CN107103033A - The preference Forecasting Methodology and device of cold start-up user - Google Patents
The preference Forecasting Methodology and device of cold start-up user Download PDFInfo
- Publication number
- CN107103033A CN107103033A CN201710170343.6A CN201710170343A CN107103033A CN 107103033 A CN107103033 A CN 107103033A CN 201710170343 A CN201710170343 A CN 201710170343A CN 107103033 A CN107103033 A CN 107103033A
- Authority
- CN
- China
- Prior art keywords
- user
- categories
- cold start
- preferences
- scoring
- 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
- G06F16/60—Information retrieval; Database structures therefor; File system structures therefor of audio data
- G06F16/63—Querying
- G06F16/635—Filtering based on additional data, e.g. user or group profiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
- G06F16/43—Querying
- G06F16/435—Filtering based on additional data, e.g. user or group profiles
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application provides a kind of preference Forecasting Methodology of cold start-up user, including：Obtain in N number of frame of reference, scoring of the cold start-up user in M categories of preferences；N is the natural number more than 1, and M is natural number；The frame of reference is the other systems that cold start-up user uses in addition to the system；According to some categories of preferences in the scoring of N number of frame of reference, prediction score value of the cold start-up user in the categories of preferences is obtained.The technical scheme of the application can sum up the preference of user from the used frame of reference of cold start-up user on the premise of without cold start-up user operation, while cold start-up user time energy is saved, add the degree of accuracy of preference prediction.
Description
Technical field
The application is related to technical field of data processing, more particularly to a kind of cold start-up user preference Forecasting Methodology and dress
Put.
Background technology
With the extensive use of information technology, increasing enterprise provides a user various online service systems.Online
The advantage of service system is the behavior of user can be tracked and be stored, and mining analysis is used from the historical record of user
The preference at family, to provide more targetedly service for each user.The display interface of online service system, push content, arrive
Operation flow etc., can by user preference adjust automatically so that more match user demand.
Excavation and analysis to user preference, based on the historical record of user.There is no historical record or can not be from
The user that few historical record analyzes preference is referred to as cold start-up user.For example, to just reached the standard grade system, APP apply, institute
Some users are cold start-up users；For another example, the system to having run a period of time, the user of new registration is that cold start-up is used
Family.
In the prior art, the preference of cold start-up user is obtained generally by the way of investigation to cold start-up user, i.e., by
The application form that cold start-up user is filled in and submission system is provided, from cold start-up user fill in result in summarize cold start-up user
Preference.When just having begun to use some web film such as user, system, which can push some films, allows user to choose whether to like；Or
When user has just enter into certain music site, system can allow user to select liked music type etc..This mode needs to take
The time and efforts of cold start-up user, and obtained information is very limited, is difficult often that reaction cold start-up user is really emerging
Interest and preference.
The content of the invention
In view of this, the application provides a kind of preference Forecasting Methodology of cold start-up user, including：
Obtain in N number of frame of reference, scoring of the cold start-up user in M categories of preferences；N is the natural number more than 1,
M is natural number；The frame of reference is the other systems with the historical record for analyzing the cold start-up user preference；
According to some categories of preferences in the scoring of N number of frame of reference, the cold start-up user is obtained in the categories of preferences
Prediction score value.
A kind of preference Forecasting Methodology for cold start-up user that the application is provided, is applied with the analysis cold start-up user
In the system of the historical record of preference, including：
The request that cold starting system obtains the classification scoring of cold start-up user preference is received, the request includes cold start-up use
The personal information at family；The cold starting system is the system for lacking the historical record for analyzing the cold start-up user preference；
Scoring of the cold start-up user in M categories of preferences is generated according to the personal information of cold start-up user, M are commented
Divide and return to cold starting system, cooling starts according to some categories of preferences in the scoring of the system and other frames of reference, obtains
Prediction score value of the cold start-up user in the categories of preferences；M is natural number.
