CN108228642A - A kind of project recommendation and device based on memristor - Google Patents
A kind of project recommendation and device based on memristor Download PDFInfo
- Publication number
- CN108228642A CN108228642A CN201611192441.1A CN201611192441A CN108228642A CN 108228642 A CN108228642 A CN 108228642A CN 201611192441 A CN201611192441 A CN 201611192441A CN 108228642 A CN108228642 A CN 108228642A
- Authority
- CN
- China
- Prior art keywords
- recommendation
- value set
- project
- preference value
- user
- 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.)
- Granted
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
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the invention discloses a kind of item recommendation method and device based on memristor, for making for user's recommended project.Present invention method includes:Preference value set is obtained, the preference value set includes multiple preference values, wherein, a preference value in the multiple preference value is a user to the preference value of a project;The preference value set is stored in a storage module, the memory module includes multiple memristors, wherein, there is always a memristors in the multiple memristor to store any one of preference value set mesh preference value;It is determined as one or more destination items of target user's recommendation according to the preference value set.
Description
Technical field
The present invention relates to the communications field more particularly to a kind of project recommendations and device based on memristor.
Background technology
In internet, the preference information of user is collected in commending system application, to user's recommended products or service, extensively should
For daily life, such as the recommendation of the various aspects such as shopping, film, books.It makes businessman precisely determine the demand of user, subtracts
Few meaningless advertisement expenditure, user can be also received to oneself valuable commodity or information on services.
Common commending system has the recommendation of search-engine results and the recommendation of shopping website.The recommendation of search-engine results
Very more search results can be obtained by being embodied in when user searches for information, and commending system can select to best suit user demand
As a result, before being presented in several result options.Particularly, different search can be presented in commending system according to the characteristics of user
Hitch fruit.The recommendation of shopping website is specially according to the information such as the browsing record of user, personal like, age, occupation, shopping network
Standing can targetedly Recommendations.Instantly the commending system based on software is generally used, its realization is needed on hardware
Large-scale computing cluster needs responsible distributed software frame on software, so as to fulfill a set of distributed computing environment.
However, since the commending system based on software needs to store a large amount of information and a large amount of calculates so that its into
This height, speed are slow, volume is big, power consumption is big, difficult in maintenance.
Invention content
An embodiment of the present invention provides a kind of item recommendation method and device based on memristor, for making to push away for user
Recommend project.
First aspect of the embodiment of the present invention provides a kind of item recommendation method based on memristor, including:
Preference value set is obtained, which includes multiple preference values, wherein, one in multiple preference value is partially
Good value is preference value of the user to a project, can pass through the preference value to disparity items that user is obtained on internet
Data, can also acquisition be interacted by the information with terminal, can also by server interaction obtain, do not limit herein
It is fixed.The preference value set is stored in a storage module, which includes multiple memristors, wherein, in multiple memristor
In there is always a memristors to store any one of preference value set mesh preference value, wherein it is possible to be used for a memristor
In one preference value of storage.There is always a memristors in multiple memristor to store any one of preference value set mesh
Preference value, and multiple memristor may be constructed a memristor array, due to output preference value set.Specifically, memristor
Different electric conductivity values correspond to different preference values, wherein corresponding to a user per a line, each row correspond to a kind of project.
It is determined as one or more destination items of target user's recommendation according to the preference value set.
Due to realizing such commending system using hardware, it can reduce volume while advisory speed is improved, reduce
Power consumption simplifies and safeguards.
With reference to first aspect present invention, the first embodiment of first aspect present invention, including:
Determine candidate recommendation project set, which includes multiple candidate recommendation projects;According to this partially
Good data acquisition system calculates the target user respectively to the recommendation of each project in the candidate recommendation project, obtains recommendation collection
It closes;It is determined as the one or more destination item of target user recommendation according to the recommendation value set.
The target user can be calculated according to the preference data set respectively to each project in the candidate recommendation project
Recommendation, obtain recommend value set, and using the recommendation value set as determine whether recommendation one project foundation.
With reference to the first embodiment of first aspect present invention, second of embodiment of first aspect present invention, packet
It includes:
Determine that one or more, to the preference value of each project in the candidate recommendation project, obtains candidate item with reference to user
Mesh preference value set, the candidate items preference value set are contained in the preference value set, and the project of each candidate can be if not
Same books, film or dining room.The one or more is calculated with reference to each in user according to the candidate items preference value set
With reference to user respectively with the hobby deviation of the target user, hobby deviation value set is obtained, can be made between two users
Calculate hobby deviation of two users for same project with computing unit, the computing unit include absolute value circuit and
Memristor.Wherein, the absolute value circuit is for calculating the absolute value of the deviation between two users, and memristor is used to store
The hobby deviation.The target user is calculated to the candidate recommendation project according to the preference value set and the hobby deviation value set
In each project recommendation.
In some feasible embodiments, after candidate items preference value set and hobby deviation value set is obtained,
Can the larger user of some hobby deviations be screened out according to hobby deviation value set, be left some close, roots of hobby
Determine to calculate the recommendation of each project to the preference value of candidate items according to the close user of these hobbies.
With reference to the first embodiment of first aspect present invention, the third embodiment of first aspect present invention, packet
It includes:
It determines to be higher than in the recommendation value set one or more projects corresponding to the recommendation of threshold value, is this or more
A destination item.
With reference to the first embodiment of first aspect present invention, the 4th kind of embodiment of first aspect present invention, packet
It includes:
The highest one or more projects of recommendation in the recommendation value set are determined, for the one or more target item
Mesh.
