CN107679945A - Method for establishing consumer objects recommended models and associated method and device - Google Patents
Method for establishing consumer objects recommended models and associated method and device Download PDFInfo
- Publication number
- CN107679945A CN107679945A CN201710891771.8A CN201710891771A CN107679945A CN 107679945 A CN107679945 A CN 107679945A CN 201710891771 A CN201710891771 A CN 201710891771A CN 107679945 A CN107679945 A CN 107679945A
- Authority
- CN
- China
- Prior art keywords
- consumer objects
- user
- feature
- consumer
- sample
- 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
- 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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Marketing (AREA)
- Accounting & Taxation (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Finance (AREA)
- Theoretical Computer Science (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Educational Administration (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (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 invention provides a kind of method for establishing consumer objects recommended models and associated method and device, is related to modeling field and information recommendation field.Wherein, the method for establishing consumer objects recommended models includes：Obtain consumption sample；The characteristic parameter of the consumption sample is determined according to user characteristics corresponding with the consumption sample and consumer objects feature；Based on model training is carried out comprising the training data including the characteristic parameter, consumer objects recommended models are obtained.Technical scheme provided in an embodiment of the present invention can take into account the individual demand of user, improve Consumer's Experience.
Description
Technical field
The present embodiments relate to modeling field and information recommendation field, more particularly to one kind to establish consumer objects recommendation
Method of model and associated method and device.
Background technology
In existing information recommendation field, by taking take-away field as an example, positive negative sample is extracted based on user's history user behaviors log,
Then rule is clicked on or bought using linear either nonlinear model Algorithm Learning user, and using the model learnt to reality
When scene carry out commodity (trade company or vegetable) sequence.
The recommendation results that this method obtains are difficult to the individual demand for taking into account user, and recommendation effect is bad.
The content of the invention
The embodiment of the present invention provides a kind of method for establishing consumer objects recommended models and associated method and device, to
Solve the problems, such as that the recommendation results of consumer objects in the prior art are not good enough.
In a first aspect, a kind of method for establishing consumer objects recommended models, methods described are provided in the embodiment of the present invention
Including：
Obtain consumption sample；
The feature of the consumption sample is determined according to user characteristics corresponding with the consumption sample and consumer objects feature
Parameter；
Based on model training is carried out comprising the training data including the characteristic parameter, consumer objects recommended models are obtained.
With reference in a first aspect, of the invention in the first implementation of first aspect, the basis and the consumption sample
User characteristics and consumer objects feature corresponding to this determine the characteristic parameter of the consumption sample, including：
The first consumer objects in the first user characteristics, the consumer objects feature in the user characteristics are special
Sign, is calculated the first matching degree parameter in the characteristic parameter.
Alternatively, first user characteristics is the price preference of user, and first consumer objects are characterized as consumption pair
The price of elephant.
With reference in a first aspect, of the invention in second of implementation of first aspect, the basis and the consumption sample
User characteristics and consumer objects feature corresponding to this determine the characteristic parameter of the consumption sample, including：
Second user in user characteristics described in the second consumer objects characteristic query in the consumer objects feature
Feature, determine the second matching degree parameter in the characteristic parameter.
Alternatively, the second user feature includes the category and the matching degree of user and category of user preferences, described
Second consumer objects are characterized as the affiliated category of the consumer objects.
With reference to the third implementation of first aspect, methods described also includes:It is true according to the historical behavior data of user
Surely the user characteristics of user preferences is reflected.
With reference to the 4th kind of implementation of first aspect, the consumption sample includes sample label；The training data bag
Include the sample label.
Second aspect, the embodiments of the invention provide a kind of device for establishing consumer objects recommended models, including：
Sample acquisition module, sample is consumed for obtaining；
Parameter determination module, for determining institute according to user characteristics corresponding with the consumption sample and consumer objects feature
State the characteristic parameter of consumption sample；
Model training module, for based on model training is carried out comprising the training data including the characteristic parameter, obtaining
Consumer objects recommended models.
With reference to second aspect, in the first implementation of second aspect, the parameter determination module includes the present invention：
First matching degree submodule, in the first user characteristics in the user characteristics, the consumer objects feature
One consumer objects feature, the first matching degree parameter in the characteristic parameter is calculated.First user characteristics is user
Price preference, first consumer objects are characterized as the price of consumer objects.
With reference to second aspect, in second of implementation of second aspect, the parameter determination module includes the present invention：
Second matching degree submodule, for user characteristics described in the second consumer objects characteristic query in the consumer objects feature
In second user feature, determine the second matching degree parameter in the characteristic parameter.The second user feature includes user
The category of hobby and user and the matching degree of category, second consumer objects are characterized as the affiliated category of the consumer objects.
With reference to the third implementation of second aspect, described device also includes：Feature mining module, for according to user
Historical behavior data determine reflection user preferences the user characteristics.
With reference to the 4th kind of implementation of second aspect, the consumption sample includes sample label；The training data bag
Include the sample label.
The function of described device can be realized by hardware, and corresponding software can also be performed by hardware and is realized.It is described
Hardware or software include one or more modules corresponding with above-mentioned function phase.
The third aspect, the embodiments of the invention provide a kind of method for recommending consumer objects, including：
Obtain the user characteristics of user and the consumer objects feature of consumer objects to be recommended；
Based on the user characteristics, the consumer objects feature and using such as aforementioned first aspect or its any realization side
The consumer objects recommended models that method described in formula is established, real recommendation consumer objects are chosen from the consumer objects to be recommended；
Recommend the real recommendation consumer objects.For example, the reality is recommended to push away to user using existing information recommendation mode
Recommend consumer objects.