Present invention also provides a kind of preference prediction meanss of cold start-up user, including：
Classification scores acquiring unit, and for obtaining in N number of frame of reference, the cold start-up user is in M categories of preferences
Scoring；N is the natural number more than 1, and M is natural number；The frame of reference is going through with the analysis cold start-up user preference
The other systems of Records of the Historian record；
Score value computing unit is predicted, for, in the scoring of N number of frame of reference, obtaining described cold open according to some categories of preferences
Employ prediction score value of the family in the categories of preferences.
A kind of preference prediction meanss for cold start-up user that the application is provided, are applied with the analysis cold start-up user
In the system of the historical record of preference, including：
Scored request reception unit, and the request of cold start-up user preference classification scoring is obtained for receiving cold starting system,
The request includes the personal information of cold start-up user；The cold starting system analyzes the cold start-up user preference to lack
Historical record system；
Score request-response unit, for generating the cold start-up user at M according to the personal information of cold start-up user
The scoring of categories of preferences, cold starting system is returned to by M scoring, and cooling starts according to some categories of preferences in the system and its
The scoring of his frame of reference, obtains prediction score value of the cold start-up user in the categories of preferences；M is natural number.
In above technical scheme, embodiments herein, the referential used from least two cold start-up users
In system, the scoring of each categories of preferences of cold start-up user is obtained, then some categories of preferences is comprehensive in the scoring of each frame of reference
Be combined into prediction score value of the cold start-up user in the categories of preferences, so as to without cold start-up user operation on the premise of,
The preference of user is summed up from the used frame of reference of cold start-up user, while cold start-up user time energy is saved,
Add the degree of accuracy of preference prediction.
Brief description of the drawings
Fig. 1 is that one kind is applied in cold starting system in the embodiment of the present application, the preference Forecasting Methodology of cold start-up user
Flow chart；
Fig. 2 is that one kind is applied in frame of reference in the embodiment of the present application, the stream of the preference Forecasting Methodology of cold start-up user
Cheng Tu；
Fig. 3 is the schematic diagram of the pre- flow gauge of preference in the application application example；
Fig. 4 is a kind of hardware structure diagram for the equipment for running the embodiment of the present application；
Fig. 5 is that one kind is applied in cold starting system in the embodiment of the present application, the preference prediction meanss of cold start-up user
Building-block of logic；
Fig. 6 is that one kind is applied in frame of reference in the embodiment of the present application, and the preference prediction meanss of cold start-up user are patrolled
Collect structure chart.
Embodiment
With online service and the diversification of application, increasing user's selection is participated in multiple service systems.When
When one user turns into new user in one runtime, the user to more systems that other have run then usually
It is old user.To the firm system reached the standard grade soon, all users are cold start-up users, but these users at other
In system through even running, it is the old user for possessing abundant historical record to also tend to.The preference of user be it is stable, will not
Different and different with used system, therefore, the preference of the cold start-up user of some system generally has been reflected in the use
Family is using (to these other systems, the user has no longer been cold start-up user) in other systems after a while.Thing
In reality, these abundant cross-system historical datas are the precious resources of study user preference in the case of cold start-up, can be borrowed cold
Start historical record of the user in other systems, summarize the preference of cold start-up user, then for predicting that the cold start-up is used
The preference of family in the present system.
For example, Taobao user is likely to the user for being also Sina weibo or dried shrimp music.User has just begun to use dried shrimp sound
When happy, he there is no historical record in the system of dried shrimp music, but he is likely to leave in other Taobao or Sina weibo
Abundant behavior history.If the user often forwards or commented on the microblogging of certain singer in Sina weibo, then it can predict
The user is the loyal bean vermicelli of the singer, the music of the singer musically to be recommended into the user in dried shrimp, even if
The user is the cold start-up user of dried shrimp music.