Second aspect of the embodiment of the present invention provides a kind of project recommendation device based on memristor, including:
Acquisition module, for obtaining preference value set, which includes multiple preference values, wherein, it is multiple inclined
A preference value in good value is a user to the preference value of a project;Memory module, including multiple memristors, for depositing
The preference value set is stored up, wherein, there is always a memristors in multiple memristor to store any in the preference value set
Project preference value;Recommending module, for being determined as one or more target items of target user's recommendation according to the preference value set
Mesh.
With reference to second aspect of the present invention, the first embodiment of second aspect of the present invention, including:
Determination sub-module, for determining candidate recommendation project set, which pushes away including multiple candidates
Recommend project;Computational submodule, for calculating the target user respectively in the candidate recommendation project according to the preference data set
Each project recommendation, obtain recommend value set;Recommend submodule, for being determined as the target according to the recommendation value set
The one or more destination item that user recommends.
With reference to the first embodiment of second aspect of the present invention, second of embodiment of second aspect of the present invention, packet
It includes:
First determination unit, for determining one or more with reference to user to each project in the candidate recommendation project
Preference value obtains candidate items preference value set, which is contained in the preference value set;First calculates
Unit, for calculating the one or more with reference to each with reference to user's difference in user according to the candidate items preference value set
With the hobby deviation of the target user, hobby deviation value set is obtained;Second computing unit, for according to the preference value set
Recommendation of the target user to each project in the candidate recommendation project is calculated with the hobby deviation value set.
With reference to the first embodiment of second aspect of the present invention, the third embodiment of second aspect of the present invention, packet
It includes:
Second determination unit, for determining to be higher than in the recommendation value set one or more corresponding to the recommendation of threshold value
Project, for the one or more destination item.
With reference to the first embodiment of second aspect of the present invention, the 4th kind of embodiment of second aspect of the present invention, packet
It includes:
Third determination subelement for determining the highest one or more projects of the recommendation in the recommendation value set, is
The one or more destination item.
The third aspect of the embodiment of the present invention provides a kind of project recommendation device based on memristor, including:
Bus, transceiver, processor and memory;The bus is used to connect with the transceiver, the processor and the memory
It connects;The transceiver, for obtaining preference value set, which includes multiple preference values, wherein, in multiple preference value
A preference value be user to the preference value of a project;The memory includes multiple memristors, for storing this partially
Good value set, wherein, it is inclined in multiple memristor there is always a memristor to store any one of preference value set mesh
Good value;The processor, for being determined as one or more destination items of target user's recommendation according to the preference value set.
As can be seen from the above technical solutions, the embodiment of the present invention has the following advantages:
By obtaining preference value set, which includes multiple preference values, wherein, one in multiple preference value
A preference value for a user to the preference value of a project, store the preference value set, the memory module in a storage module
Including multiple memristors, wherein, there is always a memristors in multiple memristor to store any in the preference value set
Project preference value is determined as one or more destination items of target user's recommendation according to the preference value set, due to using firmly
Part realizes such commending system, can reduce volume, reduce power consumption, simplify and safeguard improve advisory speed while.
Description of the drawings
Fig. 1 is an a kind of configuration diagram of item recommendation system based on memristor in the embodiment of the present application;
Fig. 2 is a kind of one embodiment schematic diagram of the item recommendation method based on memristor in the embodiment of the present application;
Fig. 2-1 is the operation principle schematic diagram of memristor in the embodiment of the present application;
Fig. 2-2 is the schematic diagram of memristor array in the embodiment of the present application;
Fig. 2-3 is the schematic diagram of preference value set in the embodiment of the present application;
Fig. 2-4 is the schematic diagram of the computing unit comprising absolute value circuit in the embodiment of the present application;
Fig. 2-5 is the schematic diagram that recommendation method is calculated in the embodiment of the present application;
Fig. 3 is a kind of one embodiment schematic diagram of the project recommendation device based on memristor in the embodiment of the present application;
Fig. 4 is a kind of another embodiment schematic diagram of the project recommendation device based on memristor in the embodiment of the present application;
Fig. 5 is a kind of another embodiment schematic diagram of the project recommendation device based on memristor in the embodiment of the present application;
Fig. 6 is a kind of another embodiment schematic diagram of the project recommendation device based on memristor in the embodiment of the present application;
Fig. 7 is a kind of another embodiment schematic diagram of the project recommendation device based on memristor in the embodiment of the present application;
Fig. 8 is a kind of another embodiment schematic diagram of the project recommendation device based on memristor in the embodiment of the present application.
Specific embodiment
An embodiment of the present invention provides a kind of item recommendation method and device based on memristor, for making to push away for user
Recommend project.
In order to which those skilled in the art is made to more fully understand the embodiment of the present invention, implement below in conjunction with the present invention
Attached drawing in example, is clearly and completely described the technical solution in the embodiment of the present invention, it is clear that described embodiment
The only embodiment of a present invention part, instead of all the embodiments.
Term " first ", " second ", " third " in description and claims of this specification and above-mentioned attached drawing, "
The (if present)s such as four " are the objects for distinguishing similar, and specific sequence or precedence are described without being used for.It should manage
The data that solution uses in this way can be interchanged in the appropriate case, so that the embodiments described herein can be in addition to illustrating herein
Or the sequence other than the content of description is implemented.In addition, term " comprising " and " having " and their any deformation, it is intended that
Cover it is non-exclusive include, for example, containing the process of series of steps or unit, method, system, product or equipment need not limit
In those steps or unit for clearly listing, but may include not listing clearly or for these processes, method, production
The intrinsic other steps of product or equipment or unit.
It please refers to Fig.1, for the framework of the item recommendation system based on memristor, including project recommendation device and one or more
A terminal.
