CN107730343A - A kind of user's merchandise news method for pushing and equipment based on picture attribute extraction - Google Patents
A kind of user's merchandise news method for pushing and equipment based on picture attribute extraction Download PDFInfo
- Publication number
- CN107730343A CN107730343A CN201710833188.1A CN201710833188A CN107730343A CN 107730343 A CN107730343 A CN 107730343A CN 201710833188 A CN201710833188 A CN 201710833188A CN 107730343 A CN107730343 A CN 107730343A
- Authority
- CN
- China
- Prior art keywords
- user
- commodity
- picture
- item property
- merchandise news
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/5838—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Library & Information Science (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of user's merchandise news method for pushing based on picture attribute extraction, it is characterised in that methods described includes：Obtain the instruction that triggering is recommended；Item property is extracted according to above-mentioned instruction, the end article for including above-mentioned item property is searched for from user-interest library；Form Recommendations list and push merchandise news to user.The invention also discloses the equipment using the above method.The inventive method is simple and easy to do, compact equipment, using the picture attribute combination knowledge mapping of commodity, the user individual database of comparatively perfect is established, using the collaborative filtering based on user and commodity, the accurate push of similar commodity is realized, novel and comfortable purchase experiences are brought for user.
Description
Technical field
The present invention relates to field of computer technology, more particularly to a kind of user's merchandise news based on picture attribute extraction pushes away
Delivery method and equipment.
Background technology
Electric business enters booming high speed stage in recent years, some large-scale B2C it is numerous and confused pour into capital market while,
Some are absorbed in shopping guide's platform towards B2C also in numerous and confused projection.The preliminary stage of electric business shopping guide's platform development is mainly merely
User provides flow guiding and commercial product recommending, and the market being absorbed in also multidigit female group, therefore can not expire in all directions
Sufficient user's is directed to shopping need.The later stage more vertical and precisely effort in positioning always, to be pushed away specifically for user
Recommend the commodity of tight demand so that the commodity recommended comply fully with user's request, thus more accurate and Consumer's Experience.
Moreover, existing electric business shopping guide finds target business typically by the way of user search queries or classified inquiry
Product, but the not easy-to-use simple word such as pattern, the material of suitcase of the pattern such as clothes of commodity is described, it is user
Accurately retrieve commodity interested and improve difficulty；Moreover, only consider the character description information of commodity, it is easy to ignore commodity
Other characteristics, the simple word description based on commodity can not accurately pick out suitable commodity；In addition, in general commodity push away
Send on the basis of being all built upon user's purchase and browsing record, according to the purchase interest of these discovery users, push content
It is relatively single, it can do nothing to help active user and find and meet other commodity of oneself.
And in recent years, the computer technology such as computer identification and article characteristics extraction is risen, while artificial intelligence, nerve net
The popularization and application of network provide new R＆D direction for the guidance that user buys.And knowledge mapping is in the foundation of electric business platform,
The information source of personalization is provided for vertical precise positioning customer demand.Knowledge mapping corresponding to image is established, me can be helped
The implication of keyword is described, keyword is expanded and come out for image appearance；Picture search simultaneously, help to find most like
Image.So need a kind of shopping guide based on image badly, help to find the purchase interest of user and the knowledge of deeper, for
Convenient, efficient, accurately commercial articles searching and recommendation that family is brought, so as to enjoy the purchase experiences of novelty.
The content of the invention
In order to solve the problems, such as that user's commercial product recommending is not vertical accurate enough in the prior art, the embodiments of the invention provide one
User merchandise news method for pushing and equipment of the kind based on picture attribute extraction.The technical scheme is as follows：
On the one hand, the invention provides a kind of user's merchandise news method for pushing based on picture attribute extraction, the side
Method includes：
Obtain the instruction that triggering is recommended；
Item property is extracted according to above-mentioned instruction, the target business for including above-mentioned item property is searched for from user-interest library
Product；
Form Recommendations list and push merchandise news to user.
Wherein, the instruction for triggering recommendation is touched including the passive triggering command based on system or businessman and the active based on user
Send instructions；
The passive triggering command includes weather, season, birthday, red-letter day or system self-defined technical dates；
The active triggering command includes user's active typing and presets scene information.
Further, described the step of extracting item property according to active triggering command, includes：
User inputs default scene information in electric business platform, and the default scene information is picture, text or picture and text
The combination of this information；
Picture feature is extracted using the neutral net of training in advance, it is special to obtain the semanteme corresponding with above-mentioned picture feature
Sign, end article attribute is defined as by semantic feature；
The active triggering command extraction item property also includes user and inputted in electric business platform in default scene information
The item property that text message is extracted.
Further, the creation method of the user-interest library includes：
According to user electric business platform picture searching, browse, buy and collection record, obtain in electric business platform information storehouse
The merchandise news of user's commodity interested, is made user/commodity basic information table, and the merchandise news comprises at least commodity figure
Piece；
Using training in advance neutral net extraction user/commodity basic information table in commodity picture feature, obtain with
The corresponding semantic feature of above-mentioned picture feature, item property is defined as by semantic feature；
Create using incidence relation of the item property as ABC unit and using between merchandise classification, item property type as
The user-interest library of framework.
Preferably, the concrete mode for the end article that the search from user-interest library includes item property is with basis
The item property of triggering command extraction scans for for keyword to the commodity in user-interest library；The item property includes face
Color, texture material, quality, type, style, brand, the place of production.
Wherein, the mode for forming commercial product recommending list includes user collaborative filtering and commodity collaborative filtering.
Specifically, the specific method of user collaborative filtering includes：
Item property classification is extracted using the text message of the item property of commodity in user-interest library, confirms to use
The current commodity preference at family；
Define the basic standard value that each item property value in the user-interest library of targeted customer is the item property, meter
The business of the value of identical item property and the basic standard value in other users interest storehouse is calculated, and defines this business and is weighed for similarity
Weight；
Similarity weight of the other users relative to targeted customer is calculated, and confirms that the current commodity of other users is inclined
It is good；
Items list of the targeted customer relative to other users currently without preference is obtained, other users are calculated in foundation
It is ranked up relative to targeted customer's similarity weight, predicts the potential commodity preference of targeted customer, obtain Recommendations list.
Specifically, commodity collaborative filtering includes：
By the image of end article in user-interest library and corresponding word description and the image in electric business platform information storehouse and
Corresponding word description is compared, and is respectively compared image similarity and word repetitive rate, and define image similarity and word weight
Multiple rate plus and be commodity similar value；
It is ranked up from high to low according to commodity approximation is calculated, obtains Recommendations list.
