CN106874522A - Information recommendation method, device, storage medium and processor - Google Patents
Information recommendation method, device, storage medium and processor Download PDFInfo
- Publication number
- CN106874522A CN106874522A CN201710198948.6A CN201710198948A CN106874522A CN 106874522 A CN106874522 A CN 106874522A CN 201710198948 A CN201710198948 A CN 201710198948A CN 106874522 A CN106874522 A CN 106874522A
- Authority
- CN
- China
- Prior art keywords
- data
- information
- behavior data
- historical behavior
- recommended models
- 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
- 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
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of information recommendation method, device, storage medium and processor.Wherein, the method includes：The collection user behavioral data produced at access target website and the first attribute data of targeted website, wherein, behavioral data includes the first current behavior data and the first historical behavior data；First historical behavior data and the first attribute data are respectively processed and obtain the second historical behavior data and the second attribute data, and then according to the second historical behavior data and the second attribute data generation User Information Database, wherein, comprising multiple data samples in User Information Database；Presupposed information recommended models are trained according to data sample, obtain first information recommended models；The first current behavior data are processed according to first information recommended models, so as to generate the recommendation information of targeted website.The present invention solves the network information present in prior art and recommends the degree of accuracy and less efficient technical problem.
Description
Technical field
The present invention relates to internet arena, in particular to a kind of information recommendation method, device, storage medium and place
Reason device.
Background technology
With the arriving in big data epoch, the information content in network is presented Exponential growth, information overload is brought therewith
Problem.Intelligent recommendation system is to solve one of information overload most effective way, and the commending system based on big data is by adopting
Collect the behavioural information of user, mass data of analyzing and researching simultaneously therefrom excavates useful information, and then draw the interest preference of user
And the information being associated with its interest preference is pushed to user.Intelligent recommendation system has been increasingly becoming the research heat of message area
One of point.
In intelligent recommendation system, the validity of recommending data, accuracy, real-time are the important indicators of systematic function.
The scheme that existing intelligent recommendation system is used is broadly divided into following three class：
First, the recommendation based on colony, using collaborative filtering, collects the information of user and has similar behavior to user
The information of colony, the recommendation of unique user is realized according to corresponding community information.
2nd, content-based recommendation, collects the information of user and the dominant correlation attribute information of commodity, is that user recommends it
Bought or the similar product of browsed commodity.
3rd, hybrid plan, collects multiparty data, with reference to number of mechanisms such as the recommendation based on colony, content-based recommendations,
Recommended jointly using multiple technologies.
Wherein, the explanatory difference of recommendation mechanisms based on colony, and Sparse Problem is faced with, and content-based recommendation
Its performance of mechanism is limited by the dominant attribute of product, cannot effectively be pushed away during or shortage imperfect for attribute tags
Recommend.Additionally, the proposed algorithm in above-mentioned three kinds of schemes has cold start-up, the new user for no buying behavior recommends
As a result precision is very low.To sum up, there is the network information in the prior art and recommend the degree of accuracy and less efficient technical problem.
For above-mentioned problem, effective solution is not yet proposed at present.
The content of the invention
A kind of information recommendation method, device, storage medium and processor are the embodiment of the invention provides, it is existing at least to solve
There is the network information present in technology to recommend the degree of accuracy and less efficient technical problem.
One side according to embodiments of the present invention, there is provided a kind of information recommendation method, the method includes：Collection user
The first attribute data of produced behavioral data and above-mentioned targeted website at access target website, wherein, above-mentioned behavior number
According to including the first current behavior data and the first historical behavior data；To above-mentioned first historical behavior data and above-mentioned first attribute
Data are respectively processed and obtain the second historical behavior data and the second attribute data, and then according to above-mentioned second historical behavior number
User Information Database is generated according to above-mentioned second attribute data, wherein, comprising multiple data in above-mentioned User Information Database
Sample；Presupposed information recommended models are trained according to above-mentioned data sample, obtain first information recommended models, wherein, on
It is the above-mentioned presupposed information recommended models after convergence to state first information recommended models；According to above-mentioned first information recommended models to upper
State the first current behavior data to be processed, so as to generate the recommendation information of above-mentioned targeted website.
Further, it is above-mentioned that above-mentioned first historical behavior data and above-mentioned first attribute data process obtaining second
Historical behavior data and the second attribute data include：Line number is entered to above-mentioned first historical behavior data and above-mentioned first attribute data
According to cleaning, the 3rd historical behavior data and the 3rd attribute data are obtained；To above-mentioned 3rd historical behavior data and above-mentioned 3rd category
Property data carry out data prediction, obtain above-mentioned second historical behavior data and above-mentioned second attribute data, wherein, above-mentioned data
The mode of pretreatment at least includes one of the following：The loading of data pick-up, data conversion and data.
Further, it is above-mentioned presupposed information recommended models are trained according to above-mentioned data sample, obtain the first information
Recommended models include：Above-mentioned data sample is processed according to default training algorithm, obtains user characteristics vector, wherein, on
Stating default training algorithm at least includes one of the following：Collaborative filtering, data parallel algorithm and model based on matrix decomposition
Parallel algorithm；Above-mentioned first information recommended models are obtained according to above-mentioned user characteristics vector.