The present embodiments relate to terminal, can refer to the equipment for providing a user voice and/or data connectivity, have
There is the portable equipment of wireless connecting function or be connected to other processing equipments of radio modem.Wireless terminal can be with
It communicates through wireless access network (RAN Radio Access Network) with one or more core nets, wireless terminal can be with
It is mobile terminal, such as mobile phone (or for " honeycomb " phone) and the computer with mobile terminal, for example, it may be portable
Formula, pocket, hand-held, built-in computer or vehicle-mounted mobile device, they exchanged with wireless access network language and/or
Data.For example, personal communication service (PCS, Personal Communication Service) phone, wireless phone, session
Initiation protocol (SIP) phone, wireless local loop (WLL, Wireless Local Loop) are stood, personal digital assistant (PDA,
Personal Digital Assistant) etc. equipment.Wireless terminal is referred to as system, subscriber unit (Subscriber
Unit), subscriber station (Subscriber Station), movement station (Mobile Station), mobile station (Mobile), distant station
(Remote Station), access point (Access Point), remote terminal (Remote Terminal), access terminal
(Access Terminal), user terminal (User Terminal), terminal device, user agent (User Agent), user
Equipment (User Device) or user equipment (User Equipment).
By taking mobile phone as an example, the structure of mobile phone can include:It is radio frequency (Radio Frequency, RF) circuit, memory, defeated
Enter unit, display unit, sensor, voicefrequency circuit, Wireless Fidelity (wireless fidelity, WiFi) module, processor,
And the components such as power supply.It will be understood by those skilled in the art that Yi Shang structure does not form the restriction to mobile phone, can include than
It illustrates more or fewer components and either combines certain components or different components arrangement.
In some feasible embodiments, project recommendation device can be a server, or in server
One function module, or third-party device is not construed as limiting herein.
By taking server as an example, server can generate bigger difference due to configuration or different performance, can include one
Or more than one central processing unit (central processing units, CPU) (for example, one or more processors)
And memory, the storage medium of one or more storage application programs or data (such as deposit by one or more magnanimity
Store up equipment).Wherein, memory and storage medium can be of short duration storage or persistent storage.The program for being stored in storage medium can
To include one or more modules (diagram does not mark), each module can include grasping the series of instructions in server
Make.Further, central processing unit could be provided as communicating with storage medium, perform one in storage medium on the server
Series of instructions operates.Server can also include one or more power supplys, one or more wired or wireless networks
Interface, one or more input/output interfaces and/or, one or more operating systems, such as Windows
ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM etc..
In some feasible embodiments, project recommendation device and terminal technology can be handed over into row information by radio communication
Mutually, then the structure of terminal can include RF circuits and pass through RF circuits realization wireless communication.Specifically, RF circuits can be used for receiving
In photos and sending messages or communication process, signal sends and receivees, particularly, after the downlink information of base station is received, at processor
Reason;In addition, the data for designing uplink are sent to base station.In general, RF circuits include but not limited to antenna, at least one amplification
Device, transceiver, coupler, low-noise amplifier (Low Noise Amplifier, LNA), duplexer etc..In addition, RF circuits
It can also communicate by radio communication with network and other equipment.Above-mentioned wireless communication can use any communication standard or association
View, including but not limited to global system for mobile communications (Global System of Mobile communication, GSM) are led to
With grouping wireless service (General Packet Radio Service, GPRS), CDMA (Code Division
Multiple Access, CDMA), wideband code division multiple access (Wideband Code Division Multiple Access,
WCDMA), long term evolution (Long Term Evolution, LTE), Email, short message service (Short Messaging
Service, SMS) etc..
It please refers to Fig.2, includes for a kind of one embodiment of the recommendation method based on memristor in the embodiment of the present application, wrap
It includes:
201st, preference value set is obtained, which includes multiple preference values, wherein, one in multiple preference value
A preference value is a user to the preference value of a project.
In some feasible embodiments, the number for the preference value to disparity items that user is obtained on internet can be passed through
According to, can also acquisition be interacted by the information with terminal, can also by server interaction obtain, be not construed as limiting herein.
In the data of acquisition, scoring or preference value of the different terminals for different projects can be obtained, and as data
Recommend corresponding project or service for user in basis.
The method of common acquisition preference Value Data has the data of search-engine results and shopping website.Wherein, it is searching for
In the method for engine results, user can obtain very more search results when searching for information, and project recommendation device can select most
Meet user demand as a result, before being presented in several result options.Particularly, project recommendation device can be according to the spy of user
It puts that different search results is presented.In the data for obtaining shopping website, it can be recorded according to the browsing of user, personal happiness
The information such as good, age, occupation.
202nd, the preference value set is stored in a storage module, which includes multiple memristors, wherein, it is more at this
There is always a memristors in a memristor to store any one of preference value set mesh preference value.
In embodiments of the present invention, memory module can include multiple memristors, and multiple memristor is due to storing preference
Value set.
In some feasible embodiments, -1 is please referred to Fig.2, is the operation principle of memristor, is related to four electric appliance ginsengs
Number, is resistance, capacitance, inductance and memristor respectively.Memristor describes the relationship between magnetic flux and charge q, and this relationship is determined
Justice is d=Mdq, and wherein M represents memristor value, this is a variable with resistance with common physical dimension, is numerically equal to applied
It is added in the voltage at memristor both ends and flows through the ratio between electric current.
Wherein it is possible to it is used to store a preference value for a memristor.There is always one in multiple memristor to recall
Resistance device stores any one of preference value set mesh preference value, and multiple memristor may be constructed a memristor array,
Due to output preference value set.Specifically, the different electric conductivity values of memristor correspond to different preference values, wherein being corresponded to per a line
A user, each row correspond to a kind of project.