Embodiments herein proposes a kind of preference Forecasting Methodology of new cold start-up user, and the preference of user is divided into M
Individual categories of preferences, calculates from least two frames of reference that cold start-up user has used and obtains cold start-up user each
The scoring of categories of preferences, then draw the user as the cold start-up user of the system all scorings synthesis of one categories of preferences
In the prediction score value of the categories of preferences, fill in questionnaire without cold start-up user and more can accurately position this and cold open
The preference at family is employed, reduces and cold start-up user is bothered, to solve problems of the prior art.
Embodiments herein may operate in any with calculating in the equipment with storage capacity, such as mobile phone, flat board electricity
The equipment such as brain, PC (Personal Computer, PC), notebook, server；Can also be by operating in two or two
The logical node of individual above equipment realizes the various functions in the embodiment of the present application.
In embodiments herein, some user not yet has enough historical records in a system, can make this be
System analyzes the preference of the user, and the user is cold start-up user for the system, and the system can just be reached the standard grade, reach the standard grade
Soon or the system for a long time of even running, do not limit.Above-mentioned the system is referred to as cold starting system in the embodiment of the present application,
By in addition to the system, the user uses has enough historical records to make for analyzing N number of other systems of the user preference
For frame of reference, using historical record of the user in N number of frame of reference, to predict that the user is inclined when using the system
It is good.
The preference Forecasting Methodology of cold start-up user applies the flow on cold starting system as shown in figure 1, applying in reference
Flow in system is as shown in Figure 2.
On cold starting system, step 110, obtain in N (N is the natural number more than 1) individual frame of reference, the cold start-up
Scoring of the user in M (M is natural number) individual categories of preferences.
In the embodiment of the present application, the N number of frame of reference of identical can be all used to all cold start-up users, can also be right
Different cold start-up users uses different N number of frames of reference, does not limit.For example, it is possible to use the registration of cold start-up user
The used N number of system of the information searching user, and as frame of reference.
It can be determined to divide how many categories of preferences according to the business demand of cold starting system, and how to divide preference
Classification.The scoring of respective M categories of preferences can be calculated by the historical record of each frame of reference using identical algorithm,
Can also according to each frame of reference business characteristic, the scoring of respective M categories of preferences is calculated using different algorithms.
Embodiments herein is not construed as limiting.
In addition, the mode of division categories of preferences, the algorithm of M categories of preferences scoring of calculating can be with reference in the prior art
The analysis method for excavating user preference using historical record in the present system is realized.For example, can be belonged to based on cold start-up user
The statistics of the behavior record of some categories of preferences come provide the categories of preferences scoring (such as music preferences classification can according to
The sum of family play list in some fixed time period provides scoring)；For another example, existing collaborative filtering, square can be used
Battle array decomposes scheduling algorithm to calculate the scoring of M categories of preferences.
In the first implementation, cold starting system can obtain the history of cold start-up user from some frame of reference
Record, voluntarily calculates the scoring to M categories of preferences of the frame of reference according to the historical record of cold start-up user.
In two kinds of implementations, can by some frame of reference using cold start-up user the system historical record as foundation, calculate
Scoring of the cold start-up user to M categories of preferences of the frame of reference is drawn, and M scoring is passed into cold starting system.
In the third implementation, above two implementation can be respectively adopted to different frames of reference, pass through above two
The combination of implementation obtains scoring of the cold start-up user in M categories of preferences in different frames of reference respectively.
Specifically, in above-mentioned second or the third implementation, in some frame of reference, step 210, receive
Cold starting system obtains the request of cold start-up user preference classification scoring, and the request includes the personal letter of cold start-up user
Breath.
Cold starting system can ask the preference point of the user using the personal information of cold start-up user to frame of reference
Class scores.Which or which information of cold start-up user can be used as by personal information according to practical application scene.Example
Such as, user can use the social account of oneself or log in many exist in the login account of some large-scale websites in the prior art
In line service system, this application scenarios, cold starting system can using the login account of user as cold start-up user individual
Information.For another example, cold starting system can be by the passport NO. of user, Terminal Equipment Identifier, wearable device mark, cell-phone number
The information that code etc. can represent user identity is used as the personal information of cold start-up user.
