CN109711931A - Method of Commodity Recommendation, device, equipment and storage medium based on user's portrait - Google Patents
Method of Commodity Recommendation, device, equipment and storage medium based on user's portrait Download PDFInfo
- Publication number
- CN109711931A CN109711931A CN201811540327.2A CN201811540327A CN109711931A CN 109711931 A CN109711931 A CN 109711931A CN 201811540327 A CN201811540327 A CN 201811540327A CN 109711931 A CN109711931 A CN 109711931A
- Authority
- CN
- China
- Prior art keywords
- user
- recommended
- commodity
- current
- portrait
- 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
- 238000000034 method Methods 0.000 title claims abstract description 44
- 230000003542 behavioural effect Effects 0.000 claims abstract description 43
- 239000013598 vector Substances 0.000 claims description 71
- 239000013065 commercial product Substances 0.000 claims description 22
- 238000005070 sampling Methods 0.000 claims description 17
- 230000006399 behavior Effects 0.000 claims description 7
- 239000000047 product Substances 0.000 claims description 4
- 238000012706 support-vector machine Methods 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 2
- 239000004744 fabric Substances 0.000 claims description 2
- 235000013399 edible fruits Nutrition 0.000 claims 1
- 238000007405 data analysis Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 8
- 238000004891 communication Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 238000004088 simulation Methods 0.000 description 2
- 235000005116 Stachys sieboldii Nutrition 0.000 description 1
- 244000057214 Stachys sieboldii Species 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000017105 transposition Effects 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to data analysis technique fields, disclose a kind of Method of Commodity Recommendation, device, equipment and storage medium based on user's portrait.The purchaser record further according to user itself or the commodity of collection do not recommend same or similar commodity to the present invention, but commodity to be recommended are determined according to the behavioural characteristic of of a sort other users is belonged to user to be recommended, so that the type of commodity to be recommended is more, and content is richer, avoid the homogeneous problem of recommendation, the more potential interested commodity of user can be excavated simultaneously improves user experience so that commodity be promoted to strike a bargain.
Description
Technical field
The present invention relates to data analysis technique field more particularly to a kind of Method of Commodity Recommendation based on user's portrait, dress
It sets, equipment and storage medium.
Background technique
In order to the commodity for allowing user to recognize that oneself may be needed, some shopping websites or shopping APP can according to
The purchaser record at family or the commodity of collection recommend same or similar commodity.
But the commodity recommended using the purchaser record of user or the commodity of collection, it will usually which there are serious homogeneities to ask
Topic, and type is single, can Recommendations range it is narrow, cause the commodity recommended that can seldom meet user's actual need,
Poor user experience.
Above content is only used to facilitate the understanding of the technical scheme, and is not represented and is recognized that above content is existing skill
Art.
Summary of the invention
The main purpose of the present invention is to provide a kind of Method of Commodity Recommendation based on user's portrait, device, equipment and deposit
Storage media, it is intended to solve the problems, such as that the commodity homogeneity of recommendation exists in the prior art, and type is single, can Recommendations model
It encloses narrow, causes the commodity recommended that can seldom meet user's actual need, the technical issues of poor user experience.
To achieve the above object, the present invention provides a kind of Method of Commodity Recommendation based on user's portrait, described to be based on user
The Method of Commodity Recommendation of portrait the following steps are included:
Obtain the current Figure Characteristics of user to be recommended;
The current signature vector of the user to be recommended is generated according to the current Figure Characteristics;
Classified by hyperplane formula to the user to be recommended according to the current signature vector, obtains current point
Class result;
It obtains and belongs to the behavioural characteristic of of a sort other users with the current class result, and the behavior that will acquire is special
Sign is as with reference to behavioural characteristic;
Commodity to be recommended are determined according to the reference behavioural characteristic of acquisition, and the commodity to be recommended are pushed to described wait push away
Recommend the user equipment of user.
Preferably, described that the user to be recommended is divided by hyperplane formula according to the current signature vector
Class obtains current class result, comprising:
According to the current signature vector by the corresponding hyperplane formula of non-leaf nodes each in tree to described
User to be recommended classifies, and obtains current class result.
Preferably, the tree is binary tree structure, and the non-leaf nodes of the binary tree structure respectively corresponds not
Same hyperplane formula, the leaf node of the binary tree structure correspond to different class of subscribers.
Preferably, described that the corresponding hyperplane of non-leaf nodes each in tree is passed through according to the current signature vector
Formula classifies to the user to be recommended, obtains current class result, comprising:
Using the root node in binary tree structure as present node;
Obtain current hyperplane formula corresponding with the present node;
It brings the current signature vector into the current hyperplane formula, obtains division result;
The child node of the present node is chosen from the binary tree structure according to the division result;
Using the child node of selection as new present node, and it is corresponding with the present node current to return to the acquisition
The step of hyperplane formula, until making the corresponding class of subscriber of the leaf node when present node is leaf node
For current class result.
Preferably, before the current Figure Characteristics for obtaining user to be recommended, the commodity based on user's portrait are pushed away
Recommend method further include:
Obtain several sampling feature vectors and corresponding sample classification result;
Support vector machines model is instructed according to the sampling feature vectors and corresponding sample classification result
Practice, obtains hyperplane formula.