It should be noted that when there are user, to the preference value of project, memristor is just stored with data, if nothing, accordingly
The data of memristor can be without (none).Preference value of the corresponding user of memristor to project that then data are none,
It can be seen as needing to seek its recommendation.
In embodiments of the present invention, the preference value set stored by memristor can regard a storage matrix as, please refer to
Fig. 2-2 are the schematic diagram of memristor array, and wherein item1 is project 1, and item2 is project 2, and so on, likewise,
User1 is user 1, and user2 is user 2, and so on, it repeats no more.
For matrix manipulation, structure is similar can so that reconfigurable arrays are more efficient, also promote more to study and concentrate
In in corresponding circuit design, such as flash memory transistor array, however the computing capability of these designs and can expansion capability it is restricted
In the large scale of unit module tremendously, it is impossible to obtain being promoted.And the discovery of memristor element causes neuromorphic to calculate
The design of system produces revolutionary change, since memristor has memory characteristic, can remember the electric current passed through, it can be with
Cynapse in simulative neural network, while memristor cross array structure can greatly improve integration density in 2D/3D designs.
The connection matrix of neural network can be well realized in memristor crossed array, it can be provided in the size of very little
Large-scale signal connection, additionally it is possible to different weighed combinations is carried out to input signal.It is as shown below, N × M memristor intersecting maneuvers
Row are a basic modules, to realize the approximate calculation function of matrix-vector multiplication.One group of input voltage is applied to the N of array
In wordline (word-lines, WLs), then M bit line is flowed through to collect by measuring the voltage at fixed value resistance rs both ends
The electric current of (bit-lines, BLs).At the node that memristor is located at wordline and bit line intersects.Memristor in crossed array can lead to
It crosses and directly applies program voltage pulse to be programmed to it.Ensure to make by selecting the bias voltage of memristor array WLs and BLs
It is higher than programming thresholds with the voltage on target memristor or electric current.This scheme is realized simply, but resolution ratio is relatively low.Article is set
It has counted based on electric signal amplitude layered design method, but this method is not built upon in the behavior of intrinsic biological synapse, such as arteries and veins
Rush mechanism and space-time characterisation.
In some feasible embodiments, memristor array has the characteristics that following several:1st, memristor resistance value depends on stream
The quantity of electric charge through this device, therefore can realize the storage of analog quantity;2nd, the memory chip storage density ratio based on memristor
Currently based on chip height at least an order of magnitude of transistor, and as technology maturation gap can be widened further;3rd, storage speed
Degree 3 orders of magnitude higher than flash memory speed;4th, it is non-volatile, no refresh power consumption.
203rd, candidate recommendation project set is determined, which includes multiple candidate recommendation projects.
In some feasible embodiments, candidate recommendation project set can be first determined, the candidate recommendation project set packet
Include multiple candidate recommendation projects.In some feasible embodiments, which does not include the target user to multiple
The preference value of any one in candidate recommendation project because if including, then can use the preference value, but at other
In embodiment, perhaps the preference value is out-of-date data, after can also calculating its recommendation or calculating its recommendation, is used for
With preference value compare, make to the recommendation it is accurate whether assessment.In embodiments of the present invention, candidate recommendation project set
It can include the candidate recommendation project for same type of product or service, such as film, books or dining room, not limit herein
It is fixed, can also be different types of product or service, such as in a candidate recommendation project set, include film, books and
Dining room is not construed as limiting herein.
- 3 are please referred to Fig.2, is the schematic diagram of preference value set, as shown in the figure, it is assumed that userM is target user, the target
User possesses the preference value of project 1 (item1) and project 2 (item2), but without project 3 (item3) and project 4 (item4)
Preference value, then can be using setting item 3 and project 4 as a part for candidate items set.In other feasible embodiments
In, it can be herein not construed as limiting using setting item 1 and project 2 as a part for candidate items set.
204th, determine that one or more, to the preference value of each project in the candidate recommendation project, is waited with reference to user
Option preference value set, the candidate items preference value set are contained in the preference value set.
In some feasible embodiments, after candidate recommendation project is determined, it can obtain one or more with reference to use
Family to the preference value of candidate recommendation project as reference.- 3 are please referred to Fig.2, can be made with user 1- user 5 (user1-user5)
To refer to user, the preference value to candidate recommendation project with reference to user is obtained, it is assumed that candidate recommendation project is project 3 (item3)
With project 4 (item4), then user 1- user 5 (user1-user5) can be obtained respectively for project 3 (item3) and project 4
(item4) preference value, as shown in the figure, for for project 3 (item3):4th, 4,4,2,5, for project 4 (item4):2、1、2、
5、1。
205th, according to the preference value set calculate the one or more with reference in user it is each with reference to user respectively with this
The hobby deviation of target user obtains hobby deviation value set.
In some feasible embodiments, if to be user's recommended project, such as books, film or dining room, it can pass through
The following steps:1st, the crowd similar to the user, such as business associate, friend, classmate, age professional identical people are found
Deng, after collecting a large amount of information, classification find out similar crowd;2nd, the project that similar crowd uses or consumes is found out;It 3rd, will be similar
In book/film/dining room that crowd often reads, book/film/dining room of the user and the data that scoring is relatively low are got rid of, it then will be surplus
Under data by scoring height recommend user.
In some feasible embodiments, it is specific recommend method can also there are two types of method, 1, the collaboration based on user
Recommend;2nd, the collaborative filtering recommending based on article.
Wherein, the Collaborative Recommendation based on user can be that user A likes article A/C, user B to like article B, user C happinesses
Joyous article A/C/D.It can visually see, user A is more similar a kind of crowd to user C, therefore can push away article D
It recommends and gives user A.