In above-mentioned frame of reference, step 220, cold start-up user is generated in M according to the personal information of cold start-up user
The scoring of individual categories of preferences, cold starting system is returned to by M scoring, cooling start according to some categories of preferences in the system and
The scoring of other frames of reference, obtains prediction score value of the cold start-up user in the categories of preferences.
Frame of reference can obtain the historical record of the user according to the personal information of cold start-up user, and historical record is entered
After row data cleansing and filtering, scoring of the user in M categories of preferences is obtained using predetermined preference categories algorithm, and by M
Individual scoring returns to cold starting system.Specific data cleansing and filter method, preference categories algorithm can use prior art
Realize, repeat no more.
On cold starting system, step 120, according to some categories of preferences in the scoring of N number of frame of reference, obtain this and cold open
Employ prediction score value of the family in the categories of preferences.
In cold start-up user in obtaining N number of frame of reference after the scoring of M categories of preferences, to each categories of preferences, altogether
There is the N number of scoring for corresponding respectively to N number of frame of reference.Based on this N number of scoring, cold open can be obtained using certain algorithm
Employ prediction score value of the family in single categories of preferences.
Different categories of preferences can predict score value using identical algorithm by N number of score calculation, it would however also be possible to employ no
Same algorithm.In a kind of application scenarios, the business characteristic of different frames of reference is different, therefore the historical record pair of different system
The contribution of the different categories of preferences of cold start-up user is also different., can be according to some categories of preferences in N in this application scenarios
The scoring of individual frame of reference and the degree of correlation of each frame of reference and the categories of preferences, to obtain cold start-up user in the preference
The prediction score value of classification；In other words, the degree of correlation of each frame of reference and the categories of preferences is reflected in the algorithm of use, with
The higher frame of reference of the categories of preferences degree of correlation, has bigger influence to prediction score value.
Scoring of the embodiments herein to some categories of preferences in comprehensive N number of different frames of reference, to obtain the preference
The specific algorithm of the prediction score value of classification is not limited.In one example, can be to some categories of preferences in N number of frame of reference
Scoring carry out linear weighted function, to obtain prediction score value of the cold start-up user in the categories of preferences, such as formula 1：
In formula 1, EPtFor the prediction score value of t-th of categories of preferences, t is the natural number from 1 to M；Pi,tFor i-th of referential
The scoring of t-th of categories of preferences in system；ωi,tTo calculate EPtWhen Pi,tWeight.
In this example, different categories of preferences, can combine to calculate prediction score value using different weights.Change speech
It, is when calculating the prediction score value of single categories of preferences, N number of weight that the scoring of N number of frame of reference is used in linear weighted function,
Can be because of categories of preferences it is different and different.Pair frame of reference high with categories of preferences degree of correlation, can use larger power
Calculate the prediction score value of the categories of preferences again；For example, being film respectively for two frames of reference to categories of preferences " film "
Booking class website and the situation of sports news class website, can assign film booking class website larger weight, and assign physical culture
The less weight in news category website.
After the prediction score value of M categories of preferences of cold start-up user is calculated, it can be learnt according to these prediction score values
Where interest of the cold start-up user between difference preference's classification.So, cold starting system can be according to the interest of cold start-up user
Personalized display interface is customized for the user, content, and/or operation flow etc. is pushed, and for cold start-up, user provides more
Targetedly service.
, can commenting to M categories of preferences of each frame of reference for the ease of carrying out the contrast between M categories of preferences
Divide and be normalized and (carried out by cold starting system or carried out by frame of reference)；Can to by N number of frame of reference to same
The scoring of one categories of preferences draw the categories of preferences prediction score value algorithm be normalized (such as to linear weighted function algorithm,
It can make when calculating the prediction score value of some categories of preferences, use N number of weight of the scoring of N number of frame of reference and for 1, i.e.,
In formula 1)；Above-mentioned two normalized, the prediction point of such difference preference's classification can also be used simultaneously
The span of value can will be contrasted by simple size between zero and one, learn cold start-up user to which categories of preferences
It is interested.