On the other hand, present invention also offers a kind of user's merchandise news pushing equipment based on picture attribute extraction, bag
Include：
Scene Recognition module, the default scene information inputted for identifying user in electric business platform, the default scene letter
Breath includes the combination of picture, text or picture and text information；
Acquisition module, for obtaining the text message and pictorial information of user's commodity interested in electric business platform information storehouse,
User/commodity basic information table is made；
Image characteristics extraction module, the characteristics of image for the neutral net extraction commodity using training in advance；The figure
As characteristic extracting module includes color of image extraction unit, picture shape extraction unit, image texture extraction unit；
Item property definition module, the semantic feature corresponding for obtaining above-mentioned characteristics of image, semantic feature is defined
For item property；
User-interest library builds module, for building comprising the incidence relation between merchandise classification, item property type
The user-interest library of knowledge mapping framework and item property ABC unit；
Recommending module, for forming Recommendations list and pushing merchandise news to user；
The output end of the scene Recognition module is connected with the input of acquisition module, the output end of the acquisition module with
The input of picture feature extraction module is connected, and output end and the item property of the picture feature extraction module define mould
Block input is connected, the input and the input phase of user-interest library structure module of the item property definition module
Even；The input of the recommending module is connected with user-interest library structure module.
Wherein, recommending module includes：User collaborative recommends submodule and commodity Collaborative Recommendation submodule；
The user collaborative recommends submodule, including frequency statistics unit, definition basic standard value cell, similarity weight
Computing unit, similarity weight sequencing unit；
Frequency statistics unit, for counting user electric business platform search, browse, buy and the frequency of collecting commodities,
Confirm the current preference of user；
Define basic standard value cell, some item property in user-interest library for defining some targeted customer
It is worth the basic standard value for the item property；
Similarity weight calculation unit, for calculating the value of identical item property and the basis in other users interest storehouse
The business of standard value；
Similarity weight sequencing unit, other users are calculated for foundation and enter relative to targeted customer's similarity weight
The sequence of row commodity to be recommended, forms recommendation list；
The commodity Collaborative Recommendation submodule, including comparing unit, similar value computing unit, similar value sequencing unit；
Comparing unit, for by the image of end article in user-interest library and corresponding word description and electric business platform information
Image and corresponding word description in storehouse are compared, and are respectively compared image similarity and word repetitive rate；
Similar value computing unit, for calculate image similarity and word repetitive rate plus and, and define this and add and be business
Product similar value；
Similar value sequencing unit, commodity approximation is calculated for foundation and is ranked up from high to low, obtain recommending business
Product list；
The frequency statistics unit, the input of definition basic standard value cell are connected with user-interest library, the frequency
The input of statistic unit, the output end of definition basic standard value cell respectively with similarity weight calculation unit is connected, described
The output end of similarity weight calculation unit is connected with the similarity weight sequencing unit；
The output end of the comparing unit is connected with the input of the similar value computing unit, and the similar value calculates single
The output end of member is connected with the input of the similar value sequencing unit.
The beneficial effect that technical scheme provided in an embodiment of the present invention is brought is：
1st, user's merchandise news method for pushing of the present invention based on picture attribute extraction, utilizes the picture category of commodity
Property combination knowledge mapping, establish the user individual database of comparatively perfect, it is real using the collaborative filtering based on user and commodity
The accurate push of existing similar commodity, novel and comfortable purchase experiences are brought for user；
2nd, user's merchandise news method for pushing of the present invention based on picture attribute extraction is establishing user individual
After database, by the collaborative filtering of user, other preferences of user are excavated, the commodity that push may be interested, pass through commodity
Collaborative filtering, the information expressed by the word description and picture of composite merchandise, push similar commodity；
3rd, user's merchandise news method for pushing of the present invention based on picture attribute extraction avoids prior art list
The pure limitation for considering commodity character description information, the information source searched for be used as by pictorial information, avoid ignoring commodity its
His characteristic, realize and more comprehensively, accurately recommend；
4th, it is of the present invention based on picture attribute extraction user's merchandise news method for pushing in the prior art with
On the basis of the purchase interest that record finds user is bought and browsed at family, by the collaborative filtering of user, avoid in push
Hold single, be more conducive to help active user to find and meet other commodity of oneself.
Brief description of the drawings
Technical scheme in order to illustrate the embodiments of the present invention more clearly, make required in being described below to embodiment
Accompanying drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the present invention, for
For those of ordinary skill in the art, on the premise of not paying creative work, other can also be obtained according to these accompanying drawings
Accompanying drawing.
Fig. 1 is a kind of user's merchandise news method for pushing flow based on picture attribute extraction provided in an embodiment of the present invention
Figure；
Fig. 2 be the present embodiments relate to user actively triggering recommend method flow diagram；
Fig. 3 be the present embodiments relate to formation recommendation list method flow diagram；
Fig. 4 is a kind of knot of user's merchandise news pushing equipment based on picture attribute extraction provided in an embodiment of the present invention
Structure schematic diagram；
Sequence number is expressed as in figure：
41- scene Recognition modules, 42- acquisition modules, 43- image characteristics extraction modules, 44- item property definition modules,
45- user-interest libraries build module, 46- recommending modules, and 461- user collaboratives recommend submodule, 462- commodity Collaborative Recommendation submodules
Block, 463- frequency statistics units, 464- define basic standard value cell, 465- similarity weight calculation units, 466- similarities
Weight sequencing unit, 467- comparing units, 468- similar value computing units, 469- similar value sequencing units.
Embodiment
To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with attached in the embodiment of the present invention
Figure, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is only this
Invention part of the embodiment, rather than whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art exist
The every other embodiment obtained under the premise of creative work is not made, belongs to the scope of protection of the invention.
Embodiment 1
The embodiments of the invention provide a kind of user's merchandise news method for pushing based on picture attribute extraction, such as Fig. 1 institutes
Show, methods described includes：
S101：Obtain the instruction that triggering is recommended；
The instruction can be the passive triggering based on system or the active triggering based on user, be keyed in relative to simple
Word, retrieved in database, above-mentioned instruction can provide more rich information, such as picture, picture and text message
With reference to etc., the limitation that prior art considers merely commodity character description information is avoided, the letter of search is used as by pictorial information
Breath source, other characteristics for ignoring commodity are avoided, realize and more comprehensively, accurately recommend；
S201：Item property is extracted according to above-mentioned instruction, the mesh for including above-mentioned item property is searched for from user-interest library
Mark commodity；
Employ new shopping guide's pattern of picture search, and experienced image typing, Attribute Recognition, image retrieval are patrolled
Collect chain so that more accurately obtain search key, and keyword is expanded and come out for image appearance, be advantageous to rapid and convenient
Image and its corresponding commodity that user is most interested in accurately are found, novel and comfortable purchase experiences are brought to user；
The present invention using by computer identify establishes knowledge mapping corresponding to image, with reference to user electric business platform purchase
The behavior such as buy and browse, extraction shopping interest preference, establish the user-interest library of personalization；
S301：Form Recommendations list and push merchandise news to user；
On the basis of the user-interest library of foundation, using the collaborative filtering based on user and commodity, similar commodity are realized
Accurate push, bring novel and comfortable purchase experiences for user.