Preferably, the current Figure Characteristics include the age of the user to be recommended, gender, educational background, income information, disappear
Charge information and assets distribution pattern;
The current signature vector that the user to be recommended is generated according to the current Figure Characteristics, comprising:
The age of the user to be recommended, gender, educational background, income information, consumption information and assets distribution pattern are carried out
Digitized processing;
By the age of the user to be recommended after digitized processing, gender, educational background, income information, consumption information and money
Produce the current signature vector that distribution pattern group is combined into the user to be recommended.
Preferably, the reference behavioural characteristic according to acquisition determines commodity to be recommended, and the commodity to be recommended are pushed away
It send to the user equipment of the user to be recommended, comprising:
Commodity in the reference behavioural characteristic of acquisition are counted, the frequency of occurrence of each commodity is obtained;
The commodity that frequency of occurrence is more than preset quantity are pushed into institute as commodity to be recommended, and by the commodity to be recommended
State the user equipment of user to be recommended.
In addition, to achieve the above object, the present invention also proposes a kind of device for recommending the commodity based on user's portrait, the dress
It sets and includes:
Feature obtains module, for obtaining the current Figure Characteristics of user to be recommended;
Vector generation module, for generated according to the current Figure Characteristics current signature of the user to be recommended to
Amount;
Vector categorization module, for according to the current signature vector by hyperplane formula to the user to be recommended into
Row classification, obtains current class result;
Behavior the reference module, it is special for obtaining the behavior for belonging to of a sort other users with the current class result
Sign, and the behavioural characteristic that will acquire is used as and refers to behavioural characteristic;
Commodity pushing module determines commodity to be recommended for the reference behavioural characteristic according to acquisition, and will be described to be recommended
Commodity push to the user equipment of the user to be recommended.
In addition, to achieve the above object, the present invention also proposes a kind of commercial product recommending equipment based on user's portrait, described to set
It is standby to include: memory, processor and be stored on the memory and what be run on the processor is drawn a portrait based on user
Commercial product recommending program, it is described based on user portrait commercial product recommending program be arranged for carrying out as described above based on user draw a portrait
Method of Commodity Recommendation the step of.
In addition, to achieve the above object, the present invention also proposes a kind of storage medium, it is stored with and is based on the storage medium
The commercial product recommending program of user's portrait, the commercial product recommending program based on user's portrait realize institute as above when being executed by processor
The step of Method of Commodity Recommendation based on user's portrait stated.
The present invention generates the current signature vector of the user to be recommended according to the current Figure Characteristics of user to be recommended, then
Classified by hyperplane formula to the user to be recommended according to the current signature vector, obtain current class as a result,
Then the behavioural characteristic for obtaining and belonging to the behavioural characteristic of of a sort other users with the current class result, and will acquire is made
Finally to determine commodity to be recommended according to the reference behavioural characteristic of acquisition, and the commodity to be recommended are pushed away with reference to behavioural characteristic
It send to the user equipment of the user to be recommended, the purchaser record further according to user itself or the commodity of collection are not identical to recommend
Or similar commodity, but quotient to be recommended is determined according to the behavioural characteristic of of a sort other users is belonged to user to be recommended
Product, so that the type of commodity to be recommended is more, and content is richer, avoids the homogeneous problem of recommendation, while can excavate user more
More potential interested commodity improve user experience so that commodity be promoted to strike a bargain.
Detailed description of the invention
Fig. 1 is that the fund style drift degree for the hardware running environment that the embodiment of the present invention is related to determines the knot of equipment
Structure schematic diagram;
Fig. 2 is the flow diagram that fund style of the present invention drift degree determines method first embodiment;
Fig. 3 is the flow diagram that fund style of the present invention drift degree determines method second embodiment;
Fig. 4 is the schematic diagram of binary tree structure in the embodiment of the present invention;
Fig. 5 is the structural block diagram of fund style of the present invention drift degree determining device first embodiment.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific embodiment
It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not intended to limit the present invention.
Referring to Fig.1, Fig. 1 is that the commodity based on user's portrait for the hardware running environment that the embodiment of the present invention is related to push away
Recommend device structure schematic diagram.
As shown in Figure 1, should may include: processor 1001, such as centre based on the commercial product recommending equipment that user draws a portrait
It manages device (Central Processing Unit, CPU), communication bus 1002, user interface 1003, network interface 1004, storage
Device 1005.Wherein, communication bus 1002 is for realizing the connection communication between these components.User interface 1003 may include showing
Display screen (Display), input unit such as keyboard (Keyboard), optional user interface 1003 can also include the wired of standard
Interface, wireless interface.Network interface 1004 optionally may include standard wireline interface and wireless interface (such as Wireless Fidelity
(WIreless-FIdelity, WI-FI) interface).Memory 1005 can be the random access memory (Random of high speed
Access Memory, RAM) memory, be also possible to stable nonvolatile memory (Non-Volatile Memory,
), such as magnetic disk storage NVM.Memory 1005 optionally can also be the storage device independently of aforementioned processor 1001.