It can be seen that, in embodiments herein, the preference of user is divided into M categories of preferences, used from least two cold start-ups
In the frame of reference that family is used, the scoring of M categories of preferences of cold start-up user is obtained, then by all scorings of a categories of preferences
Synthesis draws prediction score value of the cold start-up user in the categories of preferences, without investigating cold start-up user, you can
The preference of cold start-up user is more accurately drawn, reduces and cold start-up user is bothered, cold start-up user time is being saved
While energy, the degree of accuracy of preference prediction is improved.
In the application example of the application, user turns into new after being registered on a service platform (cold starting system)
User (cold start-up user).Cold starting system uses the personal information provided during the cold start-up user's registration (such as social account, hand
Machine number, passport NO. etc.), the preference categories for inquiring about the user to other service platforms score.Provided with other N number of service platforms
There is the user to be used for analyzing the historical record of its preference enough, this other N number of service platform turns into frame of reference 1, frame of reference
2, until frame of reference N.
Refer to Fig. 3, it is assumed that preference categories have M classes, and it is flat in this service that each frame of reference extracts cold start-up user
The historical record of platform, is cleaned and is filtered to historical record data and (remove invalid and noise data, can refer to prior art
Realize) after, respective preference categories algorithm is run, after arithmetic result is normalized, cold start-up user is drawn
In the scoring (summation of this M scoring is 1) of M preference categories, and M scoring is supplied to cold starting system.
Cold starting system is obtained after N*M scoring, by M scoring of each preference categories, inputs the line of the preference categories
Property weighted formula, obtains the prediction score value of the categories of preferences.The linear weighted function formula of each categories of preferences is using at normalization
Reason, i.e. N number of weight sum corresponding to N number of frame of reference are 1.In the linear weighted function formula of the classification of each preference, according to ginseng
Test system sets the weight corresponding to the scoring of the frame of reference with the degree of correlation of the categories of preferences, and degree of correlation is higher,
Weight is bigger.
Cold start-up user is obtained after the prediction score value of M preference categories in cold starting system, cold can be opened according to this
Several most interested preference categories of family are employed, to be directed to the business rule that cold start-up user runs cold starting system.
It is corresponding with the realization of above-mentioned flow, embodiments herein additionally provide it is a kind of apply in cold starting system cold open
The preference prediction meanss at family are employed, and a kind of preference prediction meanss of the cold start-up user applied in frame of reference.Both
Device can be realized by software, can also be realized by way of hardware or software and hardware combining.Exemplified by implemented in software,
As the device on logical meaning, being will be right by the CPU (Central Process Unit, central processing unit) of place equipment
The computer program instructions answered read what operation in internal memory was formed.For hardware view, except the CPU shown in Fig. 4, internal memory
And outside nonvolatile memory, the equipment where the preference prediction meanss of cold start-up user generally also includes being used to carry out nothing
Other hardware such as chip of line signal transmitting and receiving, and/or other hardware such as board for realizing network communicating function.
Fig. 5 show the preference prediction meanss of cold start-up user of the embodiment of the present application offer a kind of, including classification scoring
Acquiring unit and prediction score value computing unit, wherein：Classification scoring acquiring unit is used to obtain in N number of frame of reference, described cold
Start scoring of the user in M categories of preferences；N is the natural number more than 1, and M is natural number；The frame of reference is with analysis
The other systems of the historical record of the cold start-up user preference；Prediction score value computing unit is used to be existed according to some categories of preferences
The scoring of N number of frame of reference, obtains prediction score value of the cold start-up user in the categories of preferences.
Optionally, the scoring of M categories of preferences of each frame of reference is the score value after normalized.