The beneficial effect that technical scheme provided in an embodiment of the present invention is brought is：
1st, user's merchandise news method for pushing of the present invention based on picture attribute extraction, utilizes the picture category of commodity
Property combination knowledge mapping, establish the user individual database of comparatively perfect, it is real using the collaborative filtering based on user and commodity
The accurate push of existing similar commodity, novel and comfortable purchase experiences are brought for user；
2nd, user's merchandise news method for pushing of the present invention based on picture attribute extraction is establishing user individual
After database, by the collaborative filtering of user, other preferences of user are excavated, the commodity that push may be interested, pass through commodity
Collaborative filtering, the information expressed by the word description and picture of composite merchandise, push similar commodity.
Embodiment 2
The invention provides a kind of user's merchandise news method for pushing based on picture attribute extraction, as shown in figure 1, described
Method includes：
S101：Obtain the instruction that triggering is recommended；
Wherein, the instruction that triggering is recommended in S101 includes the passive triggering command S101a based on system or businessman and is based on
The active triggering command S101b of user；
Specifically, the passive triggering command includes weather, season, birthday, red-letter day or system self-defined technical dates；
The active triggering command S101b also includes：
S102：User inputs default scene information in electric business platform, and the default scene information is picture, text or picture
With the combination of text message.
S201：Item property is extracted according to above-mentioned instruction, being searched for from the user-interest library for creating completion includes above-mentioned business
The end article of product attribute；
Further, as shown in Fig. 2 S201 includes the step of extracting item property according to active triggering command：
S202：Picture feature is extracted using the neutral net of training in advance, obtains the language corresponding with above-mentioned picture feature
Adopted feature, semantic feature is defined as end article attribute；
S203：Extract item property of the user in the text message that electric business platform inputs in default scene information.
After the extraction step, it is necessary to have been based on to user the image procossing of commodity interested, image recognition with knowing
Know vertical, the personalized knowledge mapping for user that the combination of map construction technology creates, i.e., enter in user-interest library
Row retrieval.
User-interest library is the individuation data storehouse based on user's request preference, and essence is covering for user individual
The knowledge of goods collection of illustrative plates of user's request preference information, in specific implementation, creating knowledge of goods collection of illustrative plates includes application image processing
Technology, natural language processing technique carry out target identification and image characteristics extraction, obtain entity information, relation information, attribute letter
Breath, knowledge information and applied ontology technology, entity alignment techniques, knowledge reasoning Technical Architecture knowledge mapping.
Specifically, the creation method of the user-interest library includes：
A according to user electric business platform picture searching, browse, buy and collection record, obtain electric business platform information storehouse
The merchandise news of middle user's commodity interested, is made user/commodity basic information table, and the merchandise news comprises at least commodity figure
Piece；
B using training in advance neutral net extraction user/commodity basic information table in commodity picture feature, obtain with
The corresponding semantic feature of above-mentioned picture feature, item property is defined as by semantic feature；
C is created using item property as ABC unit and the incidence relation between merchandise classification, item property type
For the user-interest library of framework.
Wherein, the acquisition of the user/commodity basic information table is based on that user is daily to be browsed and purchaser record, extracts electric business
Text message in platform webpage is classified for commodity, described to be classified based on a certain item property or any a variety of commodity
The combination of attribute.
Preferably, the concrete mode for the end article that the search from user-interest library includes item property is with basis
The item property of triggering command extraction scans for for keyword to the commodity in user-interest library；Text is keyed in relative to simple
Word, the retrieval of picture being carried out in database, the instruction that this picture or picture are combined with word provides more rich information,
The limitation that prior art considers merely commodity character description information is avoided, the information source of search is used as by pictorial information,
Other characteristics for ignoring commodity are avoided, realizes and more comprehensively, accurately recommends.
Specifically, item property described in the embodiment of the present invention include but is not limited to color, texture material, quality, type,
Style, brand, the place of production.
S301：Form Recommendations list and push merchandise news to user.
Further, as shown in figure 3, it is described formed commercial product recommending list mode include user collaborative filtering S301a and
Commodity collaborative filtering S301b.
Specifically, user collaborative filtering S301a specific steps include：
S302a：Item property classification is extracted using the text message of the item property of commodity in user-interest library,
Confirm the current commodity preference of user；
S303a：It is the basis mark of the item property to define each item property value in the user-interest library of targeted customer
Quasi- value, calculates the business of the value of identical item property and the basic standard value in other users interest storehouse, and defines this business and be
Similarity weight；
S304a：Similarity weight of the other users relative to targeted customer is calculated, and confirms the current of other users
Commodity preference；
S305a：Items list of the targeted customer relative to other users currently without preference is obtained, it is calculated in foundation
His user is ranked up relative to targeted customer's similarity weight, predicts the potential commodity preference of targeted customer, obtains Recommendations
List.
Specifically, commodity collaborative filtering S301b specific steps include：
S302b：By in the image of end article in user-interest library and corresponding word description and electric business platform information storehouse
Image and corresponding word description are compared, and are respectively compared image similarity and word repetitive rate, and define image similarity and
Word repetitive rate plus and be commodity similar value；
S303b：It is ranked up from high to low according to commodity approximation is calculated, obtains Recommendations list.
On the basis of the user-interest library of foundation, using the collaborative filtering based on user and commodity, similar commodity are realized
Accurate push, bring novel and comfortable purchase experiences for user.
It should be noted that the present embodiments relate to based on picture attribute extraction user's merchandise news method for pushing
Based on housebroken neutral net, in the present embodiment using depth convolutional neural networks (DCNN), 16 layers of VGG moulds are utilized
Type is modeled to image, carries out image expression.Convolutional neural networks utilize multilayer neural network picture engraving in from simple to
Complicated series of features, as lower level learns the simple modes such as simple shape, color, texture, constantly combination forms gradual
The complicated pattern with semantic information, such as style and features, collar feature etc..This completes image attributes is taken out
Take, such as cultivate one's moral character, academic attitude, five points of sleeves, lengthening money：
User individual Database is, it is necessary to complete to commodity, the foundation of the knowledge mapping of user.Image technique is at this
In be to the vision of each user property supplement.It is just complete by description of several dimensions to user and item property
Into being drawn a portrait to commodity, user is drawn a portrait.With reference to commodity and the attribute of user, it is seen that, it can realize here and user is purchased
The accurate recommendation bought.