It is pushed away it will be understood by those skilled in the art that structure shown in Fig. 1 is not constituted to based on the commodity that user draws a portrait
The restriction for recommending equipment may include perhaps combining certain components or different component cloth than illustrating more or fewer components
It sets.
As shown in Figure 1, as may include operating system, network communication mould in a kind of memory 1005 of storage medium
Block, Subscriber Interface Module SIM and the commercial product recommending program based on user's portrait.
In commercial product recommending equipment based on user's portrait shown in Fig. 1, network interface 1004 is mainly used for setting with other
It is standby to carry out data communication;User interface 1003 is mainly used for carrying out data interaction with user;The present invention is based on the quotient of user's portrait
Processor 1001, memory 1005 in product recommendation apparatus can be set in the commercial product recommending equipment drawn a portrait based on user, institute
State the commercial product recommending equipment based on user's portrait calls what is stored in memory 1005 to draw a portrait based on user by processor 1001
Commercial product recommending program, and execute it is provided in an embodiment of the present invention based on user portrait Method of Commodity Recommendation.
The embodiment of the invention provides a kind of Method of Commodity Recommendation based on user's portrait, are the present invention referring to Fig. 2, Fig. 2
A kind of flow diagram of the Method of Commodity Recommendation first embodiment based on user's portrait.
In the present embodiment, it is described based on user portrait Method of Commodity Recommendation the following steps are included:
S10: the current Figure Characteristics of user to be recommended are obtained.
It should be noted that the executing subject of the method for the present embodiment is specially the server for being able to carry out commercial product recommending.
It should be understood that the user to be recommended is the user for needing to carry out commercial product recommending, can be selected at random by server
The user taken can also be the user for accessing shopping website, the user of shopping website, the present embodiment can be also accessed for before
It is without restriction to this.
It will be appreciated that the current Figure Characteristics are that can reflect user's purchasing habits to be recommended to a certain extent
Feature, may include user base feature and user's finance feature, in the concrete realization, the user base feature can wrap
It includes: the information such as age, gender and educational background, user's finance feature can include: income information, consumption information and assets distributional class
The information such as type, to guarantee more accurately to reflect the feature of user's purchasing habits to be recommended, so, in the present embodiment, institute
State age, gender, educational background that current Figure Characteristics include the user to be recommended, income information, consumption information and assets distribution
Type.
S20: the current signature vector of the user to be recommended is generated according to the current Figure Characteristics.
Since above-mentioned current Figure Characteristics are similar to simulation feature, for convenient for generation current signature vector, the present embodiment
In, when generating current signature vector, first the age of the user to be recommended, gender, educational background, income information, consumption can be believed
Breath and assets distribution pattern carry out digitized processing, then by the age of the user to be recommended after digitized processing, gender,
It goes through, take in the current signature vector that information, consumption information and assets distribution pattern group are combined into the user to be recommended.
When carrying out digitized processing to the age, treated that the age can refer to the characteristic value in following table for numeralization:
When carrying out digitized processing to gender, treated that gender can refer to the characteristic value in following table for numeralization:
When carrying out digitized processing to educational background, treated that educational background can refer to characteristic value in following table for numeralization:
To income information carry out digitized processing when, numeralization treated income information can refer to the feature in following table
Value:
When carrying out digitized processing to consumption information, treated that consumption information can refer to the feature in following table for numeralization
Value:
When carrying out digitized processing to assets distribution pattern, treated that assets distribution pattern can refer in following table for numeralization
Characteristic value:
Assuming that numeralization treated age is a, numeralization treated gender is b, treated that educational background is for numeralization
C, numeralization treated income information is d, numeralization treated consumption information are e, numeralization treated assets distribution
Type is f, at this point, the current signature vector of the user to be recommended is (a, b, c, d, e, f).
S30: classified by hyperplane formula to the user to be recommended according to the current signature vector, worked as
Preceding classification results.
It should be noted that for convenient for obtaining hyperplane formula, in the present embodiment, before step S10, it may include: it obtains
Take several sampling feature vectors and corresponding sample classification result;According to the sampling feature vectors and corresponding sample point
Class result is trained support vector machines model, obtains hyperplane formula.
In the concrete realization, the hyperplane formula is WTX+b=0, wherein x is the vector on the hyperplane, W
It is constant vector with b, T is transposition symbol.
It correspondingly, can be by the current signature vector when being classified by hyperplane formula to the user to be recommended
It is substituted into the hyperplane formula as x, if WTUser to be recommended can be classified as the first kind, if W by x+b > 0TX+b < 0, can will be to
Recommended user is classified as the second class.
S40: the row that belongs to the behavioural characteristic of of a sort other users with the current class result, and will acquire is obtained
It is characterized as reference behavioural characteristic.
It should be noted that the behavioural characteristic is commodity purchasing record, articles storage record and/or goods browse note
The information such as record, and each user can usually have these behavioural characteristics, at this point, first obtaining and the current class result category
In the behavioural characteristic of of a sort other users, and the behavioural characteristic that will acquire is used as and refers to behavioural characteristic.
S50: determining commodity to be recommended according to the reference behavioural characteristic of acquisition, and the commodity to be recommended is pushed to described
The user equipment of user to be recommended.