It is described based on the user characteristics, the consumer objects feature in a kind of implementation for combining the third aspect
And the consumer objects recommended models using the method foundation as described in aforementioned first aspect or its any implementation, from described
Real recommendation consumer objects are chosen in consumer objects to be recommended, including：
User is with respectively treating according to the consumer objects feature calculation of the user characteristics and the consumer objects to be recommended
Recommend the characteristic parameter of consumer objects；
The characteristic parameter of the user and each consumer objects to be recommended is inputted into the consumer objects recommended models；
Based on the ranking results to the consumer objects to be recommended of consumer objects recommended models output, treated from described
Recommend to choose the real recommendation consumer objects in consumer objects.
Fourth aspect, the embodiment of the present invention provide a kind of electronic equipment, including processor and memory.Wherein, it is described to deposit
Reservoir is used to store one or more computer instruction, and one or more computer instruction is called for the processor and held
OK；The processor is used to perform one or more computer instruction to realize such as above-mentioned first aspect or its various realization
The method that mode provides, or, realize the method such as the above-mentioned third aspect or its various implementation.
The electronic equipment can also include communication interface, lead to for the electronic equipment with other equipment or communication network
Letter.
5th aspect, it is real for storing the electronic equipment the embodiments of the invention provide a kind of computer-readable storage medium
Computer software instructions used in its existing function.Such as above-mentioned first aspect is realized when the computer instruction is performed or its is various
The method that implementation provides, or, realize the method such as the above-mentioned third aspect or its various implementation.
The embodiment of the present invention can take into account the individual demand of user, improve Consumer's Experience.
The aspects of the invention or other aspects can more straightforwards in the following description.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
There is the required accompanying drawing used in technology description to be briefly described, it should be apparent that, drawings in the following description are this hairs
Some bright embodiments, for those of ordinary skill in the art, on the premise of not paying creative work, can be with root
Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 shows the flow signal of the method according to an embodiment of the invention for establishing consumer objects recommended models
Figure；
Fig. 2 shows the schematic flow sheet of the method for recommendation consumer objects according to an embodiment of the invention；
Fig. 3 shows the flow signal of the method in accordance with another embodiment of the present invention for establishing consumer objects recommended models
Figure；
Fig. 4 shows the block diagram of the device according to an embodiment of the invention for establishing consumer objects recommended models；
Fig. 5 shows one of the block diagram of the parameter determination module in Fig. 4 shown devices；
Fig. 6 shows the block diagram of electronic equipment according to an embodiment of the invention.
Embodiment
In order that those skilled in the art more fully understand the present invention program, below in conjunction with the embodiment of the present invention
Accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described.
In some flows of description in description and claims of this specification and above-mentioned accompanying drawing, contain according to
Particular order occur multiple operations, but it should be clearly understood that these operation can not occur herein according to it is suitable
Sequence is performed or performed parallel, the sequence number such as 101,102 etc. of operation, is only used for distinguishing each different operation, sequence number
Any execution sequence is not represented for itself.In addition, these flows can include more or less operations, and these operations can
To perform or perform parallel in order.It should be noted that the description such as " first " herein, " second ", is to be used to distinguish not
Message together, equipment, module etc., do not represent sequencing, it is different types also not limit " first " and " second ".
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation describes, it is clear that described embodiment is only part of the embodiment of the present invention, rather than whole embodiments.Ability
The every other embodiment that field technique personnel are obtained based on the embodiment in the present invention, belongs to the scope of protection of the invention.
First, to the present invention relates to or the title that may relate to explain.These explanations are intended to make that the present invention's is various
Embodiment is more easily understood, rather than is limited the scope of the invention.
Information retrieval technique, refer to the demand of the input comprehension of information user by acquisition, examined from magnanimity information
Rope and the technology for returning to relevant information.
Recommend (or commending system), refer to the hobby and behavioural habits by excavating user, according to Result to
Recommend content interested or article in family.
CTR (Click-Through-Rate), click-through-rate.By taking the web advertisement as an example, i.e. the web advertisement clicks
Up to rate, CTR=(amount of showing of actual click number ÷ advertisements) × 100%.
CVR (Conversion Rate), conversion ratio.It is the index of a measurement advertising results by taking the web advertisement as an example,
CVR=(inversion quantity ÷ click volumes) × 100%.
Order models, refer to for a large amount of positive negative samples, trained using certain model algorithm come with to unknown sample
Originally it is ranked up the model of ability.Sell outside under scene, refer to the business recalled to a point of interest (point of interest)
Family or vegetable carry out the model of marking sequence.
Fig. 1 is the schematic flow sheet of the method according to an embodiment of the invention for establishing consumer objects recommended models.Ginseng
According to Fig. 1, methods described includes：
10：Obtain consumption sample.Wherein, " acquisition " include pass through mode obtained from data processing, also include reception and
Obtained mode.
Alternatively, in a kind of implementation of the present embodiment, processing 10 is realized in the following ways：Based on user's history
Behavioral data (for example, consumer behavior in a period of time) extracts the consumption sample.