In summary, the beneficial effect that technical scheme provided in an embodiment of the present invention is brought is：
1st, user's merchandise news method for pushing of the present invention based on picture attribute extraction, utilizes the picture category of commodity
Property combination knowledge mapping, establish the user individual database of comparatively perfect, it is real using the collaborative filtering based on user and commodity
The accurate push of existing similar commodity, novel and comfortable purchase experiences are brought for user；
2nd, user's merchandise news method for pushing of the present invention based on picture attribute extraction is establishing user individual
After database, by the collaborative filtering of user, other preferences of user are excavated, the commodity that push may be interested, pass through commodity
Collaborative filtering, the information expressed by the word description and picture of composite merchandise, push similar commodity；
3rd, user's merchandise news method for pushing of the present invention based on picture attribute extraction avoids prior art list
The pure limitation for considering commodity character description information, the information source searched for be used as by pictorial information, avoid ignoring commodity its
His characteristic, realize and more comprehensively, accurately recommend；
4th, it is of the present invention based on picture attribute extraction user's merchandise news method for pushing in the prior art with
On the basis of the purchase interest that record finds user is bought and browsed at family, by the collaborative filtering of user, avoid in push
Hold single, be more conducive to help active user to find and meet other commodity of oneself.
Embodiment 3
The embodiments of the invention provide a kind of user's merchandise news pushing equipment based on picture attribute extraction, such as Fig. 4 institutes
Show, including：
Scene Recognition module 41, the default scene information inputted for identifying user in electric business platform, the default scene
Information includes the combination of picture, text or picture and text information；
Acquisition module 42, for obtaining the text message of user's commodity interested and picture letter in electric business platform information storehouse
Breath, user/commodity basic information table is made；
Image characteristics extraction module 43, the characteristics of image for the neutral net extraction commodity using training in advance；It is described
Image characteristics extraction module includes color of image extraction unit, picture shape extraction unit, image texture extraction unit；
Item property definition module 44, the semantic feature corresponding for obtaining above-mentioned characteristics of image, semantic feature is determined
Justice is item property；
User-interest library builds module 45, for building comprising the incidence relation between merchandise classification, item property type
Knowledge mapping framework and item property ABC unit user-interest library；
Recommending module 46, for forming Recommendations list and pushing merchandise news to user；
The output end of the scene Recognition module 41 is connected with the input of acquisition module 42, the acquisition module 42 it is defeated
Go out end with the input of picture feature extraction module 43 to be connected, output end and the commodity of the picture feature extraction module 43
The input of attribute definition module 44 is connected, input and the user-interest library structure mould of the item property definition module 44
The input of block 45 is connected；The input of the recommending module 46 is connected with user-interest library structure module 45.
Wherein, recommending module 46 includes：User collaborative recommends submodule 461 and commodity Collaborative Recommendation submodule 462；
Further, the user collaborative recommends submodule 461, including frequency statistics unit 463, definition basic standard value
Unit 464, similarity weight calculation unit 465, similarity weight sequencing unit 466；
Frequency statistics unit 463, for counting user electric business platform search, browse, buy and the frequency of collecting commodities
It is secondary, confirm the current preference of user；
Define basic standard value cell 462, some commodity in user-interest library for defining some targeted customer
Property value is the basic standard value of the item property；
Similarity weight calculation unit 463, for calculate in other users interest storehouse the value of identical item property with it is described
The business of basic standard value；
Similarity weight sequencing unit 464, other users are calculated for foundation and are weighed relative to targeted customer's similarity
The sequence of commodity to be recommended is carried out again, forms recommendation list；
Specifically, the commodity Collaborative Recommendation submodule 462, including comparing unit 467, similar value computing unit 468, phase
Like value sequencing unit 469；
Comparing unit 467, for by the image of end article in user-interest library and corresponding word description and electric business platform
Image and corresponding word description in information bank are compared, and are respectively compared image similarity and word repetitive rate；
Similar value computing unit 468, for calculate image similarity and word repetitive rate plus and, and define this and add and be
Commodity similar value；
Similar value sequencing unit 469, commodity approximation is calculated for foundation and is ranked up from high to low, is recommended
Items list；
The frequency statistics unit 463, the input of definition basic standard value cell 464 are connected with user-interest library 465,
The frequency statistics unit 463, define the output end of basic standard value cell 464 respectively with similarity weight calculation unit 463
Input be connected, the output end of the similarity weight calculation unit 463 and the unit phase of similarity weight sequencing 464
Even；
The output end of the comparing unit 467 is connected with the input of the similar value computing unit 468, the similar value
The output end of computing unit 468 is connected with the input of the similar value sequencing unit 469.
What deserves to be explained is user's merchandise news pushing equipment involved in the present invention based on picture attribute extraction, also
Neural metwork training module can be included, for by the study to object features, entering road wheel to the object in video and image
Exterior feature detection, carry out classification based training and identify repeatedly, create for commodity identification and image characteristics extraction in electric business platform information storehouse
Neutral net；Created in the embodiment of the present invention for the nerve of commodity identification and image characteristics extraction in electric business platform information storehouse
Network is depth convolutional neural networks, utilizes multilayer neural network application simple shape, color, texture bottom layer image pattern and height
Layer semantics information pattern carries out image characteristics extraction.
In summary, the beneficial effect that technical scheme provided in an embodiment of the present invention is brought is：
1st, user's merchandise news pushing equipment of the present invention based on picture attribute extraction, compact-sized, function is complete
It is kind, using the picture attribute combination knowledge mapping of commodity, the user individual database of comparatively perfect is established, using based on user
With the collaborative filtering of commodity, the accurate push of similar commodity is realized, novel and comfortable purchase experiences are brought for user；
2nd, user's merchandise news pushing equipment of the present invention based on picture attribute extraction is establishing user individual
After database, by the collaborative filtering of user, other preferences of user are excavated, the commodity that push may be interested, pass through commodity
Collaborative filtering, the information expressed by the word description and picture of composite merchandise, push similar commodity；
3rd, user's merchandise news pushing equipment of the present invention based on picture attribute extraction avoids prior art list
The pure limitation for considering commodity character description information, the information source searched for be used as by pictorial information, avoid ignoring commodity its
His characteristic, realize and more comprehensively, accurately recommend；
4th, it is of the present invention based on picture attribute extraction user's merchandise news pushing equipment in the prior art with
On the basis of the purchase interest that record finds user is bought and browsed at family, by the collaborative filtering of user, avoid in push
Hold single, be more conducive to help active user to find and meet other commodity of oneself.