It should be noted that the user equipment is equipment used by a user to be recommended, it can be smart phone, pen
Remember the equipment such as this computer, PC and tablet computer, the present embodiment is without restriction to this.
It will be appreciated that one or more commodity would generally be corresponded to for reference behavioural characteristic, at this point, can be first
Commodity in the reference behavioural characteristic of acquisition are counted, the frequency of occurrence of each commodity is obtained;It is more than again pre- by frequency of occurrence
If the commodity to be recommended are pushed to the user equipment of the user to be recommended as commodity to be recommended by the commodity of quantity.
In the concrete realization, the preset quantity can be configured as needed, such as: it may be configured as 50 or 100, this reality
It is without restriction to this to apply example.
Assuming that preset quantity is 50 times, the frequency of occurrence of certain a smart phone is 64 times, certain a laptop
Frequency of occurrence is 89 times, and the frequency of occurrence of certain a Intelligent bracelet is 10 times, at this point, can be by the smart phone and laptop
Relevant information push to the user equipment of the user to be recommended.
By foregoing description it is not difficult to find that the present embodiment is described wait push away according to the generation of the current Figure Characteristics of user to be recommended
The current signature vector for recommending user carries out the user to be recommended by hyperplane formula further according to the current signature vector
Classification obtains current class as a result, then obtaining the behavior spy for belonging to of a sort other users with the current class result
Sign, and the behavioural characteristic that will acquire is used as and refers to behavioural characteristic, finally determines quotient to be recommended according to the reference behavioural characteristic of acquisition
Product, and the commodity to be recommended are pushed to the user equipment of the user to be recommended, do not remember further according to the purchase of user itself
The commodity of record or collection recommend same or similar commodity, but belong to of a sort other users according to user to be recommended
Behavioural characteristic determine commodity to be recommended so that the type of commodity to be recommended is more, and content is richer, avoid recommendation homogeneity
Change problem, while the more potential interested commodity of user can be excavated and improve user experience so that commodity be promoted to strike a bargain.
With reference to Fig. 3, Fig. 3 is a kind of process signal of the Method of Commodity Recommendation second embodiment based on user's portrait of the present invention
Figure.
Based on above-mentioned first embodiment, in the present embodiment, step S30 includes:
S30 ': the corresponding hyperplane formula pair of non-leaf nodes each in tree is passed through according to the current signature vector
The user to be recommended classifies, and obtains current class result.
It should be noted that when being classified due to SVM algorithm, it is typically only capable to be divided into two classes, and what shopping website was related to
User volume is usually larger, if user is only divided into two classes, commodity to be recommended can be made inaccurate, to avoid the problem, this reality
Apply in example, can according to the current signature vector by the corresponding hyperplane formula of non-leaf nodes each in tree to described
User to be recommended classifies, and obtains current class as a result, therefore, it is necessary to first establish a tree, to supporting vector
When machine SVM model is trained, establish multiple hyperplane formula, and by each hyperplane formula respectively with it is each in the tree
Non-leaf nodes is corresponded to.
Specifically, when obtaining several sampling feature vectors and corresponding sample classification result, if dry sample can be obtained first
The consuming capacity index is divided into two according to first metrics-thresholds by eigen vector and corresponding consuming capacity index
Class obtains sample classification result.
Wherein, the consuming capacity index is the index for measuring the consuming capacity of sampling feature vectors, can be according to reality
Border situation is that the consuming capacity index is arranged in each sampling feature vectors, corresponding as long as guaranteeing that sampling feature vectors are more similar
Consuming capacity index is closer.
It will be appreciated that being based on the characteristics of SVM algorithm is divided into two classes, in the present embodiment, the tree can be
Binary tree structure, referring to Fig. 4, the non-leaf nodes of the binary tree structure respectively corresponds different hyperplane formula, and described two
The leaf node of fork tree construction corresponds to different class of subscribers.
Assuming that the first metrics-thresholds be 20, sampling feature vectors have 50,000, wherein 20,000 be greater than the first metrics-thresholds, 3
Ten thousand less than the first metrics-thresholds, at this point, the corresponding classification results of the sampling feature vectors that can will be greater than the first metrics-thresholds are set
It is set to the first kind, the second class is set by the corresponding classification results of sampling feature vectors less than the first metrics-thresholds, to SVM mould
After type is trained, corresponding first hyperplane formula is obtained, it correspondingly, can be corresponding to described by the first hyperplane formula
Root node (i.e. node " 1 " in Fig. 4) in binary tree structure.
Then, the sampling feature vectors that classification results are the first kind can be further divided into the according to the second metrics-thresholds
Three classes and the 4th class, after being trained to SVM model, obtaining corresponding second hyperplane formula correspondingly can be by described second
The corresponding left child node (i.e. node " 2 " in Fig. 4) to the root node of hyperplane formula, at this point, the two of the left child node
A child node (i.e. leaf node " 4 " and leaf node " 5 " in Fig. 4) can respectively correspond the third class and the 4th class;It can incite somebody to action
Classification results are that the sampling feature vectors of the second class are further divided into the 5th class and the 6th class according to third metrics-thresholds, right
After SVM model is trained, corresponding third hyperplane formula is obtained, it correspondingly, can be corresponding by the third hyperplane formula
To the right child node (i.e. node " 3 " in Fig. 4) of the root node, at this point, two child node (i.e. Fig. 4 of the left child node
In leaf node " 6 " and leaf node " 7 ") the 5th class and the 6th class can be respectively corresponded.