In addition, the collaborative filtering recommending based on article can be, user A likes article A/C, user B to like article A/B/
C, user C like article A, it is seen that article A and article C are liked by most user, they have in user group A/B/C
Similitude, therefore article C can be recommended to not yet used user C in crowd.
In some feasible embodiments, the not very identical user of some hobbies can also be rejected first, is left behind
The consistent user of some hobbies, and with reference to the preference value to candidate items of these users.In some feasible embodiments
In, one or more can be first determined with reference to user, and the one or more is with reference to user to each in the candidate recommendation project
The preference value of project is stored in the preference value set.
In some feasible embodiments, optionally, can also by calculate it is each with reference in user with target user's
Like deviation, for screening with reference to user, leave some reference users more consistent with target user's hobby.Specifically
, can user 1 be obtained to user 5 (user1-user5) and target user (userM) for 1 He of project with reference chart 2-2
The preference value of project 2 (item1 and item2), obtains 3,4,5,1,4,3 and 3,3,5,1,5,4.Then from user 1 to user
5 one by one comparison user M for the deviation of the preference value of project 1 or project 2, obtain hobby deviation, which can be with
Absolute value of the difference for two preference values, or the difference of two preference values square or square root or other be used for table
Show the numerical value of deviation, be not construed as limiting herein.
By taking absolute value of the difference as an example, as the embodiment can obtain the hobby deviation of such as user 1 and target user for project 1
Value is | 3-3 |=0, and user 2 and target user are for the hobby deviation of project 1 | 4-3 |=1, user 3 and target user couple
In the hobby deviation of project 1, it is | 5-3 |=2, user 4 and target user be for the hobby deviation of project 1 | 1-3 |=
2, user 5 and target user be for the hobby deviation of project 1 | 4-3 |=1, similarly, can it is inclined for the hobby of project 2
Difference is respectively 1,1,2,4,1.It may be considered that the hobby deviation of user 1 and target user are 1, user 2 and target user
Hobby deviation for 2, the hobby deviation of user 3 and target user are 3, and the hobby deviation of user 4 and target user are
5, the hobby deviation of user 4 and target user are 2, and so on, then it can obtain hobby deviation value set.
In some feasible embodiments, can between two users using computing unit come calculate two users for
The hobby deviation of same project.Specifically, please referring to Fig.2-4, the computing unit comprising absolute value circuit is such as user 1
(user1) computing unit between user 2 (user2).The computing unit includes absolute value circuit and memristor MR0.Wherein,
The absolute value circuit is used to calculate the absolute value of the deviation between user 1 and user 2, if user 1 is to the preference of destination item
It is 3 to be worth, and user 1 is 1 to the preference value of destination item, and the deviation between user1 and user2 is | 1-3 |=| -2 |=2.
In other feasible embodiments, absolute value circuit can change squaring circuit or square root circuit into, as long as can be used for
Represent that user1 and user2 to the deviation of the preference value of same project, is not limited herein.It should be noted that this is absolutely
The specific implementation for being worth circuit is the prior art, and details are not described herein again.It, can be in project 1 in some feasible embodiments
Place inputs a burst pulse, and the transverse direction of 1~user of user M has electric current and passes through at this time, and each absolute calculators will be counted respectively
The absolute value of the difference between 1~user of user 5 and user M is calculated, after absolute calculators, result is had into recalling for the right
It hinders in device MR, respectively obtains the value of MR0, MR1, MR2, MR3, MR4, MR5.
206th, the target user is calculated to the candidate according to institute's candidate items preference value set and the hobby deviation value set
The recommendation of each project in recommended project.
In some feasible embodiments, after candidate items preference value set and hobby deviation value set is obtained,
Can the larger user of some hobby deviations be screened out according to hobby deviation value set, be left some close, roots of hobby
Determine to calculate the recommendation of each project to the preference value of candidate items according to the close user of these hobbies.Specifically,
- 5 are please referred to Fig.2, to calculate recommendation method, the hobby deviation as user 3 and user 4 obtain is larger, does not consider it partially
Good value, remaining user 1, user 2 and user 5 are to be closer to the hobby of target user M.Then can use user 1,
User 2 and user 5 determine its recommendation to the preference value of project 3 and project 4, specifically, recommendation can be added, it can also
Deviation will be used to be weighted as weight to each value, obtain recommendations of the target user M to candidate items 3 and project 4.
207th, it is determined as the one or more destination item of target user recommendation according to the recommendation value set.
In some feasible embodiments, the target user can be calculated according to the preference data set respectively to the candidate
The recommendation of each project in recommended project obtains recommending value set, and using the recommendation value set as determining whether to push away
Recommend the foundation of a project.
The recommendation that project 3 is such as obtained by calculation is 13, and the recommendation to project 4 is 4, if then project 3 and project 4
Recommend one, be then likely to a recommended project 3.
In some feasible embodiments, can according to recommend value set be determined as the target user recommendation this or
Multiple destination items, specifically, can determine first in the recommendation value set higher than threshold value recommendation corresponding to one or
Multiple projects for the one or more destination item, then determine highest one or more of recommendation in the recommendation value set
A project for the one or more destination item, can also determine the destination item recommended, not limit herein by other means
It is fixed.
It please refers to Fig.3, the present invention also provides a kind of project recommendation device 300 based on memristor, including:
Acquisition module 301, for obtaining preference value set, which includes multiple preference values, wherein, it is multiple
A preference value in preference value is a user to the preference value of a project.
Memory module 302, including multiple memristors, for storing the preference value set, wherein, in multiple memristor
There is always a memristors to store any one of preference value set mesh preference value.
Recommending module 303, for being determined as one or more target items of target user's recommendation according to the preference value set
Mesh.
It please refers to Fig.4, which includes:
Determination sub-module 3031, for determining candidate recommendation project set, which includes multiple times
Select recommended project.