Wherein, the consumption sample can be understood as consumer record, including user id and consumer objects id.It is in addition, described
Label information (positive/negative), real time information (comprising dispatching duration, dispatching distance etc.) etc. can also be included by consuming sample.Wherein, institute
Label information is stated, for example, based on CTR (clicking rate), it is believed that the sample that behavior is clicked on comprising user is positive sample, not
What it is by user's click is then negative sample；Based on CVR (conversion ratio), it is believed that conversion (for example, be converted into real consumption or
Actually place an order) sample be positive sample, unconverted sample is then negative sample.
12：The spy of the consumption sample is determined according to user characteristics corresponding with the consumption sample and consumer objects feature
Levy parameter.
Alternatively, in a kind of implementation of the present embodiment, user characteristics and consumer objects feature are obtained in advance.Its
In, the user characteristics is corresponding with user id, for reflecting consumer characteristics (for example, hobby, shopping preferences etc.)；It is described to disappear
Take that characteristics of objects is corresponding with consumer objects id, for reflecting the feature of consumer objects, the consumer objects for example can be trade company,
Commodity, vegetable etc..
14：Based on model training is carried out comprising the training data including the characteristic parameter, obtain consumer objects and recommend mould
Type.
The method provided using the present embodiment, the feature ginseng of consumption sample is determined based on user characteristics and consumer objects feature
Number, and based on model training is carried out comprising the training data including the characteristic parameter, the consumer objects so obtained recommend mould
Type can synthetic user feature and consumer objects feature carry out consumer objects recommendation, so as to consider the same of the feature of consumer objects
When, take into account user individual (for example, consumption habit, taste custom, brand custom etc.).
In a kind of implementation of the present embodiment, using the user characteristics as the user preferences for reflecting user preferences
Exemplified by feature, the consumer objects are commodity, the consumption that can be had the following advantages that of method that is provided using the present embodiment
Object recommendation model：Recommendation results (or ranking results) more conform to the individual demand of user；It can improve in user's history
The exposure rate for the commodity (that is, long-tail commodity) do not consumed.
Alternatively, in a kind of implementation of the present embodiment, at least one of the user characteristics includes user characteristics, institute
State consumer objects feature and include at least one consumer objects feature, the characteristic parameter includes at least one parameter.Alternatively, exist
In processing 12, a user characteristics is corresponding with a consumer objects feature, forms one group, the matching degree of the two is then used as the spy
Levy a parameter in parameter.
Alternatively, in a kind of implementation of the present embodiment, in processing 12, according to first in the user characteristics
The first consumer objects feature in user characteristics, the consumer objects feature, is calculated first in the characteristic parameter
With degree parameter.The characteristic parameter can only include the first matching degree parameter, can also include other specification.
Exemplarily, first user characteristics is the price preference of user, and first consumer objects are characterized as consuming
The price (for example, average visitor's unit price) of object.
Alternatively, in a kind of implementation of the present embodiment, in processing 12, according in the consumer objects feature
Second user feature in user characteristics described in second consumer objects characteristic query, determine the second matching in the characteristic parameter
Spend parameter.The characteristic parameter can only include the second matching degree parameter, can also include other specification.
Exemplarily, the second user feature includes the category and the matching degree of user and category of user preferences, institute
State the second consumer objects and be characterized as the affiliated category of the consumer objects.
Alternatively, in a kind of implementation of the present embodiment, the consumption sample includes sample label, the training number
According to including the sample label.Wherein, the sample label is used for the classification for representing the consumption sample, can be used as implementation model
Supervised learning processing in training.
Alternatively, in a kind of implementation of the present embodiment, the training data in addition to comprising the characteristic parameter,
Also include the parameter of sales volume (for example, total sales volume, average sales volume etc.), favorable comment number etc, the parameter is used to quantify user couple
In the possibility of the purchase of consumer objects, it can be used as the recurrence processing in implementation model training.
Alternatively, in a kind of implementation of the present embodiment, the effect of the consumer objects recommended models is, according to
The characteristic parameter of family and both consumer objects is given a mark (or, evaluation) to consumer objects.Fraction is higher, represents consumer objects quilt
User's receiving, the possibility liked or placed an order are bigger.
The method for establishing consumer objects recommended models according to an embodiment of the present invention is said above in association with Fig. 1
It is bright.It will be appreciated by those skilled in the art that the present invention also protects recommends consumer objects based on the consumer objects recommended models
Method.Exemplarily, Fig. 2 shows the schematic flow sheet of the method for recommendation consumer objects according to an embodiment of the invention,
Reference picture 2, methods described includes：
20：Obtain the user characteristics of user and the consumer objects feature of consumer objects to be recommended.
Wherein, the consumer objects to be recommended refer to the set of predetermined consumer objects, for therefrom choosing real push away
Recommend consumer objects (i.e. the consumer objects of actual recommendation).For example, the consumer objects to be recommended can be called together according to customer location
The consumer objects of the locality condition for the satisfaction setting returned.
22：Based on the user characteristics, the consumer objects feature and consumer objects recommended models, from described to be recommended
Real recommendation consumer objects are chosen in consumer objects.Wherein, the consumer objects recommended models are using embodiment illustrated in fig. 1 or reality
The consumer objects recommended models that the method that existing mode is provided is established.
24：Recommend the real recommendation consumer objects.For example, the real recommendation is consumed using existing information recommendation technology
Object recommendation is to user.