The above is certainly, can also to carry out further division according to similar process as needed for being divided into four classes,
This is repeated no more.
To classify convenient for quickly treating recommended user, above-mentioned steps S30 ' can further comprise: by binary tree structure
In root node as present node;Obtain current hyperplane formula corresponding with the present node;By the current signature
Vector brings the current hyperplane formula into, obtains division result;It is selected from the binary tree structure according to the division result
Take the child node of the present node;Using the child node of selection as new present node, and returns to the acquisition and work as with described
The step of front nodal point corresponding current hyperplane formula, until when the present node is leaf node, by the leaf node
Corresponding class of subscriber is as current class result.
Such as: assuming that the user to be recommended belongs to the classification of leaf node in Fig. 4 " 6 ", at this point, can first save described
Point is used as present node, takes out the first hyperplane formula of the root node as current hyperplane formula, obtains described current
Vector characteristics belong to the second class;Further according to second class result using the right child node of the root node as present node,
Using the third hyperplane formula of the right child node as current hyperplane formula, the current vector characteristics are substituted into third and are surpassed
Plane formula, at this time, it may be determined that the user to be recommended belongs to the classification of leaf node in Fig. 4 " 6 ".
In addition, the embodiment of the present invention also proposes a kind of storage medium, it is stored on the storage medium and is drawn a portrait based on user
Commercial product recommending program, it is described based on user portrait commercial product recommending program be executed by processor when realize base as described above
In user portrait Method of Commodity Recommendation the step of.
Referring to Fig. 5, Fig. 5 is the structural block diagram for the device for recommending the commodity first embodiment drawn a portrait the present invention is based on user.
As shown in figure 5, the device for recommending the commodity based on user's portrait that the embodiment of the present invention proposes includes:
Feature obtains module 5001, for obtaining the current Figure Characteristics of user to be recommended;
Vector generation module 5002, for generating the current signature of the user to be recommended according to the current Figure Characteristics
Vector;
Vector categorization module 5003, for passing through hyperplane formula to the use to be recommended according to the current signature vector
Family is classified, and current class result is obtained;
Behavior the reference module 5004, for obtaining the behavior for belonging to of a sort other users with the current class result
Feature, and the behavioural characteristic that will acquire is used as and refers to behavioural characteristic;
Commodity pushing module 5005 determines commodity to be recommended for the reference behavioural characteristic according to acquisition, and will it is described to
Recommendations push to the user equipment of the user to be recommended.
It should be noted that when being classified due to SVM algorithm, it is typically only capable to be divided into two classes, and what shopping website was related to
User volume is usually larger, if user is only divided into two classes, commodity to be recommended can be made inaccurate, to avoid the problem, this reality
It applies in example, the vector categorization module 5003, is also used to pass through each non-leaf in tree according to the current signature vector
The corresponding hyperplane formula of node classifies to the user to be recommended, obtains current class result.
It will be appreciated that being based on the characteristics of SVM algorithm is divided into two classes, in the present embodiment, the tree is two
Tree construction is pitched, the non-leaf nodes of the binary tree structure respectively corresponds different hyperplane formula, the binary tree structure
Leaf node corresponds to different class of subscribers.
To classify convenient for quickly treating recommended user, the vector categorization module 5003 is also used to binary tree knot
Root node in structure is as present node;Obtain current hyperplane formula corresponding with the present node;By the current spy
Sign vector brings the current hyperplane formula into, obtains division result;According to the division result from the binary tree structure
Choose the child node of the present node;Using the child node of selection as new present node, and return it is described acquisition with it is described
The step of present node corresponding current hyperplane formula, until when the present node is leaf node, by the leaf section
The corresponding class of subscriber of point is as current class result.
It should be noted that in the present embodiment, the commodity based on user's portrait are pushed away for convenient for obtaining hyperplane formula
Recommend device further include:
Model training module, for obtaining several sampling feature vectors and corresponding sample classification result;According to described
Sampling feature vectors and corresponding sample classification result are trained support vector machines model, obtain hyperplane formula.
It is described current for convenient for generating current signature vector since above-mentioned current Figure Characteristics are similar to simulation feature
Figure Characteristics include age, gender, educational background, income information, consumption information and the assets distribution pattern of the user to be recommended;
The vector generation module 5002, be also used to by the age of the user to be recommended, gender, educational background, income information,
Consumption information and assets distribution pattern carry out digitized processing;By the age of the user to be recommended after digitized processing, property
Not, educational background, income information, consumption information and assets distribution pattern group are combined into the current signature vector of the user to be recommended.
It will be appreciated that one or more commodity would generally be corresponded to for reference behavioural characteristic, at this point, described
Commodity pushing module 5005, the commodity being also used in the reference behavioural characteristic to acquisition count, and obtain the appearance of each commodity
Number;Using frequency of occurrence be more than preset quantity commodity as commodity to be recommended, and the commodity to be recommended are pushed to described
The user equipment of user to be recommended.