It should be noted that：User's merchandise news pushing equipment based on picture attribute extraction that above-described embodiment provides exists
, can be as needed only with the division progress of above-mentioned each functional module for example, in practical application during to user's Recommendations
And complete above-mentioned function distribution by different functional modules, i.e., the internal structure of equipment is divided into different functional modules,
To complete all or part of function described above.In addition, the user based on picture attribute extraction that above-described embodiment provides
Merchandise news pushing equipment belongs to same design with user's merchandise news method for pushing embodiment based on picture attribute extraction, its
Specific implementation process refers to embodiment of the method, repeats no more here.
One of ordinary skill in the art will appreciate that hardware can be passed through by realizing all or part of step of above-described embodiment
To complete, by program the hardware of correlation can also be instructed to complete, described program can be stored in a kind of computer-readable
In storage medium, storage medium mentioned above can be read-only storage, disk or CD etc..
The foregoing is only presently preferred embodiments of the present invention, be not intended to limit the invention, it is all the present invention spirit and
Within principle, any modification, equivalent substitution and improvements made etc., it should be included in the scope of the protection.
Claims (10)
1. a kind of user's merchandise news method for pushing based on picture attribute extraction, it is characterised in that methods described includes：
Obtain the instruction that triggering is recommended；
Item property is extracted according to above-mentioned instruction, the end article for including above-mentioned item property is searched for from user-interest library；
Form Recommendations list and push merchandise news to user.
2. user's merchandise news method for pushing according to claim 1 based on picture attribute extraction, it is characterised in that institute
Stating the instruction that triggering is recommended includes the passive triggering command based on system or businessman and the active triggering command based on user；
The passive triggering command includes weather, season, birthday, red-letter day or system self-defined technical dates；
The active triggering command includes user's active typing and presets scene information.
3. user's merchandise news method for pushing according to claim 2 based on picture attribute extraction, it is characterised in that institute
Stating the step of extracting item property according to active triggering command includes：
User inputs default scene information in electric business platform, and the default scene information is picture, text or picture and text letter
The combination of breath；
Picture feature is extracted using the neutral net of training in advance, obtains the semantic feature corresponding with above-mentioned picture feature, will
Semantic feature is defined as end article attribute；
The active triggering command extraction item property also includes the text that user is inputted in default scene information in electric business platform
The item property that information is extracted.
4. user's merchandise news method for pushing according to claim 1 based on picture attribute extraction, it is characterised in that institute
Stating the creation method of user-interest library includes：
According to user electric business platform picture searching, browse, buy and collection record, obtain user in electric business platform information storehouse
The merchandise news of commodity interested, is made user/commodity basic information table, and the merchandise news comprises at least commodity picture；
Using training in advance neutral net extraction user/commodity basic information table in commodity picture feature, obtain with it is above-mentioned
The corresponding semantic feature of picture feature, item property is defined as by semantic feature；
It is ABC unit and incidence relation between merchandise classification, item property type as framework to create using item property
User-interest library.
5. user's merchandise news method for pushing according to claim 1 based on picture attribute extraction, it is characterised in that institute
The concrete mode for stating the end article that the search from user-interest library includes item property is the business to be extracted according to triggering command
Product attribute is that keyword scans for the commodity in user-interest library；The item property includes color, texture material, product
Matter, type, style, brand, the place of production.
6. user's merchandise news method for pushing according to claim 1 based on picture attribute extraction, it is characterised in that institute
State to be formed commercial product recommending list mode include user collaborative filtering and commodity collaborative filtering.
7. user's merchandise news method for pushing according to claim 6 based on picture attribute extraction, it is characterised in that institute
Stating the specific method of user collaborative filtering includes：
Item property classification is extracted using the text message of the item property of commodity in user-interest library, confirms user's
Current commodity preference；
The basic standard value that each item property value in the user-interest library of targeted customer is the item property is defined, calculates it
The business of the value of identical item property and the basic standard value in his user-interest library, and it is similarity weight to define this business；
Similarity weight of the other users relative to targeted customer is calculated, and confirms the current commodity preference of other users；
Items list of the targeted customer relative to other users currently without preference is obtained, it is relative according to other users are calculated
It is ranked up in targeted customer's similarity weight, predicts the potential commodity preference of targeted customer, obtain Recommendations list.
8. user's merchandise news method for pushing according to claim 6 based on picture attribute extraction, it is characterised in that institute
Stating commodity collaborative filtering includes：
By the image of end article in user-interest library and corresponding word description and the image in electric business platform information storehouse and corresponding
Word description is compared, and is respectively compared image similarity and word repetitive rate, and defines image similarity and word repetitive rate
Plus and be commodity similar value；
It is ranked up from high to low according to commodity approximation is calculated, obtains Recommendations list.
9. a kind of user's merchandise news pushing equipment based on picture attribute extraction, it is characterised in that the equipment includes：
Scene Recognition module, the default scene information inputted for identifying user in electric business platform, the default scene information bag
Include the combination of picture, text or picture and text information；
Acquisition module, for obtaining the text message and pictorial information of user's commodity interested in electric business platform information storehouse, it is made
User/commodity basic information table；
Image characteristics extraction module, the characteristics of image for the neutral net extraction commodity using training in advance；Described image is special
Levying extraction module includes color of image extraction unit, picture shape extraction unit, image texture extraction unit；
Item property definition module, the semantic feature corresponding for obtaining above-mentioned characteristics of image, business is defined as by semantic feature
Product attribute；
User-interest library builds module, and the knowledge of the incidence relation between merchandise classification, item property type is included for building
The user-interest library of collection of illustrative plates framework and item property ABC unit；
Recommending module, for forming Recommendations list and pushing merchandise news to user；
The output end of the scene Recognition module is connected with the input of acquisition module, the output end and picture of the acquisition module
The input of characteristic extracting module is connected, and the output end of the picture feature extraction module and the item property definition module are defeated
Enter end to be connected, the input of the item property definition module is connected with the input of user-interest library structure module；Institute
The input for stating recommending module is connected with user-interest library structure module.