Further, the above method also includes：Above-mentioned first current behavior data are processed, the second current line is obtained
It is data；Above-mentioned User Information Database is updated according to above-mentioned second current behavior data；Above-mentioned user letter after according to renewal
The above-mentioned data sample included in breath database is trained to above-mentioned first information recommended models, obtains the second information recommendation mould
Type, wherein, above-mentioned second information recommendation model is the above-mentioned first information recommended models after convergence.
Further, it is above-mentioned presupposed information recommended models are trained according to above-mentioned data sample before, above-mentioned side
Method also includes：Model parameter to above-mentioned presupposed information recommended models carries out initialization process.
Another aspect according to embodiments of the present invention, additionally provides a kind of information recommending apparatus, and the device includes：Collection is single
Unit, the first attribute data for gathering user's behavioral data produced at access target website and above-mentioned targeted website,
Wherein, above-mentioned behavioral data includes the first current behavior data and the first historical behavior data；First processing units, for
State the first historical behavior data and above-mentioned first attribute data is respectively processed and obtains the second historical behavior data and the second category
Property data, and then according to above-mentioned second historical behavior data and above-mentioned second attribute data generation User Information Database, wherein,
Comprising multiple data samples in above-mentioned User Information Database；Second processing unit, for according to above-mentioned data sample to default
Information recommendation model is trained, and obtains first information recommended models, wherein, above-mentioned first information recommended models are after restraining
Above-mentioned presupposed information recommended models；3rd processing unit, for current to above-mentioned first according to above-mentioned first information recommended models
Behavioral data is processed, so as to generate the recommendation information of above-mentioned targeted website.
Further, above-mentioned first processing units include：First treatment subelement, for above-mentioned first historical behavior number
Data cleansing is carried out according to above-mentioned first attribute data, the 3rd historical behavior data and the 3rd attribute data is obtained；Second processing
Subelement, for carrying out data prediction to above-mentioned 3rd historical behavior data and above-mentioned 3rd attribute data, obtains above-mentioned
Two historical behavior data and above-mentioned second attribute data, wherein, the mode of above-mentioned data prediction at least includes one of the following：Number
Loaded according to extraction, data conversion and data.
Further, above-mentioned second processing unit includes：3rd treatment subelement, for according to default training algorithm to
State data sample to be processed, obtain user characteristics vector, wherein, above-mentioned default training algorithm at least includes one of the following：Base
Collaborative filtering, data parallel algorithm and model parallel algorithm in matrix decomposition；Subelement is obtained, for according to above-mentioned use
Family characteristic vector obtains above-mentioned first information recommended models.
Another aspect according to embodiments of the present invention, additionally provides a kind of storage medium, and above-mentioned storage medium includes storage
Program, wherein, control equipment where above-mentioned storage medium to perform above-mentioned information recommendation method when said procedure runs.
Another aspect according to embodiments of the present invention, additionally provides a kind of processor, and above-mentioned processor is used for operation program,
Wherein, above-mentioned information recommendation method is performed when said procedure runs.
In embodiments of the present invention, using the collection user behavioral data produced at access target website and target network
The mode of the first attribute data stood, wherein, behavioral data includes the first current behavior data and the first historical behavior data, leads to
Cross to be respectively processed the first historical behavior data and the first attribute data and obtain the second historical behavior data and the second attribute
Data, and then according to the second historical behavior data and the second attribute data generation User Information Database, wherein, user profile number
According to, comprising multiple data samples, and then being trained to presupposed information recommended models according to data sample in storehouse, the first letter is obtained
Breath recommended models, wherein, first information recommended models are the presupposed information recommended models after convergence, have been reached according to the first information
Recommended models are processed so as to generate the purpose of the recommendation information of targeted website the first current behavior data, it is achieved thereby that
The degree of accuracy and efficiency, the technique effect of lifting Consumer's Experience that the network information is recommended are improved, and then is solved and is deposited in the prior art
The network information recommend the degree of accuracy and less efficient technical problem.
Brief description of the drawings
Accompanying drawing described herein is used for providing a further understanding of the present invention, constitutes the part of the application, this hair
Bright schematic description and description does not constitute inappropriate limitation of the present invention for explaining the present invention.In the accompanying drawings：
Fig. 1 is the schematic flow sheet of a kind of optional information recommendation method according to embodiments of the present invention；
Fig. 2 is the schematic flow sheet of the optional information recommendation method of another kind according to embodiments of the present invention；
Fig. 3 is the schematic flow sheet of another optional information recommendation method according to embodiments of the present invention；
Fig. 4 is the schematic flow sheet of another optional information recommendation method according to embodiments of the present invention；
Fig. 5 is the structural representation of a kind of optional information recommending apparatus according to embodiments of the present invention.
Specific embodiment
In order that those skilled in the art more fully understand the present invention program, below in conjunction with the embodiment of the present invention
Accompanying drawing, is clearly and completely described to the technical scheme in the embodiment of the present invention, it is clear that described embodiment is only
The embodiment of a part of the invention, rather than whole embodiments.Based on the embodiment in the present invention, ordinary skill people
The every other embodiment that member is obtained under the premise of creative work is not made, should all belong to the model of present invention protection
Enclose.