Alternatively, in a kind of implementation of the present embodiment, processing 22 is realized in the following ways：According to the user
User described in feature and the consumer objects feature calculation of the consumer objects to be recommended and the feature of each consumer objects to be recommended are joined
Number；The characteristic parameter of the user and each consumer objects to be recommended is inputted into the consumer objects recommended models；Disappeared based on described
Take the ranking results to the consumer objects to be recommended of object recommendation model output, chosen from the consumer objects to be recommended
The real recommendation consumer objects.Certainly, in another implementation of the present embodiment, consumer objects recommended models can be directly defeated
Go out the real recommendation consumer objects；In other implementations of the present embodiment, consumer objects recommended models can also integrate meter
Calculate the function of the characteristic parameter.
The method of the recommendation consumer objects provided using the present embodiment, on the one hand, be advantageous to obtain and meet user individual
The recommendation results of demand；On the other hand, be advantageous to improve the exposure rate for the consumer objects for not having consumption in user's history.This area
It will be appreciated by the skilled person that the method that the present embodiment provides, can apply to recommend offline, can also be pushed away applied to real-time online
Recommend.
Fig. 3 is the schematic flow sheet of the method according to an embodiment of the invention for establishing consumer objects recommended models.Ginseng
According to Fig. 3, so that consumer objects are commodity as an example, methods described includes：
30：Excavate user preference information (that is, user characteristics).Specifically, according to user's history behavior, including row is browsed
For, click on behavior, buying behavior, search behavior etc., excavate behavioural habits of the user in nearest certain period of time and under
Single preference, including customer consumption price preference (commodity it is preferential before and it is preferential after), completely subtract preferential susceptibility, Brand preference,
Merchandise classification preference etc..
For example, user A is in nearest three months vegetables for buying 10 times " KFC ", spending amount between 30 to 50,
And it is not used any preferential.Then it is concluded that：The price preference of the user is 40 ± 10, and Brang Preference is " KFC ",
Merchandise classification preference is western fast food, and preferential insensitive to completely subtracting.
32：Excavate commodity data information (that is, product features).Specifically, according to commodity (trade company or vegetable) most
Sales situation closely in certain time, visitor's unit price (before preferential and preferential rear) of trade company is counted, completely subtracts dynamics situation, manage class
Not etc..
For example, trade company S is respectively 20 (after preferential) and 30 (before preferential) in average visitor's unit price trimestral recently, completely subtract power
Spend and subtract 3 for full 30, then it is 10% completely to subtract dynamics, and operation classification is western fast food, hamburger.
34：Calculate matching degree.Specifically, for each consumption sample, according to the trade company id and use included in sample
Family id, user profile corresponding to matching and merchant information (are excavated according to step1 and step2 and obtained), and calculate corresponding informance it
Between matching degree.It is probably multidimensional in view of user profile and merchant information, so, matching degree can also be multidimensional.
Alternatively, the computational methods of matching degree have adjusted with changing features.For example it is calculated as with prices match degree
Example, it is assumed that user A and trade company S id information is included in certain sample, then extraction user A consumption price preference is x, trade company S
Objective unit price be y, then the calculation of user and trade company prices match degree m are：
M=1, if x=0 or y>2.0x；
M=0, if y>=0.8x and y<=1.2x；
M=(| x-y | ÷ x -0.2) ÷ (0.8), if y<0.8x or (y>1.2x and y<=2.0x)；
The scope for wherein matching angle value represents matching completely for 0~1,0, and 1 represents to mismatch, without the situation of user data
Lower setting matching degree default value is 1.
Assuming that user A consumption price preference is 40, trade company S objective unit price is 60, then matching degree is 0.375.
For another example, by taking the calculating of commodity category hobby matching degree as an example, it is assumed that obtained according to user's history Behavior mining
The category hobby dictionary (that is, a kind of user characteristics) of user is tags={ tag1:W1, tag2:W2 ..., tagi:wi,…,
tagN:WN }, wherein wi is the weight that the user is liked i-th of category, and N is the category quantity that the user likes, then for master
The commodity that category is tag2 are wanted, the matching degree of the two is then w2.
36：Model training.Specifically, for each sample, matching angle value can be calculated using processing 34.
By the sequence inputting of sample label and these matching angle value compositions into order models, model training is carried out.
Alternatively, the method that model training uses includes LR (Logic Regression Models), (gradient lifts decision tree to GBDT, is
A kind of decision tree combinational algorithm) and the two combine algorithm.By the model trained can apply to online marking in real time and
Sequence.For each user-commodity pair of Real time request on each sample under line and line, above-mentioned matching degree can be calculated
Value.
Fig. 4 is the block diagram of the device according to an embodiment of the invention for establishing consumer objects recommended models, reference picture 4,
Described device includes sample acquisition module 40, parameter determination module 42 and model training module 44, is specifically described below.
In the present embodiment, sample acquisition module 40 is used to obtain consumption sample.Wherein, " acquisition " is included by data
Mode obtained from reason, also include mode obtained from receiving.
Alternatively, in a kind of implementation of the present embodiment, sample acquisition module 40 is based on user's history behavioral data
(for example, consumer behavior in a period of time) extracts the consumption sample.
Wherein, the consumption sample can be understood as consumer record, including user id and consumer objects id.It is in addition, described
Label information (positive/negative), real time information (comprising dispatching duration, dispatching distance etc.) etc. can also be included by consuming sample.Wherein, institute
Label information is stated, for example, based on CTR (clicking rate), it is believed that the sample that behavior is clicked on comprising user is positive sample, not
What it is by user's click is then negative sample；Based on CVR (conversion ratio), it is believed that conversion (for example, be converted into real consumption or
Actually place an order) sample be positive sample, unconverted sample is then negative sample.