It should be noted that workflow described above is only schematical, not to protection model of the invention
Enclose composition limit, in practical applications, those skilled in the art can select according to the actual needs part therein or
It all achieves the purpose of the solution of this embodiment, herein with no restrictions.
In addition, the not technical detail of detailed description in the present embodiment, reference can be made to provided by any embodiment of the invention
Method is determined based on the commercial product recommending of user's portrait, and details are not described herein again.
It should be noted that, in this document, the terms "include", "comprise" or its any other variant are intended to non-row
His property includes, so that the process, method, article or the system that include a series of elements not only include those elements, and
And further include other elements that are not explicitly listed, or further include for this process, method, article or system institute it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including being somebody's turn to do
There is also other identical elements in the process, method of element, article or system.
The serial number of the above embodiments of the invention is only for description, does not represent the advantages or disadvantages of the embodiments.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side
Method can be realized by means of software and necessary general hardware platform, naturally it is also possible to by hardware, but in many cases
The former is more preferably embodiment.Based on this understanding, technical solution of the present invention substantially in other words does the prior art
The part contributed out can be embodied in the form of software products, which is stored in one as described above
In storage medium (such as ROM/RAM, magnetic disk, CD), including some instructions are used so that terminal device (it can be mobile phone,
Computer, server, air conditioner or network equipment etc.) execute method described in each embodiment of the present invention.
The above is only a preferred embodiment of the present invention, is not intended to limit the scope of the invention, all to utilize this hair
Equivalent structure or equivalent flow shift made by bright specification and accompanying drawing content is applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of Method of Commodity Recommendation based on user's portrait, which is characterized in that the commercial product recommending side based on user's portrait
Method the following steps are included:
Obtain the current Figure Characteristics of user to be recommended;
The current signature vector of the user to be recommended is generated according to the current Figure Characteristics;
Classified by hyperplane formula to the user to be recommended according to the current signature vector, obtains current class knot
Fruit;
The behavioural characteristic for obtaining and belonging to the behavioural characteristic of of a sort other users with the current class result, and will acquire is made
For with reference to behavioural characteristic;
Commodity to be recommended are determined according to the reference behavioural characteristic of acquisition, and the commodity to be recommended are pushed into the use to be recommended
The user equipment at family.
2. the Method of Commodity Recommendation as described in claim 1 based on user's portrait, which is characterized in that described according to described current
Feature vector classifies to the user to be recommended by hyperplane formula, obtains current class result, comprising:
According to the current signature vector by the corresponding hyperplane formula of non-leaf nodes each in tree to described wait push away
It recommends user to classify, obtains current class result.
3. the Method of Commodity Recommendation as claimed in claim 2 based on user's portrait, which is characterized in that the tree is two
Tree construction is pitched, the non-leaf nodes of the binary tree structure respectively corresponds different hyperplane formula, the binary tree structure
Leaf node corresponds to different class of subscribers.
4. the Method of Commodity Recommendation as claimed in claim 3 based on user's portrait, which is characterized in that described according to described current
Feature vector classifies to the user to be recommended by the corresponding hyperplane formula of non-leaf nodes each in tree, obtains
Obtain current classification results, comprising:
Using the root node in binary tree structure as present node;
Obtain current hyperplane formula corresponding with the present node;
It brings the current signature vector into the current hyperplane formula, obtains division result;
The child node of the present node is chosen from the binary tree structure according to the division result;
Using the child node of selection as new present node, and it is corresponding with the present node current super flat to return to the acquisition
The step of face formula, until when the present node is leaf node, using the corresponding class of subscriber of the leaf node as working as
Preceding classification results.
5. the Method of Commodity Recommendation as described in any one of claims 1 to 4 based on user's portrait, which is characterized in that described
Before the current Figure Characteristics for obtaining user to be recommended, the Method of Commodity Recommendation based on user's portrait further include:
Obtain several sampling feature vectors and corresponding sample classification result;
Support vector machines model is trained according to the sampling feature vectors and corresponding sample classification result, is obtained
Obtain hyperplane formula.
6. the Method of Commodity Recommendation as described in any one of claims 1 to 4 based on user's portrait, which is characterized in that described
Current Figure Characteristics include age, gender, educational background, income information, consumption information and the assets distributional class of the user to be recommended
Type;
The current signature vector that the user to be recommended is generated according to the current Figure Characteristics, comprising:
The age of the user to be recommended, gender, educational background, income information, consumption information and assets distribution pattern are subjected to number
Change processing;
By the age of the user to be recommended after digitized processing, gender, educational background, income information, consumption information and assets point
Cloth type combination is the current signature vector of the user to be recommended.
7. the Method of Commodity Recommendation as described in any one of claims 1 to 4 based on user's portrait, which is characterized in that described
Commodity to be recommended are determined according to the reference behavioural characteristic of acquisition, and the commodity to be recommended are pushed to the user's to be recommended
User equipment, comprising:
Commodity in the reference behavioural characteristic of acquisition are counted, the frequency of occurrence of each commodity is obtained;
Using frequency of occurrence be more than preset quantity commodity as commodity to be recommended, and by the commodity to be recommended push to it is described to
The user equipment of recommended user.