It should be noted that term " first ", " in description and claims of this specification and above-mentioned accompanying drawing
Two " it is etc. for distinguishing similar object, without for describing specific order or precedence.It should be appreciated that so using
Data can exchange in the appropriate case, so as to embodiments of the invention described herein can with except illustrating herein or
Order beyond those of description is implemented.Additionally, term " comprising " and " having " and their any deformation, it is intended that cover
Lid is non-exclusive to be included, for example, the process, method, system, product or the equipment that contain series of steps or unit are not necessarily limited to
Those steps or unit clearly listed, but may include not list clearly or for these processes, method, product
Or other intrinsic steps of equipment or unit.
Embodiment 1
According to embodiments of the present invention, there is provided a kind of embodiment of information recommendation method, it is necessary to explanation, in accompanying drawing
The step of flow is illustrated can perform in the such as one group computer system of computer executable instructions, and, although
Logical order is shown in flow chart, but in some cases, can be performing shown different from order herein or retouch
The step of stating.
Fig. 1 is the schematic flow sheet of a kind of optional information recommendation method according to embodiments of the present invention, as shown in figure 1,
The method comprises the following steps：
First attribute of step S102, the collection user behavioral data produced at access target website and targeted website
Data, wherein, behavioral data includes the first current behavior data and the first historical behavior data；
First historical behavior data and the first attribute data are respectively processed and obtain the second historical behavior by step S104
Data and the second attribute data, and then according to the second historical behavior data and the second attribute data generation User Information Database,
Wherein, comprising multiple data samples in User Information Database；
Presupposed information recommended models are trained by step S106 according to data sample, obtain first information recommended models,
Wherein, first information recommended models are the presupposed information recommended models after convergence；
First current behavior data are processed, so as to generate target by step S108 according to first information recommended models
The recommendation information of website.
In embodiments of the present invention, using the collection user behavioral data produced at access target website and target network
The mode of the first attribute data stood, wherein, behavioral data includes the first current behavior data and the first historical behavior data, leads to
Cross to be respectively processed the first historical behavior data and the first attribute data and obtain the second historical behavior data and the second attribute
Data, and then according to the second historical behavior data and the second attribute data generation User Information Database, wherein, user profile number
According to, comprising multiple data samples, and then being trained to presupposed information recommended models according to data sample in storehouse, the first letter is obtained
Breath recommended models, wherein, first information recommended models are the presupposed information recommended models after convergence, have been reached according to the first information
Recommended models are processed so as to generate the purpose of the recommendation information of targeted website the first current behavior data, it is achieved thereby that
The degree of accuracy and efficiency, the technique effect of lifting Consumer's Experience that the network information is recommended are improved, and then is solved and is deposited in the prior art
The network information recommend the degree of accuracy and less efficient technical problem.
Alternatively, during step S102 is performed, collection user behavior information can be recorded by Web service
The merchandise news of data and different field, the modes such as collection, packet sniffer are embedded in by related data information using Javascript
Collection is pushed in message queue kafka.Wherein, Web service are a platform independence, lower coupling, self-contained,
Based on the application program of programmable Web, open XML (a subset under standard generalized markup language) standard can be used
These application programs are described, issue, find, coordinate and configure, the application program for developing distributed interoperability.Kafka
It is a kind of distributed post subscription message system of high-throughput, it can process the everything in the website of consumer's scale
Flow data.For example, the first attribute data of targeted website can include the attribute data of the different field involved by the website, should
Field can be books, food, film etc..
Alternatively, User Information Database can include the behavioural information of user, it is also possible to the target access including user
The merchandise news of the aspects such as books, food, film associated by website.
Alternatively, the presupposed information recommended models can be the user profile depth network model DNN based on deep learning
(deep-neural-network), or multiple cross-cutting goods model DNN.The plurality of cross-cutting goods model DNN can include M
Individual model, M >=3.Preferably, M=3, respectively book commodity model, video goods model, film goods model.
Alternatively, produced behavioral data, the first history when the first current behavior data are this access website of user
Behavioral data accessed behavioral data produced during website for user before this.
Alternatively, recommendation information can be pushed to user as recommendation results, it is also possible to be stored in the service of targeted website
In device.
Alternatively, Fig. 2 is the schematic flow sheet of the optional information recommendation method of another kind according to embodiments of the present invention, such as
Shown in Fig. 2, step S104 is performed, i.e., the first historical behavior data and the first attribute data process and obtain the second history row
It is that data and the second attribute data include：
First historical behavior data and the first attribute data are carried out data cleansing by step S202, obtain the 3rd history row
It is data and the 3rd attribute data；
3rd historical behavior data and the 3rd attribute data are carried out data prediction by step S204, obtain the second history
Behavioral data and the second attribute data, wherein, the mode of data prediction at least includes one of the following：Data pick-up, data turn
Change and loaded with data.
Alternatively, during step S202 is performed, the above-mentioned related data being collected into can be carried out based on Spark
Cleaning.Wherein, data cleansing refers to wrong last one program for finding and correcting and be can recognize that in data file, including is checked
Data consistency, treatment invalid value and missing values etc..Different from questionnaire examination ＆ verification, the data scrubbing after typing is usually by computer
Without being manually performed.
Alternatively, during step S204 is performed, the data cleaned can be carried out ETL (extract, conversion, plus
Carry) operate and be persisted to HDFS (Hadoop Distributed File System, distributed file system) or
MongoDB (database based on distributed document storage, write by C Plus Plus).Wherein, ETL (Extract-
Transform-Load it is) for describing data from source terminal by extracting (extract), conversion (transform), loading
(load) to destination process.