In the present embodiment, parameter determination module 42 is used for according to user characteristics corresponding with the consumption sample and consumption
Characteristics of objects determines the characteristic parameter of the consumption sample.Alternatively, in a kind of implementation of the present embodiment, the user
Feature is corresponding with user id, for reflecting consumer characteristics (for example, hobby, shopping preferences etc.)；The consumer objects feature
Corresponding with consumer objects id, for reflecting the feature of consumer objects, the consumer objects for example can be trade company, commodity, vegetable
Deng.
In the present embodiment, model training module 44 is used for based on comprising the training data progress including the characteristic parameter
Model training, obtain consumer objects recommended models.
The described device provided using the present embodiment, the spy of consumption sample is determined based on user characteristics and consumer objects feature
Parameter is levied, and based on model training is carried out comprising the training data including the characteristic parameter, the consumer objects so obtained push away
Recommend model can synthetic user feature and consumer objects feature carry out consumer objects recommendation, so as to consider the feature of consumer objects
While, take into account user individual.
In a kind of implementation of the present embodiment, using the user characteristics as the user preferences for reflecting user preferences
Exemplified by feature, the consumer objects are commodity, the consumption that can be had the following advantages that of method that is provided using the present embodiment
Object recommendation model：Recommendation results (or ranking results) more conform to the individual demand of user；It can improve in user's history
The exposure rate for the commodity (that is, long-tail commodity) do not consumed.
Alternatively, in a kind of implementation of the present embodiment, at least one of the user characteristics includes user characteristics, institute
State consumer objects feature and include at least one consumer objects feature, the characteristic parameter includes at least one parameter.One user
Feature is corresponding with a consumer objects feature, forms one group, the matching degree of the two is then as a ginseng in the characteristic parameter
Number.
Alternatively, in a kind of implementation of the present embodiment, as shown in figure 5, parameter determination module 42 includes first
With degree submodule 420, disappear for first in the first user characteristics in the user characteristics, the consumer objects feature
Take characteristics of objects, the first matching degree parameter in the characteristic parameter is calculated.The characteristic parameter can only include described
First matching degree parameter, other specification can also be included.
Exemplarily, first user characteristics is the price preference of user, and first consumer objects are characterized as consuming
The price (for example, average visitor's unit price) of object.
Alternatively, in a kind of implementation of the present embodiment, as shown in figure 5, parameter determination module 42 includes second
With degree submodule 422, in user characteristics described in the second consumer objects characteristic query in the consumer objects feature
Second user feature, determine the second matching degree parameter in the characteristic parameter.The characteristic parameter can only include described
Second matching degree parameter, other specification can also be included.
Exemplarily, the second user feature includes the category and the matching degree of user and category of user preferences, institute
State the second consumer objects and be characterized as the affiliated category of the consumer objects.
Alternatively, in a kind of implementation of the present embodiment, as shown in dotted line frame in Fig. 4, described device also includes spy
Sign excavates module 46, for determining the user characteristics of reflection user preferences according to the historical behavior data of user.
Alternatively, in a kind of implementation of the present embodiment, the consumption sample includes sample label, the training number
According to including the sample label.Wherein, the sample label is used for the classification for representing the consumption sample, can be used as implementation model
Supervised learning processing in training.
Alternatively, in a kind of implementation of the present embodiment, the training data in addition to comprising the characteristic parameter,
Also include the parameter of sales volume (for example, total sales volume, average sales volume etc.), favorable comment number etc, the parameter is used to quantify user couple
In the possibility of the purchase of consumer objects, it can be used as the recurrence processing in implementation model training.
Fig. 6 shows the block diagram of electronic equipment according to an embodiment of the invention.Reference picture 6, the electronic equipment bag
Include processor 62 and memory 60.Wherein, the memory 60 is used to store one or more computer instruction, wherein, it is described
One or more computer instruction is called for the processor and performed.The processor 62 is by performing based on described one or more
Calculation machine is instructed to realize the side for establishing consumer objects recommended models provided such as Fig. 1 or embodiment illustrated in fig. 3 or implementation
Method, or, recommend to disappear based on the consumer objects recommended models to realize for performing one or more computer instruction
Take the method (for example, method that embodiment is provided as shown in Figure 2) of object.
In addition, as shown in dotted line frame in Fig. 6, the electronic equipment can also include input/output interface, be used for into line number
According to communication.
The embodiment of the present invention also provides a kind of computer-readable storage medium, and its function institute is realized for storing aforementioned electronic devices
Computer software instructions, specifically, building as shown in Fig. 1 or Fig. 3 is realized when the computer software instructions are performed
The method of vertical consumer objects recommended models, or, realize the method (example for recommending consumer objects based on the consumer objects recommended models
Such as, the method that embodiment as shown in Figure 2 is provided).
It is apparent to those skilled in the art that for convenience and simplicity of description, the system of foregoing description,
The specific work process of device and unit, the corresponding process in preceding method embodiment is may be referred to, will not be repeated here.
Device embodiment described above is only schematical, wherein the unit illustrated as separating component can
To be or may not be physically separate, it can be as the part that unit is shown or may not be physics list
Member, you can with positioned at a place, or can also be distributed on multiple NEs.It can be selected according to the actual needs
In some or all of module realize the purpose of this embodiment scheme.Those of ordinary skill in the art are not paying creativeness
Work in the case of, you can to understand and implement.