8. a kind of device for recommending the commodity based on user's portrait, which is characterized in that described device includes:
Feature obtains module, for obtaining the current Figure Characteristics of user to be recommended;
Vector generation module, for generating the current signature vector of the user to be recommended according to the current Figure Characteristics;
Vector categorization module, for being divided by hyperplane formula the user to be recommended according to the current signature vector
Class obtains current class result;
Behavior the reference module, for obtaining the behavioural characteristic for belonging to of a sort other users with the current class result, and
The behavioural characteristic that will acquire, which is used as, refers to behavioural characteristic;
Commodity pushing module determines commodity to be recommended for the reference behavioural characteristic according to acquisition, and by the commodity to be recommended
Push to the user equipment of the user to be recommended.
9. it is a kind of based on user portrait commercial product recommending equipment, which is characterized in that the equipment include: memory, processor and
The commercial product recommending program based on user's portrait that is stored on the memory and can run on the processor, it is described to be based on
The commercial product recommending program of user's portrait is arranged for carrying out the quotient based on user's portrait as described in any one of claims 1 to 7
The step of product recommended method.
10. a kind of storage medium, which is characterized in that be stored with the commercial product recommending journey based on user's portrait on the storage medium
Sequence, the commercial product recommending program based on user's portrait are realized as described in any one of claim 1 to 7 when being executed by processor
Based on user portrait Method of Commodity Recommendation the step of.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811540327.2A CN109711931A (en) | 2018-12-14 | 2018-12-14 | Method of Commodity Recommendation, device, equipment and storage medium based on user's portrait |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811540327.2A CN109711931A (en) | 2018-12-14 | 2018-12-14 | Method of Commodity Recommendation, device, equipment and storage medium based on user's portrait |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109711931A true CN109711931A (en) | 2019-05-03 |
Family
ID=66256700
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811540327.2A Pending CN109711931A (en) | 2018-12-14 | 2018-12-14 | Method of Commodity Recommendation, device, equipment and storage medium based on user's portrait |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109711931A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110322274A (en) * | 2019-05-30 | 2019-10-11 | 深圳壹账通智能科技有限公司 | Crowd portrayal generation method, device and computer equipment based on data analysis |
CN111062785A (en) * | 2019-12-18 | 2020-04-24 | 上海良鑫网络科技有限公司 | Method and system for intelligently selecting products to recommend to matched users |
CN111523929A (en) * | 2020-04-15 | 2020-08-11 | 高邮市新世纪灯具城经营管理有限公司 | Merchant information management method and system |
CN111966904A (en) * | 2020-08-18 | 2020-11-20 | 平安国际智慧城市科技股份有限公司 | Information recommendation method based on multi-user portrait model and related device |
CN113268645A (en) * | 2021-05-07 | 2021-08-17 | 北京三快在线科技有限公司 | Information recall method, model training method, device, equipment and storage medium |
CN113378842A (en) * | 2021-05-18 | 2021-09-10 | 浙江大学 | Recommendation method based on segmented image feature extraction |
CN113436001A (en) * | 2021-06-25 | 2021-09-24 | 中国工商银行股份有限公司 | Credit card pushing method, device, equipment, storage medium and program product |
CN113643099A (en) * | 2021-08-30 | 2021-11-12 | 北京沃东天骏信息技术有限公司 | Commodity data processing method, commodity data processing device, commodity data processing apparatus, storage medium, and program product |
CN118195744A (en) * | 2024-05-14 | 2024-06-14 | 深圳环金科技有限公司 | Customer analysis system for cross-border e-commerce platform |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106294783A (en) * | 2016-08-12 | 2017-01-04 | 乐视控股(北京)有限公司 | A kind of video recommendation method and device |
CN106530001A (en) * | 2016-11-03 | 2017-03-22 | 广州市万表科技股份有限公司 | Information recommending method and apparatus |
CN106548255A (en) * | 2016-11-24 | 2017-03-29 | 山东浪潮云服务信息科技有限公司 | A kind of Method of Commodity Recommendation based on mass users behavior |
CN106651533A (en) * | 2016-12-29 | 2017-05-10 | 合肥华凌股份有限公司 | User behavior-based personalized product recommendation method and apparatus |
CN107657274A (en) * | 2017-09-20 | 2018-02-02 | 浙江大学 | A kind of y-bend SVM tree unbalanced data industry Fault Classifications based on k means |
CN107908915A (en) * | 2017-12-25 | 2018-04-13 | 济南大学 | Predict modeling and analysis method, the equipment and storage medium of tunnel crimp |
CN108090162A (en) * | 2017-12-13 | 2018-05-29 | 北京百度网讯科技有限公司 | Information-pushing method and device based on artificial intelligence |
CN108304435A (en) * | 2017-09-08 | 2018-07-20 | 腾讯科技(深圳)有限公司 | Information recommendation method, device, computer equipment and storage medium |
-
2018