Computational submodule 3032, for calculating the target user respectively to the Candidate Recommendation item according to the preference data set
The recommendation of each project in mesh obtains recommending value set.
Recommend submodule 3033, for being determined as the one or more of target user recommendation according to the recommendation value set
Destination item.
Fig. 5 is please referred to, which includes:
First determination unit 30321, for determining one or more with reference to user to each in the candidate recommendation project
The preference value of project obtains candidate items preference value set, which is contained in the preference value set.
First computing unit 30322, for calculating the one or more with reference to use according to the candidate items preference value set
In family it is each with reference to user respectively with the hobby deviation of the target user, obtain hobby deviation value set.
Second computing unit 30323, for being calculated according to institute's candidate items preference value set and the hobby deviation value set
The target user is to the recommendation of each project in the candidate recommendation project.
Fig. 6 is please referred to, which includes:
Second determination unit 30331, for determining in the recommendation value set higher than one corresponding to the recommendation of threshold value
Or multiple projects, for the one or more destination item.
Fig. 7 is please referred to, which includes:
Third determination subelement 30332, for determining the highest one or more items of the recommendation in the recommendation value set
Mesh, for the one or more destination item.
Fig. 8 is please referred to, the present invention also provides a kind of project recommendation device 800 based on memristor, including:
Bus 801, transceiver 802, processor 803 and memory 804.
The bus 801 is used to connect with the transceiver 802, the processor 803 and the memory 804.
The transceiver 802, for obtaining preference value set, which includes multiple preference values, wherein, it is multiple
A preference value in preference value is a user to the preference value of a project.
The transceiver 802 can include the communication interface (English between processor 803 and standard communication subsystem
communication interface)。
The transceiver 802 can further include the substandard communication interfaces of EIA-RS-232C, i.e. data terminal equipment
(Data Terminal Equipment, DTE) and data communications equipment (Data Circuit-terminating
Equipment, DCE) between SERIAL BINARY DATA Fabric Interface technical standard communication interface, can also include RS-485 assist
Communication interface under view, is not construed as limiting herein.
The memory 804 includes multiple memristors, for storing the preference value set, wherein, in multiple memristor
There is always a memristors to store any one of preference value set mesh preference value.
The memory 804 can include volatile memory (volatile memory), such as random access memory
(random-access memory, RAM).The memory 804 can also include nonvolatile memory (non-volatile
), such as flash memory (flash memory), hard disk (hard disk drive, HDD) or solid state disk memory
(solid-state drive, SSD).The memory 804 can also include the combination of the memory of mentioned kind, not make herein
It limits.
Optionally, which can be also used for storage program instruction, which can call the memory
The program instruction stored in 804.
The processor 803, for being determined as one or more target items of target user's recommendation according to the preference value set
Mesh.
The processor 803 can be central processing unit (central processing unit, CPU), network processing unit
The combination of (network processor, NP) or CPU and NP.
The processor 803 can further include hardware chip.Above-mentioned hardware chip can be application-specific integrated circuit
(application-specific integrated circuit, ASIC), programmable logic device (programmable
Logic device, PLD) or combination.Above-mentioned PLD can be Complex Programmable Logic Devices (complex
Programmable logic device, CPLD), field programmable gate array (field-programmable gate
Array, FPGA), Universal Array Logic (generic array logic, GAL) or its arbitrary combination.
It is apparent to those skilled in the art that for convenience and simplicity of description, the system of foregoing description,
The specific work process of device and unit can refer to the corresponding process in preceding method embodiment, and details are not described herein.
In several embodiments provided herein, it should be understood that disclosed system, device and method can be with
It realizes by another way.For example, the apparatus embodiments described above are merely exemplary, for example, the unit
It divides, only a kind of division of logic function can have other dividing mode, such as multiple units or component in actual implementation
It may be combined or can be integrated into another system or some features can be ignored or does not perform.Another point, it is shown or
The mutual coupling, direct-coupling or communication connection discussed can be the indirect coupling by some interfaces, device or unit
It closes or communicates to connect, can be electrical, machinery or other forms.
The unit illustrated as separating component may or may not be physically separate, be shown as unit
The component shown may or may not be physical unit, you can be located at a place or can also be distributed to multiple
In network element.Some or all of unit therein can be selected according to the actual needs to realize the mesh of this embodiment scheme
's.
In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, it can also
That each unit is individually physically present, can also two or more units integrate in a unit.Above-mentioned integrated list
The form that hardware had both may be used in member is realized, can also be realized in the form of SFU software functional unit.
If the integrated unit is realized in the form of SFU software functional unit and is independent product sale or uses
When, it can be stored in a computer read/write memory medium.Based on such understanding, technical scheme of the present invention is substantially
The part to contribute in other words to the prior art or all or part of the technical solution can embody in the form of software
Out, which is stored in a storage medium, and being used including some instructions (can be with so that computer equipment
Be personal computer, server or the network equipment etc.) perform all or part of step of each embodiment the method for the present invention
Suddenly.And aforementioned storage medium includes:USB flash disk, read-only memory (ROM, Read-Only Memory), is deposited mobile hard disk at random
The various media that can store program code such as access to memory (RAM, Random Access Memory), magnetic disc or CD.
The above, the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations;Although with reference to before
Embodiment is stated the present invention is described in detail, it will be understood by those of ordinary skill in the art that:It still can be to preceding
The technical solution recorded in each embodiment is stated to modify or carry out equivalent replacement to which part technical characteristic;And these
Modification is replaced, the spirit and scope for various embodiments of the present invention technical solution that it does not separate the essence of the corresponding technical solution.