Through the above description of the embodiments, those skilled in the art can be understood that each embodiment can
Realized by the mode of software plus required general hardware platform, naturally it is also possible to pass through hardware.Based on such understanding, on
The part that technical scheme substantially in other words contributes to prior art is stated to embody in the form of software product, should
Computer software product can store in a computer-readable storage medium, such as ROM/RAM, magnetic disc, CD, including some fingers
Make to cause a computer equipment (can be personal computer, server, or network equipment etc.) to perform each implementation
Method described in some parts of example or embodiment.
Finally it should be noted that：The above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although
The present invention is described in detail with reference to the foregoing embodiments, it will be understood by those within the art that：It still may be used
To be modified to the technical scheme described in foregoing embodiments, or equivalent substitution is carried out to which part technical characteristic；
And these modification or replace, do not make appropriate technical solution essence depart from various embodiments of the present invention technical scheme spirit and
Scope.
The present invention discloses A1, a kind of method for establishing consumer objects recommended models, and methods described includes：
Obtain consumption sample；
The feature of the consumption sample is determined according to user characteristics corresponding with the consumption sample and consumer objects feature
Parameter；
Based on model training is carried out comprising the training data including the characteristic parameter, consumer objects recommended models are obtained.
A2, the method as described in A1, basis user characteristics corresponding with the consumption sample and consumer objects feature
The characteristic parameter of the consumption sample is determined, including：
The first consumer objects in the first user characteristics, the consumer objects feature in the user characteristics are special
Sign, is calculated the first matching degree parameter in the characteristic parameter.
A3, the method as described in A2,
First user characteristics is the price preference of user, and first consumer objects are characterized as the valency of consumer objects
Lattice.
A4, the method as any one of A1-A3, the basis and the corresponding user characteristics of the consumption sample and disappear
Expense characteristics of objects determines the characteristic parameter of the consumption sample, including：
Second user in user characteristics described in the second consumer objects characteristic query in the consumer objects feature
Feature, determine the second matching degree parameter in the characteristic parameter.
A5, the method as described in A4,
The second user feature includes the category and the matching degree of user and category of user preferences, second consumption
Characteristics of objects is the affiliated category of the consumer objects.
A6, the method as described in A1, methods described also include：
The user characteristics of reflection user preferences is determined according to the historical behavior data of user.
A7, the method as described in A1,
The consumption sample includes sample label；
The training data includes the sample label.
Invention additionally discloses B8, a kind of device for establishing consumer objects recommended models, including：
Sample acquisition module, sample is consumed for obtaining；
Parameter determination module, for determining institute according to user characteristics corresponding with the consumption sample and consumer objects feature
State the characteristic parameter of consumption sample；
Model training module, for based on model training is carried out comprising the training data including the characteristic parameter, obtaining
Consumer objects recommended models.
B9, the device as described in B8, the parameter determination module include：
First matching degree submodule, it is special for the first user characteristics in the user characteristics, the consumer objects
The first consumer objects feature in sign, the first matching degree parameter in the characteristic parameter is calculated.
B10, the device as described in B9,
First user characteristics is the price preference of user, and first consumer objects are characterized as the valency of consumer objects
Lattice.
B11, the device as any one of B8-B10, the parameter determination module include：
Second matching degree submodule, for described in the second consumer objects characteristic query in the consumer objects feature
Second user feature in user characteristics, determine the second matching degree parameter in the characteristic parameter.
B12, the device as described in B11,
The second user feature includes the category and the matching degree of user and category of user preferences, second consumption
Characteristics of objects is the affiliated category of the consumer objects.
B13, the device as described in B8, described device also include：
Feature mining module, for determining that the user of reflection user preferences is special according to the historical behavior data of user
Sign.
B14, the device as described in B8,
The consumption sample includes sample label；
The training data includes the sample label.
Invention additionally discloses C15, a kind of method for recommending consumer objects, including：
Obtain the user characteristics of user and the consumer objects feature of consumer objects to be recommended；
Established based on the user characteristics, the consumer objects feature and using the method as any one of A1-A7
Consumer objects recommended models, chosen from the consumer objects to be recommended and real recommend consumer objects；
Recommend the real recommendation consumer objects.
C16, the method as described in C15, it is described based on the user characteristics, the consumer objects feature and using such as
The consumer objects recommended models that method any one of A1-A7 is established, real recommend is chosen from the consumer objects to be recommended
Consumer objects, including：
User is with respectively treating according to the consumer objects feature calculation of the user characteristics and the consumer objects to be recommended
Recommend the characteristic parameter of consumer objects；
The characteristic parameter of the user and each consumer objects to be recommended is inputted into the consumer objects recommended models；
Based on the ranking results to the consumer objects to be recommended of consumer objects recommended models output, treated from described
Recommend to choose the real recommendation consumer objects in consumer objects.
Invention additionally discloses D17, a kind of electronic equipment, including processor and memory,
The memory is used to store one or more computer instruction, wherein, one or more computer instruction
Call and perform for the processor；
The processor is used to perform one or more computer instruction to realize as in A1-A7 or C15-C16 appointed
Method described in one.
Invention additionally discloses E18, a kind of computer-readable storage medium, the computer-readable storage medium is stored with one or more
Computer instruction, the method as any one of A1-A7 or C15-C16 is realized when the computer instruction is performed.
Claims (10)
1. a kind of method for establishing consumer objects recommended models, methods described include：
Obtain consumption sample；
The characteristic parameter of the consumption sample is determined according to user characteristics corresponding with the consumption sample and consumer objects feature；
Based on model training is carried out comprising the training data including the characteristic parameter, consumer objects recommended models are obtained.