10. user's merchandise news pushing equipment according to claim 9 based on picture attribute extraction, it is characterised in that
The recommending module includes：User collaborative recommends submodule and commodity Collaborative Recommendation submodule；
The user collaborative recommends submodule, including frequency statistics unit, definition basic standard value cell, similarity weight calculation
Unit, similarity weight sequencing unit；
Frequency statistics unit, for counting user electric business platform search, browse, buy and the frequency of collecting commodities, confirm
The current preference of user；
Basic standard value cell is defined, some item property value in the user-interest library for defining some targeted customer is
The basic standard value of the item property；
Similarity weight calculation unit, for calculating the value of identical item property and the basic standard in other users interest storehouse
The business of value；
Similarity weight sequencing unit, other users are calculated for foundation and are treated relative to targeted customer's similarity weight
The sequence of Recommendations, form recommendation list；
The commodity Collaborative Recommendation submodule, including comparing unit, similar value computing unit, similar value sequencing unit；
Comparing unit, for by the image of end article in user-interest library and corresponding word description and electric business platform information storehouse
Image and corresponding word description be compared, be respectively compared image similarity and word repetitive rate；
Similar value computing unit, for calculate image similarity and word repetitive rate plus and, and define this and add and be commodity phase
Like value；
Similar value sequencing unit, commodity approximation is calculated for foundation and is ranked up from high to low, obtain Recommendations row
Table；
The frequency statistics unit, the input of definition basic standard value cell are connected with user-interest library, the frequency statistics
The input of unit, the output end of definition basic standard value cell respectively with similarity weight calculation unit is connected, described similar
The output end of degree weight calculation unit is connected with the similarity weight sequencing unit；
The output end of the comparing unit is connected with the input of the similar value computing unit, the similar value computing unit
Output end is connected with the input of the similar value sequencing unit.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710833188.1A CN107730343A (en) | 2017-09-15 | 2017-09-15 | A kind of user's merchandise news method for pushing and equipment based on picture attribute extraction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710833188.1A CN107730343A (en) | 2017-09-15 | 2017-09-15 | A kind of user's merchandise news method for pushing and equipment based on picture attribute extraction |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107730343A true CN107730343A (en) | 2018-02-23 |
Family
ID=61207164
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710833188.1A Pending CN107730343A (en) | 2017-09-15 | 2017-09-15 | A kind of user's merchandise news method for pushing and equipment based on picture attribute extraction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107730343A (en) |
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108520058A (en) * | 2018-03-30 | 2018-09-11 | 维沃移动通信有限公司 | A kind of Business Information recommends method and mobile terminal |
| CN109068173A (en) * | 2018-09-28 | 2018-12-21 | 石家庄永捷科技有限公司 | A kind of method for processing video frequency and video process apparatus |
| CN109739970A (en) * | 2018-12-29 | 2019-05-10 | 联想(北京)有限公司 | Information processing method and device and electronic equipment |
| CN109886781A (en) * | 2019-01-31 | 2019-06-14 | 广东智媒云图科技股份有限公司 | Products Show method, apparatus, equipment and storage medium based on drawing behavior |
| CN110020913A (en) * | 2019-02-20 | 2019-07-16 | 中国人民财产保险股份有限公司 | Products Show method, equipment and storage medium |
| CN110473040A (en) * | 2018-05-10 | 2019-11-19 | 北京三快在线科技有限公司 | A kind of Products Show method and device, electronic equipment |
| CN110473042A (en) * | 2018-05-11 | 2019-11-19 | 北京京东尚科信息技术有限公司 | For obtaining the method and device of information |
| CN110490748A (en) * | 2019-07-10 | 2019-11-22 | 阿里巴巴集团控股有限公司 | Item recommendation method and device based on order |
| CN110827099A (en) * | 2018-08-07 | 2020-02-21 | 北京优酷科技有限公司 | Household commodity recommendation method, client and server |
| CN110827063A (en) * | 2019-10-18 | 2020-02-21 | 用友网络科技股份有限公司 | Multi-strategy fused commodity recommendation method, device, terminal and storage medium |
| CN111143590A (en) * | 2019-12-25 | 2020-05-12 | 上海云从企业发展有限公司 | Image filtering method, system, device and machine readable medium |
| CN111191012A (en) * | 2018-10-25 | 2020-05-22 | 财团法人资讯工业策进会 | Knowledge graph generation apparatus, method and computer program product thereof |
| WO2020114151A1 (en) * | 2018-12-06 | 2020-06-11 | 深圳云天励飞技术有限公司 | Clothing knowledge graph display method and apparatus, and graph server and storage medium |
| CN111695922A (en) * | 2019-03-15 | 2020-09-22 | 北京沃东天骏信息技术有限公司 | Potential user determination method and device, storage medium and electronic equipment |
| CN111882409A (en) * | 2020-09-28 | 2020-11-03 | 武汉卓尔数字传媒科技有限公司 | Method and device for recommending main body and electronic equipment |
| CN112070511A (en) * | 2020-08-12 | 2020-12-11 | 上海连尚网络科技有限公司 | Method and equipment for detecting unqualified commodities |
| CN112313687A (en) * | 2018-06-11 | 2021-02-02 | 奥姆尼欧斯株式会社 | Method and device for measuring influence by using social network |
| CN112417287A (en) * | 2020-11-24 | 2021-02-26 | 乐聚(深圳)机器人技术有限公司 | Building block searching method, model training method, device, equipment and storage medium |
| CN112561581A (en) * | 2020-12-14 | 2021-03-26 | 珠海格力电器股份有限公司 | Recommendation method and device, electronic equipment and storage medium |
| CN112883281A (en) * | 2021-03-29 | 2021-06-01 | 杭州同声相应科技有限公司 | User clustering search system based on deep learning |
| CN113065922A (en) * | 2021-04-13 | 2021-07-02 | 武汉市章瀚电子商务有限公司 | Electronic commerce shopping platform commodity intelligent matching recommendation method based on mobile internet and big data analysis |
| CN113191858A (en) * | 2021-06-10 | 2021-07-30 | 数贸科技（北京）有限公司 | Commodity display method and device based on picture search |
| CN113570428A (en) * | 2021-07-23 | 2021-10-29 | 上海普洛斯普新数字科技有限公司 | System for screening consistency of online commodities |