Alternatively, Fig. 3 is the schematic flow sheet of another optional information recommendation method according to embodiments of the present invention, such as
Shown in Fig. 3, step S106 is performed, i.e., presupposed information recommended models are trained according to data sample, obtained the first information and push away
Recommending model includes：
Step S302, is processed data sample according to default training algorithm, obtains user characteristics vector, wherein, in advance
If training algorithm at least includes one of the following：Collaborative filtering, data parallel algorithm and model based on matrix decomposition are parallel
Algorithm；
Step S304, first information recommended models are obtained according to user characteristics vector.
Alternatively, during performing above-mentioned steps S302, when processing data sample, with based on matrix point
The collaborative filtering ALS of solution carries out DNN model trainings, is instructed parallel in Spark using data parallel and model parallel mechanism
Practice DNN models, training for promotion efficiency obtains the user characteristics vector of abstract extraction.
Alternatively, Fig. 4 is the schematic flow sheet of another optional information recommendation method according to embodiments of the present invention, such as
Shown in Fig. 4, the method can also include：
First current behavior data are processed by step S402, obtain the second current behavior data；
Step S404, User Information Database is updated according to the second current behavior data；
Step S406, according to the data sample included in the User Information Database after renewal to first information recommended models
It is trained, obtains the second information recommendation model, wherein, the second information recommendation model is that the first information after convergence recommends mould
Type.
Alternatively, step S402 is performed to step S406, can carry out real-time update and optimization, example to information recommendation model
Such as, can by Spark steaming in real time from message queue obtain user behavior information, and with persistence
Original user information database mixes, and retraining is carried out to depth network model DNN, and the DNN that real-time update is optimized recommends
Model.
Alternatively, for a user, the renewal of information recommendation model and optimization can lift pushing away for information recommendation model
The degree of accuracy is recommended, it is necessary to explanation, the first current data produced by this access target website of user can be used for updating to be used
Family information database, and then for the optimization to current information recommended models, so that user is in next access target website
When, experience sense is more preferably.
Alternatively, before being trained to presupposed information recommended models according to data sample, the method can also include：
Step S10, the model parameter to presupposed information recommended models carries out initialization process.
It is alternatively possible to it is initial that presupposed information recommended models are carried out with parameter successively using limited Boltzmann machine RBM
Change.
Alternatively, for the defect and deficiency of prior art, big data combination Spark depths are based on this application provides one kind
Degree study and the intelligent recommendation method and system of real-time stream calculation, with reference to big data, Spark deep learnings, real-time stream calculation skill
Art, by collecting the behavioural information such as interest, preference of user and the merchandise news of different field, builds cross-cutting depth network mould
Type, the information for effectively integrating different field avoids the cold start-up problem of new user, while by gathering the real-time behavioural information of user,
Dynamic updates user behavior data storehouse, the renewal optimization of implementation model, the precision and validity of lift scheme, so as to provide most
Good recommendation service, lifts Consumer's Experience.
In embodiments of the present invention, using the collection user behavioral data produced at access target website and target network
The mode of the first attribute data stood, wherein, behavioral data includes the first current behavior data and the first historical behavior data, leads to
Cross to be respectively processed the first historical behavior data and the first attribute data and obtain the second historical behavior data and the second attribute
Data, and then according to the second historical behavior data and the second attribute data generation User Information Database, wherein, user profile number
According to, comprising multiple data samples, and then being trained to presupposed information recommended models according to data sample in storehouse, the first letter is obtained
Breath recommended models, wherein, first information recommended models are the presupposed information recommended models after convergence, have been reached according to the first information
Recommended models are processed so as to generate the purpose of the recommendation information of targeted website the first current behavior data, it is achieved thereby that
The degree of accuracy and efficiency, the technique effect of lifting Consumer's Experience that the network information is recommended are improved, and then is solved and is deposited in the prior art
The network information recommend the degree of accuracy and less efficient technical problem.
Embodiment 2
Other side according to embodiments of the present invention, additionally provides a kind of number-plate number identifying device, as shown in figure 5,
The information recommending apparatus include：Collecting unit 501, first processing units 503, second processing unit 505, the 3rd processing unit
507。
Wherein, collecting unit 501, for gathering user's behavioral data produced at access target website and target network
The first attribute data stood, wherein, behavioral data includes the first current behavior data and the first historical behavior data；First treatment
Unit 503, the second historical behavior data are obtained for being respectively processed to the first historical behavior data and the first attribute data
With the second attribute data, and then according to the second historical behavior data and the second attribute data generation User Information Database, wherein,
Comprising multiple data samples in User Information Database；Second processing unit 505, for being pushed away to presupposed information according to data sample
Recommend model to be trained, obtain first information recommended models, wherein, first information recommended models are that the presupposed information after convergence is pushed away
Recommend model；3rd processing unit 507, for being processed the first current behavior data according to first information recommended models, from
And generate the recommendation information of targeted website.
Alternatively, first processing units 503 include：First treatment subelement, for the first historical behavior data and the
One attribute data carries out data cleansing, obtains the 3rd historical behavior data and the 3rd attribute data；Second processing subelement, is used for
Data prediction is carried out to the 3rd historical behavior data and the 3rd attribute data, the second historical behavior data and the second attribute is obtained
Data, wherein, the mode of data prediction at least includes one of the following：The loading of data pick-up, data conversion and data.