2. the method as described in claim 1, it is characterised in that the basis and the corresponding user characteristics of the consumption sample and
Consumer objects feature determines the characteristic parameter of the consumption sample, including：
The first consumer objects feature in the first user characteristics, the consumer objects feature in the user characteristics, meter
Calculation obtains the first matching degree parameter in the characteristic parameter.
3. method as claimed in claim 2, it is characterised in that
First user characteristics is the price preference of user, and first consumer objects are characterized as the price of consumer objects.
4. such as the method any one of claim 1-3, it is characterised in that the basis is corresponding with the consumption sample
User characteristics and consumer objects feature determine the characteristic parameter of the consumption sample, including：
Second user feature in user characteristics described in the second consumer objects characteristic query in the consumer objects feature,
Determine the second matching degree parameter in the characteristic parameter.
5. method as claimed in claim 4, it is characterised in that
The second user feature includes the category and the matching degree of user and category of user preferences, second consumer objects
It is characterized as the affiliated category of the consumer objects.
A kind of 6. device for establishing consumer objects recommended models, it is characterised in that including：
Sample acquisition module, sample is consumed for obtaining；
Parameter determination module, for disappearing according to user characteristics corresponding with the consumption sample and the determination of consumer objects feature
Take the characteristic parameter of sample；
Model training module, for based on model training is carried out comprising the training data including the characteristic parameter, being consumed
Object recommendation model.
7. device as claimed in claim 6, it is characterised in that the parameter determination module includes：
First matching degree submodule, in the first user characteristics in the user characteristics, the consumer objects feature
The first consumer objects feature, the first matching degree parameter in the characteristic parameter is calculated.
8. device as claimed in claim 7, it is characterised in that
First user characteristics is the price preference of user, and first consumer objects are characterized as the price of consumer objects.
9. such as the device any one of claim 6-8, it is characterised in that the parameter determination module includes：
Second matching degree submodule, for user described in the second consumer objects characteristic query in the consumer objects feature
Second user feature in feature, determine the second matching degree parameter in the characteristic parameter.
10. device as claimed in claim 9, it is characterised in that
The second user feature includes the category and the matching degree of user and category of user preferences, second consumer objects
It is characterized as the affiliated category of the consumer objects.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710891771.8A CN107679945A (en) | 2017-09-27 | 2017-09-27 | Method for establishing consumer objects recommended models and associated method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710891771.8A CN107679945A (en) | 2017-09-27 | 2017-09-27 | Method for establishing consumer objects recommended models and associated method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107679945A true CN107679945A (en) | 2018-02-09 |
Family
ID=61137683
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710891771.8A Pending CN107679945A (en) | 2017-09-27 | 2017-09-27 | Method for establishing consumer objects recommended models and associated method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107679945A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109961357A (en) * | 2019-03-25 | 2019-07-02 | 上海拉扎斯信息科技有限公司 | User data processing method, device, electronic equipment and storage medium |
| CN110046952A (en) * | 2019-01-30 | 2019-07-23 | 阿里巴巴集团控股有限公司 | A kind of training method and device, a kind of recommended method and device of recommended models |
| WO2019174395A1 (en) * | 2018-03-13 | 2019-09-19 | 阿里巴巴集团控股有限公司 | Method and apparatus for information recommendation, and device |
| CN110503498A (en) * | 2018-05-16 | 2019-11-26 | 北京三快在线科技有限公司 | A kind of order recommended method and device |
| CN111242752A (en) * | 2020-04-24 | 2020-06-05 | 支付宝(杭州)信息技术有限公司 | Method and system for determining recommended object based on multi-task prediction |
| CN111680213A (en) * | 2019-03-11 | 2020-09-18 | 阿里巴巴集团控股有限公司 | Information recommendation method, data processing method and device |
| CN112258268A (en) * | 2020-10-16 | 2021-01-22 | 百度国际科技（深圳）有限公司 | Method, apparatus, device and medium for determining recommendation model and determining price of item |
| CN112275438A (en) * | 2020-10-13 | 2021-01-29 | 成都智叟智能科技有限公司 | Dry and wet garbage separation and crushing control method and system based on data analysis |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104156450A (en) * | 2014-08-15 | 2014-11-19 | 同济大学 | Item information recommending method based on user network data |
| CN106022869A (en) * | 2016-05-12 | 2016-10-12 | 北京邮电大学 | Consumption object recommending method and consumption object recommending device |
| CN106897911A (en) * | 2017-01-10 | 2017-06-27 | 南京邮电大学 | A kind of self adaptation personalized recommendation method based on user and article |
| CN106980648A (en) * | 2017-02-27 | 2017-07-25 | 南京邮电大学 | It is a kind of that the personalized recommendation method for combining similarity is decomposed based on probability matrix |
-
2017
- 2017-09-27 CN CN201710891771.8A patent/CN107679945A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104156450A (en) * | 2014-08-15 | 2014-11-19 | 同济大学 | Item information recommending method based on user network data |