| CN113656582A (en) * | 2021-08-17 | 2021-11-16 | 北京百度网讯科技有限公司 | Training method of neural network model, image retrieval method, device and medium |
| CN113744002A (en) * | 2020-05-27 | 2021-12-03 | 北京沃东天骏信息技术有限公司 | Method, device, equipment and computer readable medium for pushing information |
| CN113869048A (en) * | 2021-09-30 | 2021-12-31 | 广州华多网络科技有限公司 | Commodity object searching method and device, equipment, medium and product thereof |
| CN115080865A (en) * | 2022-08-19 | 2022-09-20 | 山东智豆数字科技有限公司 | E-commerce data operation management system based on multidimensional data analysis |
| CN116522011A (en) * | 2023-05-16 | 2023-08-01 | 深圳九星互动科技有限公司 | Big data-based pushing method and pushing system |
| CN117808564A (en) * | 2024-03-01 | 2024-04-02 | 深圳市拓保软件有限公司 | User data recommendation method and device based on artificial intelligence |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW201216185A (en) * | 2010-10-06 | 2012-04-16 | Chunghwa Telecom Co Ltd | Multi-aspect merchandise recommendation method and system thereof |
| CN102663010A (en) * | 2012-03-20 | 2012-09-12 | 复旦大学 | Personalized image browsing and recommending method based on labelling semantics and system thereof |
| CN104504055A (en) * | 2014-12-19 | 2015-04-08 | 常州飞寻视讯信息科技有限公司 | Commodity similarity calculation method and commodity recommending system based on image similarity |
| TW201537490A (en) * | 2014-03-28 | 2015-10-01 | Richplay Information Co Ltd | Method for recommending suppliers |
| CN105528374A (en) * | 2014-10-21 | 2016-04-27 | 苏宁云商集团股份有限公司 | A commodity recommendation method in electronic commerce and a system using the same |
| CN106021551A (en) * | 2016-07-01 | 2016-10-12 | 深圳追科技有限公司 | Consumption auxiliary decision making method based on screenshot information recognition |
| CN106202362A (en) * | 2016-07-07 | 2016-12-07 | Tcl集团股份有限公司 | Image recommendation method and image recommendation device |
| US20170031952A1 (en) * | 2015-07-28 | 2017-02-02 | The Toronto-Dominion Bank | Method and system for identifying a property for purchase using image processing |
| CN106504064A (en) * | 2016-10-25 | 2017-03-15 | 清华大学 | Clothes classification based on depth convolutional neural networks recommends method and system with collocation |
| CN106557563A (en) * | 2016-11-15 | 2017-04-05 | 北京百度网讯科技有限公司 | Query statement based on artificial intelligence recommends method and device |
-
2017
- 2017-09-15 CN CN201710833188.1A patent/CN107730343A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW201216185A (en) * | 2010-10-06 | 2012-04-16 | Chunghwa Telecom Co Ltd | Multi-aspect merchandise recommendation method and system thereof |
| CN102663010A (en) * | 2012-03-20 | 2012-09-12 | 复旦大学 | Personalized image browsing and recommending method based on labelling semantics and system thereof |
| TW201537490A (en) * | 2014-03-28 | 2015-10-01 | Richplay Information Co Ltd | Method for recommending suppliers |
| CN105528374A (en) * | 2014-10-21 | 2016-04-27 | 苏宁云商集团股份有限公司 | A commodity recommendation method in electronic commerce and a system using the same |
| CN104504055A (en) * | 2014-12-19 | 2015-04-08 | 常州飞寻视讯信息科技有限公司 | Commodity similarity calculation method and commodity recommending system based on image similarity |
| US20170031952A1 (en) * | 2015-07-28 | 2017-02-02 | The Toronto-Dominion Bank | Method and system for identifying a property for purchase using image processing |
| CN106021551A (en) * | 2016-07-01 | 2016-10-12 | 深圳追科技有限公司 | Consumption auxiliary decision making method based on screenshot information recognition |
| CN106202362A (en) * | 2016-07-07 | 2016-12-07 | Tcl集团股份有限公司 | Image recommendation method and image recommendation device |
| CN106504064A (en) * | 2016-10-25 | 2017-03-15 | 清华大学 | Clothes classification based on depth convolutional neural networks recommends method and system with collocation |
| CN106557563A (en) * | 2016-11-15 | 2017-04-05 | 北京百度网讯科技有限公司 | Query statement based on artificial intelligence recommends method and device |
Cited By (44)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108520058A (en) * | 2018-03-30 | 2018-09-11 | 维沃移动通信有限公司 | A kind of Business Information recommends method and mobile terminal |
| CN108520058B (en) * | 2018-03-30 | 2021-09-24 | 维沃移动通信有限公司 | Merchant information recommendation method and mobile terminal |
| CN110473040A (en) * | 2018-05-10 | 2019-11-19 | 北京三快在线科技有限公司 | A kind of Products Show method and device, electronic equipment |
| CN110473040B (en) * | 2018-05-10 | 2021-11-19 | 北京三快在线科技有限公司 | Product recommendation method and device and electronic equipment |
| CN110473042A (en) * | 2018-05-11 | 2019-11-19 | 北京京东尚科信息技术有限公司 | For obtaining the method and device of information |
| CN110473042B (en) * | 2018-05-11 | 2022-02-01 | 北京京东尚科信息技术有限公司 | Method and device for acquiring information |
| CN112313687A (en) * | 2018-06-11 | 2021-02-02 | 奥姆尼欧斯株式会社 | Method and device for measuring influence by using social network |
| CN112313687B (en) * | 2018-06-11 | 2024-05-07 | 奥姆尼欧斯株式会社 | Method and device for measuring influence by utilizing social network |
| CN110827099B (en) * | 2018-08-07 | 2022-05-13 | 阿里巴巴（中国）有限公司 | Household commodity recommendation method, client and server |
| CN110827099A (en) * | 2018-08-07 | 2020-02-21 | 北京优酷科技有限公司 | Household commodity recommendation method, client and server |
| CN109068173A (en) * | 2018-09-28 | 2018-12-21 | 石家庄永捷科技有限公司 | A kind of method for processing video frequency and video process apparatus |
| CN111191012B (en) * | 2018-10-25 | 2023-05-30 | 财团法人资讯工业策进会 | Knowledge graph generation device and method and computer readable storage medium thereof |
| CN111191012A (en) * | 2018-10-25 | 2020-05-22 | 财团法人资讯工业策进会 | Knowledge graph generation apparatus, method and computer program product thereof |
| WO2020114151A1 (en) * | 2018-12-06 | 2020-06-11 | 深圳云天励飞技术有限公司 | Clothing knowledge graph display method and apparatus, and graph server and storage medium |
| CN109739970B (en) * | 2018-12-29 | 2021-07-16 | 联想(北京)有限公司 | Information processing method and device and electronic equipment |
| CN109739970A (en) * | 2018-12-29 | 2019-05-10 | 联想(北京)有限公司 | Information processing method and device and electronic equipment |
| CN109886781A (en) * | 2019-01-31 | 2019-06-14 | 广东智媒云图科技股份有限公司 | Products Show method, apparatus, equipment and storage medium based on drawing behavior |
| CN109886781B (en) * | 2019-01-31 | 2023-05-09 | 广东智媒云图科技股份有限公司 | Product recommendation method, device, equipment and storage medium based on painting behaviors |