Claims (11)
1. a kind of item recommendation method based on memristor, which is characterized in that including:
Preference value set is obtained, the preference value set includes multiple preference values, wherein, one in the multiple preference value is partially
Good value is a user to the preference value of a project;
The preference value set is stored in a storage module, the memory module includes multiple memristors, wherein, the multiple
There is always a memristors in memristor to store any one of preference value set mesh preference value;
It is determined as one or more destination items of target user's recommendation according to the preference value set.
2. method according to claim 1, which is characterized in that described target user is determined as according to the preference value set to push away
The one or more destination items recommended include:
Determine candidate recommendation project set, the candidate recommendation project set includes multiple candidate recommendation projects;
The target user is calculated respectively to each project in the candidate recommendation project according to the preference data set
Recommendation obtains recommending value set;
It is determined as one or more of destination items of target user's recommendation according to the recommendation value set.
3. method according to claim 2, which is characterized in that described that the target use is calculated according to the preference data set
Family respectively includes the recommendation of each project in the candidate recommendation project:
Determine that one or more, to the preference value of each project in the candidate recommendation project, obtains candidate items with reference to user
Preference value set, the candidate items preference value set are contained in the preference value set;
It is calculated according to the candidate items preference value set one or more of with reference to each with reference to user's difference in user
With the hobby deviation of the target user, hobby deviation value set is obtained;
The target user is calculated to the candidate recommendation project according to the preference value set and the hobby deviation value set
In each project recommendation.
4. method according to claim 2, which is characterized in that described that the target use is determined as according to the recommendation value set
One or more of destination items that family is recommended include:
It determines one or more projects corresponding to higher than the recommendation of threshold value in the recommendation value set, is one or more
A destination item.
5. method according to claim 2, which is characterized in that described that the target use is determined as according to the recommendation value set
One or more of destination items that family is recommended include:
It determines the highest one or more projects of recommendation in the recommendation value set, is one or more of target items
Mesh.
6. a kind of project recommendation device based on memristor, which is characterized in that including:
Acquisition module, for obtaining preference value set, the preference value set includes multiple preference values, wherein, it is the multiple inclined
A preference value in good value is a user to the preference value of a project;
Memory module, including multiple memristors, for storing the preference value set, wherein, it is total in the multiple memristor
There are a memristors to store any one of preference value set mesh preference value;
Recommending module, for being determined as one or more destination items of target user's recommendation according to the preference value set.
7. method according to claim 6, which is characterized in that the recommending module includes:
Determination sub-module, for determining candidate recommendation project set, the candidate recommendation project set includes multiple Candidate Recommendations
Project;
Computational submodule, for calculating the target user respectively to the candidate recommendation project according to the preference data set
In each project recommendation, obtain recommend value set;
Recommend submodule, for being determined as one or more of mesh of target user's recommendation according to the recommendation value set
Mark project.
8. method according to claim 7, which is characterized in that the computational submodule includes:
First determination unit, for determining one or more preferences with reference to user to each project in the candidate recommendation project
Value, obtains candidate items preference value set, which is contained in the preference value set;
First computing unit, it is one or more of with reference in user for being calculated according to the candidate items preference value set
It is each with reference to user respectively with the hobby deviation of the target user, obtain hobby deviation value set;
Second computing unit, for calculating the target user couple according to the preference value set and the hobby deviation value set
The recommendation of each project in the candidate recommendation project.
9. method according to claim 7, which is characterized in that the recommendation submodule includes:
Second determination unit, for determining to be higher than one or more items corresponding to the recommendation of threshold value in the recommendation value set
Mesh is one or more of destination items.
10. method according to claim 7, which is characterized in that the recommendation submodule includes:
Third determination subelement, for determining the highest one or more projects of recommendation in the recommendation value set, for institute
State one or more destination items.
11. a kind of project recommendation device based on memristor, which is characterized in that including:
Bus, transceiver, processor and memory;
The bus is used to connect with the transceiver, the processor and the memory;
The transceiver, for obtaining preference value set, the preference value set includes multiple preference values, wherein, it is the multiple
A preference value in preference value is a user to the preference value of a project;
The memory includes multiple memristors, for storing the preference value set, wherein, it is total in the multiple memristor
There are a memristors to store any one of preference value set mesh preference value;
The processor, for being determined as one or more destination items of target user's recommendation according to the preference value set.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611192441.1A CN108228642B (en) | 2016-12-21 | 2016-12-21 | Memristor-based project recommendation and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611192441.1A CN108228642B (en) | 2016-12-21 | 2016-12-21 | Memristor-based project recommendation and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108228642A true CN108228642A (en) | 2018-06-29 |