| CN106022869A (en) * | 2016-05-12 | 2016-10-12 | 北京邮电大学 | Consumption object recommending method and consumption object recommending device |
| CN106897911A (en) * | 2017-01-10 | 2017-06-27 | 南京邮电大学 | A kind of self adaptation personalized recommendation method based on user and article |
| CN106980648A (en) * | 2017-02-27 | 2017-07-25 | 南京邮电大学 | It is a kind of that the personalized recommendation method for combining similarity is decomposed based on probability matrix |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019174395A1 (en) * | 2018-03-13 | 2019-09-19 | 阿里巴巴集团控股有限公司 | Method and apparatus for information recommendation, and device |
| CN110503498B (en) * | 2018-05-16 | 2021-07-20 | 北京三快在线科技有限公司 | Order recommendation method and device |
| CN110503498A (en) * | 2018-05-16 | 2019-11-26 | 北京三快在线科技有限公司 | A kind of order recommended method and device |
| CN110046952A (en) * | 2019-01-30 | 2019-07-23 | 阿里巴巴集团控股有限公司 | A kind of training method and device, a kind of recommended method and device of recommended models |
| CN110046952B (en) * | 2019-01-30 | 2021-12-10 | 创新先进技术有限公司 | Recommendation model training method and device, and recommendation method and device |
| CN111680213A (en) * | 2019-03-11 | 2020-09-18 | 阿里巴巴集团控股有限公司 | Information recommendation method, data processing method and device |
| CN111680213B (en) * | 2019-03-11 | 2024-04-05 | 阿里巴巴集团控股有限公司 | Information recommendation method, data processing method and device |
| CN109961357A (en) * | 2019-03-25 | 2019-07-02 | 上海拉扎斯信息科技有限公司 | User data processing method, device, electronic equipment and storage medium |
| CN111242752A (en) * | 2020-04-24 | 2020-06-05 | 支付宝(杭州)信息技术有限公司 | Method and system for determining recommended object based on multi-task prediction |
| CN111242752B (en) * | 2020-04-24 | 2020-08-14 | 支付宝(杭州)信息技术有限公司 | Method and system for determining recommended object based on multi-task prediction |
| CN112275438A (en) * | 2020-10-13 | 2021-01-29 | 成都智叟智能科技有限公司 | Dry and wet garbage separation and crushing control method and system based on data analysis |
| CN112275438B (en) * | 2020-10-13 | 2022-03-01 | 成都智叟智能科技有限公司 | Dry and wet garbage separation and crushing control method and system based on data analysis |
| CN112258268A (en) * | 2020-10-16 | 2021-01-22 | 百度国际科技（深圳）有限公司 | Method, apparatus, device and medium for determining recommendation model and determining price of item |
| CN112258268B (en) * | 2020-10-16 | 2023-11-07 | 百度国际科技（深圳）有限公司 | Method, apparatus, device and medium for determining recommendation model and determining price of article |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107679945A (en) | Method for establishing consumer objects recommended models and associated method and device | |
| Marakanon et al. | Perceived quality, perceived risk and customer trust affecting customer loyalty of environmentally friendly electronics products | |
| Das et al. | An integrated production inventory model under interactive fuzzy credit period for deteriorating item with several markets | |
| CN110222272A (en) | A kind of potential customers excavate and recommended method | |
| CN106651418A (en) | Method of recommending add-on item for special offer when spending enough by e-business | |
| KR20120127326A (en) | Product Pricing System in E-Commerce using Internet | |
| Setiawan et al. | Data mining applications for sales information system using market basket analysis on stationery company | |
| Gangurde et al. | Building prediction model using market basket analysis | |
| Fikri et al. | Market orientation and business performance: the mediating role of total quality management and service innovation among Moslem fashion macro, small and medium enterprises in Indonesia | |
| CN110807691B (en) | Cross-commodity-class commodity recommendation method and device | |
| CN111768243A (en) | Sales prediction method, prediction model construction method, device, equipment and medium | |
| Taufik et al. | Decision support system design for determining brown sugar quality with weighted product method | |
| Priyanka et al. | Analysis Role of ML and Big Data Play in Driving Digital Marketing’s Paradigm Shift | |
| Kulshrestha et al. | Study for the prediction of E-commerce business market growth using machine learning algorithm | |
| Sharma et al. | Best seller rank (bsr) to sales: An empirical look at amazon. com | |
| Nikkhahan et al. | Customer lifetime value model in an online toy store | |
| Ganhewa et al. | Sales Optimization Solution for Fashion Retail | |
| CN112801384A (en) | Commodity quality evaluation and prediction method, system, medium and equipment | |
| Iqbal | Machine learning applications in e-commerce | |
| Al-Basha | Forecasting Retail Sales Using Google Trends and Machine Learning | |
| Stephen et al. | An intelligent cloud-based customer relationship management system to determine flexible pricing for customer retention | |
| CN111192112A (en) | Multi-platform interaction method and device | |
| Marković et al. | Analysis of Interactions of Key Stakeholders on B2C e-Markets-Agent Based Modelling and Simulation Approach | |
| da Silva Alves | Predicting product sales in fashion retailing: a data analytics approach | |
| Barbosa | Contemporary trends in innovative marketing strategies |
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 | ||
| CB02 | Change of applicant information |
Address after: 100085 Beijing, Haidian District on the road to the information on the ground floor of the 1 to the 3 floor of the 2 floor, room 11, 202 Applicant after: Beijing Xingxuan Technology Co., Ltd. Address before: 100085 Beijing, Haidian District on the road to the information on the ground floor of the 1 to the 3 floor of the 2 floor, room 11, 202 Applicant before: BEIJING XIAODU INFORMATION TECHNOLOGY CO., LTD. |
|
| CB02 | Change of applicant information | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180209 |
|
| RJ01 | Rejection of invention patent application after publication |