Alternatively, second processing unit 505 includes：3rd treatment subelement, for the default training algorithm of basis to data
Sample is processed, and obtains user characteristics vector, wherein, default training algorithm at least includes one of the following：Based on matrix decomposition
Collaborative filtering, data parallel algorithm and model parallel algorithm；Subelement is obtained, for being obtained according to user characteristics vector
First information recommended models.
Alternatively, the device also includes：Fourth processing unit, for processing the first current behavior data, obtains
Second current behavior data；Updating block, for updating User Information Database according to the second current behavior data；5th treatment
Unit, for being instructed to first information recommended models according to the data sample included in the User Information Database after renewal
Practice, obtain the second information recommendation model, wherein, the second information recommendation model is the first information recommended models after convergence.
Alternatively, the device can also include：6th processing unit, for the model parameter to presupposed information recommended models
Carry out initialization process.
In embodiments of the present invention, using the collection user behavioral data produced at access target website and target network
The mode of the first attribute data stood, wherein, behavioral data includes the first current behavior data and the first historical behavior data, leads to
Cross to be respectively processed the first historical behavior data and the first attribute data and obtain the second historical behavior data and the second attribute
Data, and then according to the second historical behavior data and the second attribute data generation User Information Database, wherein, user profile number
According to, comprising multiple data samples, and then being trained to presupposed information recommended models according to data sample in storehouse, the first letter is obtained
Breath recommended models, wherein, first information recommended models are the presupposed information recommended models after convergence, have been reached according to the first information
Recommended models are processed so as to generate the purpose of the recommendation information of targeted website the first current behavior data, it is achieved thereby that
The degree of accuracy and efficiency, the technique effect of lifting Consumer's Experience that the network information is recommended are improved, and then is solved and is deposited in the prior art
The network information recommend the degree of accuracy and less efficient technical problem.
Embodiment 3
Another aspect according to embodiments of the present invention, additionally provides a kind of storage medium, and above-mentioned storage medium includes depositing
The program of storage, wherein, equipment where above-mentioned storage medium is controlled when said procedure runs performs upper in the embodiment of the present application 1
The information recommendation method stated.
Another aspect according to embodiments of the present invention, additionally provides a kind of processor, and above-mentioned processor is used for operation program,
Wherein, the above-mentioned information recommendation method in the embodiment of the present application 1 is performed when said procedure runs.
In embodiments of the present invention, using the collection user behavioral data produced at access target website and target network
The mode of the first attribute data stood, wherein, behavioral data includes the first current behavior data and the first historical behavior data, leads to
Cross to be respectively processed the first historical behavior data and the first attribute data and obtain the second historical behavior data and the second attribute
Data, and then according to the second historical behavior data and the second attribute data generation User Information Database, wherein, user profile number
According to, comprising multiple data samples, and then being trained to presupposed information recommended models according to data sample in storehouse, the first letter is obtained
Breath recommended models, wherein, first information recommended models are the presupposed information recommended models after convergence, have been reached according to the first information
Recommended models are processed so as to generate the purpose of the recommendation information of targeted website the first current behavior data, it is achieved thereby that
The degree of accuracy and efficiency, the technique effect of lifting Consumer's Experience that the network information is recommended are improved, and then is solved and is deposited in the prior art
The network information recommend the degree of accuracy and less efficient technical problem.
The embodiments of the present invention are for illustration only, and the quality of embodiment is not represented.
In the above embodiment of the present invention, the description to each embodiment all emphasizes particularly on different fields, and does not have in certain embodiment
The part of detailed description, may refer to the associated description of other embodiment.
In several embodiments provided herein, it should be understood that disclosed technology contents, can be by other
Mode is realized.Wherein, device embodiment described above is only schematical, such as division of described unit, Ke Yiwei
A kind of division of logic function, can there is other dividing mode when actually realizing, such as multiple units or component can combine or
Person is desirably integrated into another system, or some features can be ignored, or does not perform.Another, shown or discussed is mutual
Between coupling or direct-coupling or communication connection can be the INDIRECT COUPLING or communication link of unit or module by some interfaces
Connect, can be electrical or other forms.
The unit that is illustrated as separating component can be or may not be it is physically separate, it is aobvious as unit
The part for showing can be or may not be physical location, you can with positioned at a place, or can also be distributed to multiple
On unit.Some or all of unit therein can be according to the actual needs selected to realize the purpose of this embodiment scheme.
In addition, during each functional unit in each embodiment of the invention can be integrated in a processing unit, it is also possible to
It is that unit is individually physically present, it is also possible to which two or more units are integrated in a unit.Above-mentioned integrated list
Unit can both be realized in the form of hardware, it would however also be possible to employ the form of SFU software functional unit is realized.
If the integrated unit is to realize in the form of SFU software functional unit and as independent production marketing or use
When, can store in a computer read/write memory medium.Based on such understanding, technical scheme is substantially
The part for being contributed to prior art in other words or all or part of the technical scheme can be in the form of software products
Embody, the computer software product is stored in a storage medium, including some instructions are used to so that a computer
Equipment (can be personal computer, server or network equipment etc.) perform each embodiment methods described of the invention whole or
Part steps.And foregoing storage medium includes：USB flash disk, read-only storage (ROM, Read-Only Memory), arbitrary access are deposited
Reservoir (RAM, Random Access Memory), mobile hard disk, magnetic disc or CD etc. are various can be with store program codes
Medium.