CN108228642B CN108228642B (en) | 2022-03-11 |
Family
ID=62650905
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611192441.1A Active CN108228642B (en) | 2016-12-21 | 2016-12-21 | Memristor-based project recommendation and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108228642B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2020087619A1 (en) * | 2018-10-29 | 2020-05-07 | 华中科技大学 | Reduced instruction set processor based on memristor |
CN111798280A (en) * | 2020-09-08 | 2020-10-20 | 腾讯科技(深圳)有限公司 | Multimedia information recommendation method, device and equipment and storage medium |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102231166A (en) * | 2011-07-12 | 2011-11-02 | 浙江大学 | Collaborative recommendation method based on social context |
US20120028714A1 (en) * | 2009-01-19 | 2012-02-02 | Wms Gaming, Inc. | Transporting and using wagering game data |
CN102542489A (en) * | 2011-12-27 | 2012-07-04 | 纽海信息技术(上海)有限公司 | Recommendation method based on user interest association |
CN103927673A (en) * | 2014-03-24 | 2014-07-16 | 腾讯科技(深圳)有限公司 | Method and device for recommending service information |
US20140229539A1 (en) * | 2013-02-13 | 2014-08-14 | Jason Daniel Powers | System and method for tag-based social networking |
CN104298772A (en) * | 2014-10-29 | 2015-01-21 | 吴健 | Collaborative filtering recommendation method and device optimizing neighbor selection |
CN104331411A (en) * | 2014-09-19 | 2015-02-04 | 华为技术有限公司 | Item recommendation method and item recommendation device |
US9152679B2 (en) * | 2013-01-31 | 2015-10-06 | Hewlett-Packard Development Company, L.P. | Displaying recommended entities in a relevance map |
CN105247511A (en) * | 2013-06-07 | 2016-01-13 | 苹果公司 | Intelligent automated assistant |
CN105390520A (en) * | 2015-10-21 | 2016-03-09 | 清华大学 | Parameter configuration method for memristor intersection array |
WO2016167785A1 (en) * | 2015-04-17 | 2016-10-20 | Hewlett Packard Enterprise Development Lp | Determining resistive state of memristors |
-
2016
- 2016-12-21 CN CN201611192441.1A patent/CN108228642B/en active Active
Patent Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120028714A1 (en) * | 2009-01-19 | 2012-02-02 | Wms Gaming, Inc. | Transporting and using wagering game data |
CN102231166A (en) * | 2011-07-12 | 2011-11-02 | 浙江大学 | Collaborative recommendation method based on social context |
CN102542489A (en) * | 2011-12-27 | 2012-07-04 | 纽海信息技术(上海)有限公司 | Recommendation method based on user interest association |
US9152679B2 (en) * | 2013-01-31 | 2015-10-06 | Hewlett-Packard Development Company, L.P. | Displaying recommended entities in a relevance map |
US20140229539A1 (en) * | 2013-02-13 | 2014-08-14 | Jason Daniel Powers | System and method for tag-based social networking |
CN105247511A (en) * | 2013-06-07 | 2016-01-13 | 苹果公司 | Intelligent automated assistant |
CN103927673A (en) * | 2014-03-24 | 2014-07-16 | 腾讯科技(深圳)有限公司 | Method and device for recommending service information |
CN104331411A (en) * | 2014-09-19 | 2015-02-04 | 华为技术有限公司 | Item recommendation method and item recommendation device |
CN104298772A (en) * | 2014-10-29 | 2015-01-21 | 吴健 | Collaborative filtering recommendation method and device optimizing neighbor selection |
WO2016167785A1 (en) * | 2015-04-17 | 2016-10-20 | Hewlett Packard Enterprise Development Lp | Determining resistive state of memristors |
CN105390520A (en) * | 2015-10-21 | 2016-03-09 | 清华大学 | Parameter configuration method for memristor intersection array |
Non-Patent Citations (3)
Title |
---|
张永华等: "第四种基本电路元件忆阻器及其应用", 《器件与技术》 * |
胡小方: "基于忆阻器的非易失性存储器研究", 《中国优秀硕士学位论文全文数据库 (信息科技辑)》 * |
胡舒凯等: "忆阻器存储研究与展望", 《计算机研究与发展》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2020087619A1 (en) * | 2018-10-29 | 2020-05-07 | 华中科技大学 | Reduced instruction set processor based on memristor |
US11830547B2 (en) | 2018-10-29 | 2023-11-28 | Huazhong University Of Science And Technology | Reduced instruction set processor based on memristor |
CN111798280A (en) * | 2020-09-08 | 2020-10-20 | 腾讯科技(深圳)有限公司 | Multimedia information recommendation method, device and equipment and storage medium |
CN111798280B (en) * | 2020-09-08 | 2020-12-15 | 腾讯科技(深圳)有限公司 | Multimedia information recommendation method, device and equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN108228642B (en) | 2022-03-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US9183497B2 (en) | Performance-efficient system for predicting user activities based on time-related features | |
CN103514204B (en) | Information recommendation method and device | |
CN104408043B (en) | A kind of information processing method and server | |
CN103841122B (en) | Target object information recommends method, server and client | |
EP2506522A1 (en) | Method and device for pushing data | |
CN101937547A (en) | Software and/or software information pushing method, system, acquisition device, software shop service system and mobile terminal | |
EP3070661A1 (en) | System and method for providing context driven hyper-personalized recommendation | |
US20070168208A1 (en) | Location recommendation method and system | |
CN112231350B (en) | Enterprise business opportunity mining method and device based on knowledge graph | |
Socievole et al. | Wireless contacts, Facebook friendships and interests: analysis of a multi-layer social network in an academic environment | |
US20090192872A1 (en) | Inferring environmental knowledge through near field communication and data mining | |
Mo et al. | Event recommendation in social networks based on reverse random walk and participant scale control | |
US11632702B2 (en) | Intelligent event-based network routing | |
CN107092609A (en) | A kind of information-pushing method and device | |
CN111460279A (en) | Information recommendation method and device, storage medium and computer equipment | |
CN108228642A (en) | A kind of project recommendation and device based on memristor | |
CN106777239A (en) | Recommendation information generation method, device and computer equipment | |
CN103188348A (en) | Linkman management method based on file sharing | |
CN108846708A (en) | User's buying behavior prediction technique, device, equipment and storage medium | |
CN106897898A (en) | The method and system that a kind of electric business platform intelligent is given | |
CN107612707B (en) | Preprocessing method and system for classified storage of homologous sample data in industry field | |
CN112069416B (en) | Cross-social network user identity recognition method based on community discovery | |
Yuan et al. | Partitioning social networks for fast retrieval of time-dependent queries | |
US10817511B2 (en) | Search dimensionality expansion | |
TWI518616B (en) | Method and electronic device for rating outfit |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |