264A (en) * | 2018-12-20 | 2019-05-17 | 深圳壹账通智能科技有限公司 | Product data method for pushing, device, computer equipment and storage medium |
| CN109726664B (en) * | 2018-12-24 | 2021-07-09 | 出门问问信息科技有限公司 | Intelligent dial recommendation method, system, equipment and storage medium |
| CN109726664A (en) * | 2018-12-24 | 2019-05-07 | 出门问问信息科技有限公司 | A kind of intelligence dial plate recommended method, system, equipment and storage medium |
| CN109710739A (en) * | 2018-12-24 | 2019-05-03 | 联想(北京)有限公司 | A kind of information processing method and device, storage medium |
| CN109710739B (en) * | 2018-12-24 | 2020-12-18 | 联想(北京)有限公司 | Information processing method and device and storage medium |
| CN110097394A (en) * | 2019-03-27 | 2019-08-06 | 青岛高校信息产业股份有限公司 | The latent objective recommended method of product and device |
| CN110415039A (en) * | 2019-07-31 | 2019-11-05 | 北京三快在线科技有限公司 | The method and device of business processing |
| CN112818377A (en) * | 2019-11-18 | 2021-05-18 | 广东美云智数科技有限公司 | Authority data recommendation method, authority setting method, authority data recommendation system, authority setting system, electronic device and medium |
| CN113674010A (en) * | 2020-05-15 | 2021-11-19 | 北京沃东天骏信息技术有限公司 | Information processing method, device and storage medium |
| CN111639268B (en) * | 2020-06-01 | 2023-02-17 | 上海大学 | User similarity calculation method |
| CN111639268A (en) * | 2020-06-01 | 2020-09-08 | 上海大学 | User similarity calculation method |
| CN113486249A (en) * | 2021-07-27 | 2021-10-08 | 中国银行股份有限公司 | Mobile banking business recommendation method and device |
| CN113486249B (en) * | 2021-07-27 | 2024-02-27 | 中国银行股份有限公司 | Mobile banking business recommendation method and device |
| CN113888291A (en) * | 2021-12-08 | 2022-01-04 | 北京达佳互联信息技术有限公司 | Virtual space processing method and device, electronic equipment and storage medium |
| CN113888291B (en) * | 2021-12-08 | 2022-04-22 | 北京达佳互联信息技术有限公司 | Virtual space processing method and device, electronic equipment and storage medium |
| CN115408601A (en) * | 2022-05-18 | 2022-11-29 | 安胜旗信息咨询(深圳)有限公司 | Labeling method, device, medium and equipment based on user operation behaviors |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108205775A (en) | The recommendation method, apparatus and client of a kind of business object | |
| US20210142389A1 (en) | Method and System for making Recommendation from Binary Data Using Neighbor-Score Matrix and Latent Factors | |
| CN109902708B (en) | Recommendation model training method and related device | |
| CN109191240B (en) | Method and device for recommending commodities | |
| TWI716057B (en) | Service recommendation method, device and equipment | |
| CN102789462B (en) | A kind of item recommendation method and system | |
| CN105469263A (en) | Commodity recommendation method and device | |
| EP2960849A1 (en) | Method and system for recommending an item to a user | |
| CN106251174A (en) | Information recommendation method and device | |
| CN110162693A (en) | A kind of method and server of information recommendation | |
| CN110008397B (en) | Recommendation model training method and device | |
| WO2019072128A1 (en) | Object identification method and system therefor | |
| CN109711887A (en) | Generation method and device of mall recommendation list, electronic equipment and computer medium | |
| CN111400613A (en) | Article recommendation method, device, medium and computer equipment | |
| KR102049777B1 (en) | Item recommendation method and apparatus based on user behavior | |
| CN110753920A (en) | System and method for optimizing and simulating web page ranking and traffic | |
| US20210192549A1 (en) | Generating analytics tools using a personalized market share | |
| GB2541275A (en) | Visitor identification via modeling with device and user behavior information | |
| CN107767152A (en) | Product purchase intention analysis method and server | |
| Lacic et al. | Towards a scalable social recommender engine for online marketplaces: The case of apache solr | |
| CN111680213B (en) | Information recommendation method, data processing method and device | |
| WO2020221022A1 (en) | Service object recommendation method | |
| CN117076770A (en) | Data recommendation method and device based on graph calculation, storage value and electronic equipment | |
| CN116977019A (en) | Merchant recommendation method and device, electronic equipment and storage medium | |
| JP2014115685A (en) | Profile analyzing device, method and program |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180626 |