The above is only the preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art
For member, under the premise without departing from the principles of the invention, some improvements and modifications can also be made, these improvements and modifications also should
It is considered as protection scope of the present invention.
Claims (10)
1. a kind of information recommendation method, it is characterised in that including：
The collection user behavioral data produced at access target website and the first attribute data of the targeted website, its
In, the behavioral data includes the first current behavior data and the first historical behavior data；
The first historical behavior data and first attribute data are respectively processed and obtain the second historical behavior data
With the second attribute data, and then according to the second historical behavior data and second attribute data generation user profile data
Storehouse, wherein, comprising multiple data samples in the User Information Database；
Presupposed information recommended models are trained according to the data sample, obtain first information recommended models, wherein, it is described
First information recommended models are the presupposed information recommended models after convergence；
The first current behavior data are processed according to the first information recommended models, so as to generate the target network
The recommendation information stood.
2. method according to claim 1, it is characterised in that described to the first historical behavior data and described first
Attribute data process and obtains the second historical behavior data and the second attribute data and include：
Data cleansing is carried out to the first historical behavior data and first attribute data, the 3rd historical behavior data are obtained
With the 3rd attribute data；
Data prediction is carried out to the 3rd historical behavior data and the 3rd attribute data, the second history row is obtained
It is data and second attribute data, wherein, the mode of the data prediction at least includes one of the following：Data pick-up,
Data conversion and data are loaded.
3. method according to claim 1, it is characterised in that described that mould is recommended to presupposed information according to the data sample
Type is trained, and obtaining first information recommended models includes：
The data sample is processed according to default training algorithm, obtains user characteristics vector, wherein, the default training
Algorithm at least includes one of the following：Collaborative filtering, data parallel algorithm and model parallel algorithm based on matrix decomposition；
The first information recommended models are obtained according to the user characteristics vector.
4. method according to claim 1, it is characterised in that methods described also includes：
The first current behavior data are processed, the second current behavior data are obtained；
The User Information Database is updated according to the second current behavior data；
The data sample included in the User Information Database after according to renewal is to the first information recommended models
It is trained, obtains the second information recommendation model, wherein, the second information recommendation model is the first information after convergence
Recommended models.
5. method according to claim 1, it is characterised in that recommended presupposed information according to the data sample described
Before model is trained, methods described also includes：Model parameter to the presupposed information recommended models is carried out at initialization
Reason.
6. a kind of information recommending apparatus, it is characterised in that including：
Collecting unit, for gathering user's behavioral data produced at access target website and the targeted website first
Attribute data, wherein, the behavioral data includes the first current behavior data and the first historical behavior data；
First processing units, for the first historical behavior data and first attribute data to be respectively processed and obtained
Second historical behavior data and the second attribute data, and then according to the second historical behavior data and second attribute data
Generation User Information Database, wherein, comprising multiple data samples in the User Information Database；
Second processing unit, for being trained to presupposed information recommended models according to the data sample, obtains the first information
Recommended models, wherein, the first information recommended models are the presupposed information recommended models after convergence；
3rd processing unit, for being processed the first current behavior data according to the first information recommended models,
So as to generate the recommendation information of the targeted website.
7. device according to claim 6, it is characterised in that the first processing units include：
First treatment subelement, for carrying out data cleansing to the first historical behavior data and first attribute data,
Obtain the 3rd historical behavior data and the 3rd attribute data；
Second processing subelement, for the 3rd historical behavior data and the 3rd attribute data to be carried out data and located in advance
Reason, obtains the second historical behavior data and second attribute data, wherein, the mode of the data prediction is at least wrapped
Include one of the following：The loading of data pick-up, data conversion and data.
8. device according to claim 6, it is characterised in that the second processing unit includes：
3rd treatment subelement, for being processed the data sample according to default training algorithm, obtain user characteristics to
Amount, wherein, the default training algorithm at least includes one of the following：Collaborative filtering, data parallel based on matrix decomposition
Algorithm and model parallel algorithm；
Subelement is obtained, for obtaining the first information recommended models according to the user characteristics vector.
9. a kind of storage medium, it is characterised in that the storage medium includes the program of storage, wherein, in described program operation
When control in equipment perform claim requirement 1 to claim 5 where the storage medium described in any one information recommendation side
Method.
10. a kind of processor, it is characterised in that the processor is used for operation program, wherein, right of execution when described program is run
Profit requires the information recommendation method described in any one in 1 to claim 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710198948.6A CN106874522A (en) | 2017-03-29 | 2017-03-29 | Information recommendation method, device, storage medium and processor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710198948.6A CN106874522A (en) | 2017-03-29 | 2017-03-29 | Information recommendation method, device, storage medium and processor |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106874522A true CN106874522A (en) | 2017-06-20 |
Family
ID=59159824
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710198948.6A Pending CN106874522A (en) | 2017-03-29 | 2017-03-29 | Information recommendation method, device, storage medium and processor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106874522A (en) |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107734594A (en) * | 2017-09-22 | 2018-02-23 | 平安科技（深圳）有限公司 | Personalized wifi hotspot method for pushing, device and storage medium |
| CN108304441A (en) * | 2017-11-14 | 2018-07-20 | 腾讯科技（深圳）有限公司 | Network resource recommended method, device, electronic equipment, server and storage medium |
| CN108334536A (en) * | 2017-11-30 | 2018-07-27 | 中国电子科技集团公司电子科学研究院 | A kind of information recommendation method, equipment and storage medium |
| CN108765076A (en) * | 2018-05-24 | 2018-11-06 | 蜜芽宝贝（北京）网络科技有限公司 | Mother and baby's content recommendation method, device and readable storage medium storing program for executing |
| CN108830689A (en) * | 2018-06-13 | 2018-11-16 | 深圳乐信软件技术有限公司 | Item recommendation method, device, server and storage medium |
| CN108989397A (en) * | 2018-06-26 | 2018-12-11 | 腾讯音乐娱乐科技（深圳）有限公司 | Data recommendation method, device and storage medium |
| CN109447813A (en) * | 2018-09-05 | 2019-03-08 | 平安科技（深圳）有限公司 | Products Show method, apparatus, computer equipment and storage medium |
| CN109816412A (en) * | 2017-11-21 | 2019-05-28 | 腾讯科技（深圳）有限公司 | A kind of training pattern generation method, device, equipment and computer storage medium |
| CN110032684A (en) * | 2019-04-22 | 2019-07-19 | 山东大学 | The cross-domain Parallel Sequence recommended method of information, medium and equipment based on shared account |
| CN110222894A (en) * | 2019-06-06 | 2019-09-10 | 阿里巴巴集团控股有限公司 | Advertisement placement method, device and equipment |
| CN110276018A (en) * | 2019-05-29 | 2019-09-24 | 深圳技术大学 | Personalized recommendation method, terminal and storage medium for online education system |
| WO2020037917A1 (en) * | 2018-08-22 | 2020-02-27 | 平安科技（深圳）有限公司 | User behavior data recommendation method, server and computer readable medium |
| CN111881355A (en) * | 2020-07-28 | 2020-11-03 | 北京深演智能科技股份有限公司 | Object recommendation method and device, storage medium and processor |
| CN112100558A (en) * | 2020-09-02 | 2020-12-18 | 北京字节跳动网络技术有限公司 | Method, device, equipment and storage medium for object recommendation |
| CN112364241A (en) * | 2020-10-27 | 2021-02-12 | 北京五八信息技术有限公司 | Information pushing method and device, electronic equipment and storage medium |
| CN113536140A (en) * | 2021-09-16 | 2021-10-22 | 广东众聚人工智能科技有限公司 | Real-time recommendation method and system based on small sample data |
| CN113961792A (en) * | 2020-07-15 | 2022-01-21 | 北京达佳互联信息技术有限公司 | Resource recommendation system, method, electronic device and storage medium |
| CN115374370A (en) * | 2022-10-26 | 2022-11-22 | 小米汽车科技有限公司 | Content pushing method and device based on multiple models and electronic equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105469263A (en) * | 2014-09-24 | 2016-04-06 | 阿里巴巴集团控股有限公司 | Commodity recommendation method and device |
| US20160259857A1 (en) * | 2015-03-06 | 2016-09-08 | Microsoft Technology Licensing, Llc | User recommendation using a multi-view deep learning framework |
| CN106126632A (en) * | 2016-06-22 | 2016-11-16 | 北京小米移动软件有限公司 | Recommend method and device |
| CN106485562A (en) * | 2015-09-01 | 2017-03-08 | 苏宁云商集团股份有限公司 | A kind of commodity information recommendation method based on user's history behavior and system |
-
2017
- 2017-03-29 CN CN201710198948.6A patent/CN106874522A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105469263A (en) * | 2014-09-24 | 2016-04-06 | 阿里巴巴集团控股有限公司 | Commodity recommendation method and device |
| US20160259857A1 (en) * | 2015-03-06 | 2016-09-08 | Microsoft Technology Licensing, Llc | User recommendation using a multi-view deep learning framework |
| CN106485562A (en) * | 2015-09-01 | 2017-03-08 | 苏宁云商集团股份有限公司 | A kind of commodity information recommendation method based on user's history behavior and system |
| CN106126632A (en) * | 2016-06-22 | 2016-11-16 | 北京小米移动软件有限公司 | Recommend method and device |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107734594A (en) * | 2017-09-22 | 2018-02-23 | 平安科技（深圳）有限公司 | Personalized wifi hotspot method for pushing, device and storage medium |
| CN107734594B (en) * | 2017-09-22 | 2020-12-04 | 平安科技（深圳）有限公司 | Personalized wifi hotspot pushing method and device and storage medium |
| CN108304441A (en) * | 2017-11-14 | 2018-07-20 | 腾讯科技（深圳）有限公司 | Network resource recommended method, device, electronic equipment, server and storage medium |
| CN108304441B (en) * | 2017-11-14 | 2022-04-26 | 腾讯科技（深圳）有限公司 | Network resource recommendation method and device, electronic equipment, server and storage medium |