| CN110020913A (en) * | 2019-02-20 | 2019-07-16 | 中国人民财产保险股份有限公司 | Products Show method, equipment and storage medium |
| CN111695922A (en) * | 2019-03-15 | 2020-09-22 | 北京沃东天骏信息技术有限公司 | Potential user determination method and device, storage medium and electronic equipment |
| CN110490748A (en) * | 2019-07-10 | 2019-11-22 | 阿里巴巴集团控股有限公司 | Item recommendation method and device based on order |
| CN110490748B (en) * | 2019-07-10 | 2023-11-28 | 创新先进技术有限公司 | Order-based project recommendation method and device |
| CN110827063A (en) * | 2019-10-18 | 2020-02-21 | 用友网络科技股份有限公司 | Multi-strategy fused commodity recommendation method, device, terminal and storage medium |
| CN111143590A (en) * | 2019-12-25 | 2020-05-12 | 上海云从企业发展有限公司 | Image filtering method, system, device and machine readable medium |
| CN113744002A (en) * | 2020-05-27 | 2021-12-03 | 北京沃东天骏信息技术有限公司 | Method, device, equipment and computer readable medium for pushing information |
| CN112070511A (en) * | 2020-08-12 | 2020-12-11 | 上海连尚网络科技有限公司 | Method and equipment for detecting unqualified commodities |
| CN111882409B (en) * | 2020-09-28 | 2020-12-08 | 武汉卓尔数字传媒科技有限公司 | Method and device for recommending main body and electronic equipment |
| CN111882409A (en) * | 2020-09-28 | 2020-11-03 | 武汉卓尔数字传媒科技有限公司 | Method and device for recommending main body and electronic equipment |
| CN112417287A (en) * | 2020-11-24 | 2021-02-26 | 乐聚(深圳)机器人技术有限公司 | Building block searching method, model training method, device, equipment and storage medium |
| CN112561581A (en) * | 2020-12-14 | 2021-03-26 | 珠海格力电器股份有限公司 | Recommendation method and device, electronic equipment and storage medium |
| CN112883281A (en) * | 2021-03-29 | 2021-06-01 | 杭州同声相应科技有限公司 | User clustering search system based on deep learning |
| CN113065922B (en) * | 2021-04-13 | 2021-12-07 | 南京莫愁智慧信息科技有限公司 | Electronic commerce shopping platform commodity intelligent matching recommendation method based on mobile internet and big data analysis |
| CN113065922A (en) * | 2021-04-13 | 2021-07-02 | 武汉市章瀚电子商务有限公司 | Electronic commerce shopping platform commodity intelligent matching recommendation method based on mobile internet and big data analysis |
| CN113191858A (en) * | 2021-06-10 | 2021-07-30 | 数贸科技（北京）有限公司 | Commodity display method and device based on picture search |
| CN113570428B (en) * | 2021-07-23 | 2024-02-02 | 上海普洛斯普新数字科技有限公司 | System for be used for online commodity uniformity screening |
| CN113570428A (en) * | 2021-07-23 | 2021-10-29 | 上海普洛斯普新数字科技有限公司 | System for screening consistency of online commodities |
| CN113656582A (en) * | 2021-08-17 | 2021-11-16 | 北京百度网讯科技有限公司 | Training method of neural network model, image retrieval method, device and medium |
| CN113869048A (en) * | 2021-09-30 | 2021-12-31 | 广州华多网络科技有限公司 | Commodity object searching method and device, equipment, medium and product thereof |
| CN115080865B (en) * | 2022-08-19 | 2022-11-04 | 山东智豆数字科技有限公司 | E-commerce data operation management system based on multidimensional data analysis |
| CN115080865A (en) * | 2022-08-19 | 2022-09-20 | 山东智豆数字科技有限公司 | E-commerce data operation management system based on multidimensional data analysis |
| CN116522011A (en) * | 2023-05-16 | 2023-08-01 | 深圳九星互动科技有限公司 | Big data-based pushing method and pushing system |
| CN116522011B (en) * | 2023-05-16 | 2024-02-13 | 深圳九星互动科技有限公司 | Big data-based pushing method and pushing system |
| CN117808564A (en) * | 2024-03-01 | 2024-04-02 | 深圳市拓保软件有限公司 | User data recommendation method and device based on artificial intelligence |
| CN117808564B (en) * | 2024-03-01 | 2024-05-28 | 深圳市拓保软件有限公司 | User data recommendation method and device based on artificial intelligence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107730343A (en) | A kind of user's merchandise news method for pushing and equipment based on picture attribute extraction | |
| CN110750656B (en) | Multimedia detection method based on knowledge graph | |
| CN103399883B (en) | Method and system for performing personalized recommendation according to user interest points/concerns | |
| CN102339306B (en) | Sketch-based image search | |
| CN108921670B (en) | Drug transaction recommendation method fusing potential interest, spatio-temporal data and category popularity of user | |
| CN107861972A (en) | The method and apparatus of the full result of display of commodity after a kind of user's typing merchandise news | |
| US20150242525A1 (en) | System for referring to and/or embedding posts within other post and posts within any part of another post | |
| CN107609152A (en) | Method and apparatus for expanding query formula | |
| CN110598130A (en) | Movie recommendation method integrating heterogeneous information network and deep learning | |
| CN107944898A (en) | The automatic discovery of advertisement putting building information and sort method | |
| CN107729900A (en) | It is a kind of that the method and apparatus for completing typing information completion is extracted using picture attribute | |
| CN112380453B (en) | Article recommendation method and device, storage medium and equipment | |
| CN106557480A (en) | Implementation method and device that inquiry is rewritten | |
| Vempati et al. | Enabling hyper-personalisation: Automated ad creative generation and ranking for fashion e-commerce | |
| CN101950400A (en) | Network shopping guiding method | |
| CN107808300A (en) | A kind of advertisement push system and method for pushing based on wireless domain marketing platform | |
| CN105373633A (en) | Top-k subscription inquiring and matching method of position sensing subscription/publishing system | |
| CN113505311B (en) | Scenic spot interaction recommendation method based on' potential semantic space | |
| CN105916032A (en) | Video recommendation method and video recommendation terminal equipment | |
| Li et al. | Metro-wordle: An interactive visualization for urban text distributions based on wordle | |
| CN107562966A (en) | The optimization system and method based on intelligence learning for web page interlinkage retrieval ordering | |
| CN105740387B (en) | A kind of scientific and technical literature recommended method based on author's frequent mode | |
| CN106484889A (en) | The flooding method and apparatus of Internet resources | |
| CN107622071A (en) | By indirect correlation feedback without clothes image searching system and the method looked under source | |
| Yang et al. | Design and application of handicraft recommendation system based on improved hybrid algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180223 |