- 2018-12-14 CN CN201811540327.2A patent/CN109711931A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106294783A (en) * | 2016-08-12 | 2017-01-04 | 乐视控股(北京)有限公司 | A kind of video recommendation method and device |
CN106530001A (en) * | 2016-11-03 | 2017-03-22 | 广州市万表科技股份有限公司 | Information recommending method and apparatus |
CN106548255A (en) * | 2016-11-24 | 2017-03-29 | 山东浪潮云服务信息科技有限公司 | A kind of Method of Commodity Recommendation based on mass users behavior |
CN106651533A (en) * | 2016-12-29 | 2017-05-10 | 合肥华凌股份有限公司 | User behavior-based personalized product recommendation method and apparatus |
CN108304435A (en) * | 2017-09-08 | 2018-07-20 | 腾讯科技(深圳)有限公司 | Information recommendation method, device, computer equipment and storage medium |
CN107657274A (en) * | 2017-09-20 | 2018-02-02 | 浙江大学 | A kind of y-bend SVM tree unbalanced data industry Fault Classifications based on k means |
CN108090162A (en) * | 2017-12-13 | 2018-05-29 | 北京百度网讯科技有限公司 | Information-pushing method and device based on artificial intelligence |
CN107908915A (en) * | 2017-12-25 | 2018-04-13 | 济南大学 | Predict modeling and analysis method, the equipment and storage medium of tunnel crimp |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110322274A (en) * | 2019-05-30 | 2019-10-11 | 深圳壹账通智能科技有限公司 | Crowd portrayal generation method, device and computer equipment based on data analysis |
CN111062785A (en) * | 2019-12-18 | 2020-04-24 | 上海良鑫网络科技有限公司 | Method and system for intelligently selecting products to recommend to matched users |
CN111523929A (en) * | 2020-04-15 | 2020-08-11 | 高邮市新世纪灯具城经营管理有限公司 | Merchant information management method and system |
CN111966904A (en) * | 2020-08-18 | 2020-11-20 | 平安国际智慧城市科技股份有限公司 | Information recommendation method based on multi-user portrait model and related device |
CN111966904B (en) * | 2020-08-18 | 2023-09-05 | 深圳平安智慧医健科技有限公司 | Information recommendation method and related device based on multi-user portrait model |
CN113268645A (en) * | 2021-05-07 | 2021-08-17 | 北京三快在线科技有限公司 | Information recall method, model training method, device, equipment and storage medium |
CN113378842A (en) * | 2021-05-18 | 2021-09-10 | 浙江大学 | Recommendation method based on segmented image feature extraction |
CN113436001A (en) * | 2021-06-25 | 2021-09-24 | 中国工商银行股份有限公司 | Credit card pushing method, device, equipment, storage medium and program product |
CN113643099A (en) * | 2021-08-30 | 2021-11-12 | 北京沃东天骏信息技术有限公司 | Commodity data processing method, commodity data processing device, commodity data processing apparatus, storage medium, and program product |
CN118195744A (en) * | 2024-05-14 | 2024-06-14 | 深圳环金科技有限公司 | Customer analysis system for cross-border e-commerce platform |
CN118195744B (en) * | 2024-05-14 | 2024-07-05 | 深圳环金科技有限公司 | Customer analysis system for cross-border e-commerce platform |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109711931A (en) | Method of Commodity Recommendation, device, equipment and storage medium based on user's portrait | |
Zhang et al. | Label propagation algorithm for community detection based on node importance and label influence | |
CN109697629B (en) | Product data pushing method and device, storage medium and computer equipment | |
CN105989074B (en) | A kind of method and apparatus recommend by mobile device information cold start-up | |
CN106651542A (en) | Goods recommendation method and apparatus | |
CN108829808A (en) | A kind of page personalized ordering method, apparatus and electronic equipment | |
CN108256537A (en) | A kind of user gender prediction method and system | |
CN109492180A (en) | Resource recommendation method, device, computer equipment and computer readable storage medium | |
Lu et al. | Show me the money: Dynamic recommendations for revenue maximization | |
CN106484766B (en) | Searching method and device based on artificial intelligence | |
CN108205768A (en) | Database building method and data recommendation method and device, equipment and storage medium | |
CN107273476A (en) | A kind of article search method, device and server | |
CN107424007A (en) | A kind of method and apparatus for building electronic ticket susceptibility identification model | |
CN105354203A (en) | Information display method and apparatus | |
CN109087162A (en) | Data processing method, system, medium and calculating equipment | |
CN106503224A (en) | A kind of method and device for recommending application according to keyword | |
CN106874314A (en) | The method and apparatus of information recommendation | |
CN108550046A (en) | A kind of resource and market recommendation method, apparatus and electronic equipment | |
CN110322281A (en) | The method for digging and device of similar users | |
CN107045700A (en) | Product method for pushing and device based on streaming user behavioural analysis | |
CN110263255A (en) | Acquisition methods, system, server and the storage medium of customer attribute information | |
CN106251178A (en) | Data digging method and device | |
CN106779791A (en) | A kind of generation method and device of object picture combination of arranging in pairs or groups | |
CN107391540A (en) | A kind of small routine methods of exhibiting, device and grader | |
CN109190027A (en) | Multi-source recommended method, terminal, server, computer equipment, readable medium |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190503 |