| CN109816412A (en) * | 2017-11-21 | 2019-05-28 | 腾讯科技（深圳）有限公司 | A kind of training pattern generation method, device, equipment and computer storage medium |
| CN108334536B (en) * | 2017-11-30 | 2023-10-24 | 中国电子科技集团公司电子科学研究院 | Information recommendation method, device and storage medium |
| CN108334536A (en) * | 2017-11-30 | 2018-07-27 | 中国电子科技集团公司电子科学研究院 | A kind of information recommendation method, equipment and storage medium |
| CN108765076A (en) * | 2018-05-24 | 2018-11-06 | 蜜芽宝贝（北京）网络科技有限公司 | Mother and baby's content recommendation method, device and readable storage medium storing program for executing |
| CN108830689A (en) * | 2018-06-13 | 2018-11-16 | 深圳乐信软件技术有限公司 | Item recommendation method, device, server and storage medium |
| CN108989397A (en) * | 2018-06-26 | 2018-12-11 | 腾讯音乐娱乐科技（深圳）有限公司 | Data recommendation method, device and storage medium |
| CN108989397B (en) * | 2018-06-26 | 2021-04-20 | 腾讯音乐娱乐科技（深圳）有限公司 | Data recommendation method and device and storage medium |
| WO2020037917A1 (en) * | 2018-08-22 | 2020-02-27 | 平安科技（深圳）有限公司 | User behavior data recommendation method, server and computer readable medium |
| CN109447813A (en) * | 2018-09-05 | 2019-03-08 | 平安科技（深圳）有限公司 | Products Show method, apparatus, computer equipment and storage medium |
| CN110032684A (en) * | 2019-04-22 | 2019-07-19 | 山东大学 | The cross-domain Parallel Sequence recommended method of information, medium and equipment based on shared account |
| CN110276018A (en) * | 2019-05-29 | 2019-09-24 | 深圳技术大学 | Personalized recommendation method, terminal and storage medium for online education system |
| CN110222894A (en) * | 2019-06-06 | 2019-09-10 | 阿里巴巴集团控股有限公司 | Advertisement placement method, device and equipment |
| CN113961792A (en) * | 2020-07-15 | 2022-01-21 | 北京达佳互联信息技术有限公司 | Resource recommendation system, method, electronic device and storage medium |
| CN111881355B (en) * | 2020-07-28 | 2023-03-10 | 北京深演智能科技股份有限公司 | Object recommendation method and device, storage medium and processor |
| CN111881355A (en) * | 2020-07-28 | 2020-11-03 | 北京深演智能科技股份有限公司 | Object recommendation method and device, storage medium and processor |
| CN112100558A (en) * | 2020-09-02 | 2020-12-18 | 北京字节跳动网络技术有限公司 | Method, device, equipment and storage medium for object recommendation |
| CN112364241A (en) * | 2020-10-27 | 2021-02-12 | 北京五八信息技术有限公司 | Information pushing method and device, electronic equipment and storage medium |
| CN113536140A (en) * | 2021-09-16 | 2021-10-22 | 广东众聚人工智能科技有限公司 | Real-time recommendation method and system based on small sample data |
| CN113536140B (en) * | 2021-09-16 | 2022-02-11 | 广东众聚人工智能科技有限公司 | Real-time recommendation method and system based on small sample data |
| CN115374370A (en) * | 2022-10-26 | 2022-11-22 | 小米汽车科技有限公司 | Content pushing method and device based on multiple models and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106874522A (en) | Information recommendation method, device, storage medium and processor | |
| Mori et al. | β-diversity, community assembly, and ecosystem functioning | |
| CN110717098B (en) | Meta-path-based context-aware user modeling method and sequence recommendation method | |
| CN108595494A (en) | The acquisition methods and device of reply message | |
| CN104142940B (en) | Information recommendation processing method and processing device | |
| CN110442790A (en) | Recommend method, apparatus, server and the storage medium of multi-medium data | |
| Maniya et al. | A multi-attribute selection of automated guided vehicle using the AHP/M-GRA technique | |
| Lima et al. | Domain knowledge integration in data mining using decision tables: case studies in churn prediction | |
| Sebastián‐González | Drivers of species' role in avian seed‐dispersal mutualistic networks | |
| CN108665064A (en) | Neural network model training, object recommendation method and device | |
| CN105095219B (en) | Micro-blog recommendation method and terminal | |
| CN106570008A (en) | Recommendation method and device | |
| CN110175895B (en) | Article recommendation method and device | |
| CN106940801A (en) | A kind of deeply for Wide Area Network learns commending system and method | |
| CN106951489A (en) | A kind of personalized recommendation method and device for sparse big data | |
| CN108717654B (en) | Multi-provider cross recommendation method based on clustering feature migration | |
| CN114036406A (en) | Recommendation method and system based on graph contrast learning and social network enhancement | |
| CN107239993A (en) | A kind of matrix decomposition recommendation method and system based on expansion label | |
| EP4202725A1 (en) | Joint personalized search and recommendation with hypergraph convolutional networks | |
| CN103995866A (en) | Commodity information pushing method and device based on link forecasting | |
| CN109903127A (en) | Group recommendation method and device, storage medium and server | |
| CN112149352B (en) | Prediction method for marketing activity clicking by combining GBDT automatic characteristic engineering | |
| CN107864405A (en) | A kind of Forecasting Methodology, device and the computer-readable medium of viewing behavior type | |
| CN110717089A (en) | User behavior analysis system and method based on weblog | |
| CN111259267B (en) | Distributed hybrid collaborative intelligent recommendation method based on sparsity perception |
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 |
Application publication date: 20170620 |
|
| RJ01 | Rejection of invention patent application after publication |