In one example, it is described prediction score value computing unit specifically for：To some categories of preferences in N number of frame of reference
Scoring carries out linear weighted function, obtains prediction score value of the cold start-up user in the categories of preferences.
In above-mentioned example, used in linear weighted function N number of weight of the scoring of N number of frame of reference and for 1.
In above-mentioned example, N number of weight that the scoring of N number of frame of reference is used in linear weighted function, because of categories of preferences
It is different and different.
Optionally, it is described prediction score value computing unit specifically for：According to some categories of preferences commenting in N number of frame of reference
Point and frame of reference and the categories of preferences degree of correlation, the prediction for obtaining the cold start-up user in the categories of preferences divides
Value；The higher frame of reference with the categories of preferences degree of correlation, has bigger influence to the prediction score value.
Optionally, classification scoring acquiring unit specifically for：Using the personal information of the cold start-up user to ginseng
Test system asks the preference categories scoring of the cold start-up user, receives the system that N number of frame of reference returns and cold is opened to described
Employ scoring of the family in M categories of preferences.
Fig. 6 show the embodiment of the present application offer a kind of cold start-up user preference prediction meanss, apply with point
In the system for the historical record for analysing the cold start-up user preference, including scoring request reception unit and scoring ask response single
Member, wherein：Scoring request reception unit is used to receive the request that cold starting system obtains the classification scoring of cold start-up user preference, institute
Stating request includes the personal information of cold start-up user；The cold starting system is to lack the analysis cold start-up user preference
The system of historical record；The request-response unit that scores is used to generate the cold start-up user according to the personal information of cold start-up user
In the scoring of M categories of preferences, M scoring is returned into cold starting system, cooling startup is at this according to some categories of preferences
The scoring of system and other frames of reference, obtains prediction score value of the cold start-up user in the categories of preferences；M is natural number.
Optionally, the scoring of the M categories of preferences is the score value after normalized.
Optionally, it is described scoring request-response unit specifically for：According to being obtained the personal information of cold start-up user
The historical record of cold start-up user, is carried out after data cleansing and filtering to historical record, is obtained using predetermined preference categories algorithm
To the cold start-up user in the scoring of M categories of preferences, M scoring is returned into cold starting system.
The preferred embodiment of the application is the foregoing is only, not to limit the application, all essences in the application
God is with principle, and any modification, equivalent substitution and improvements done etc. should be included within the scope of the application protection.
In a typical configuration, computing device includes one or more processors (CPU), input/output interface, net
Network interface and internal memory.
Internal memory potentially includes the volatile memory in computer-readable medium, random access memory (RAM) and/or
The forms such as Nonvolatile memory, such as read-only storage (ROM) or flash memory (flash RAM).Internal memory is computer-readable medium
Example.
Computer-readable medium includes permanent and non-permanent, removable and non-removable media can be by any method
Or technology come realize information store.Information can be computer-readable instruction, data structure, the module of program or other data.
The example of the storage medium of computer includes, but are not limited to phase transition internal memory (PRAM), static RAM (SRAM), moved
State random access memory (DRAM), other kinds of random access memory (RAM), read-only storage (ROM), electric erasable
Programmable read only memory (EEPROM), fast flash memory bank or other memory techniques, read-only optical disc read-only storage (CD-ROM),
Digital versatile disc (DVD) or other optical storages, magnetic cassette tape, the storage of tape magnetic rigid disk or other magnetic storage apparatus
Or any other non-transmission medium, the information that can be accessed by a computing device available for storage.Define, calculate according to herein
Machine computer-readable recording medium does not include temporary computer readable media (transitory media), such as data-signal and carrier wave of modulation.
It should also be noted that, term " comprising ", "comprising" or its any other variant are intended to nonexcludability
Comprising so that process, method, commodity or equipment including a series of key elements are not only including those key elements, but also wrap
Include other key elements being not expressly set out, or also include for this process, method, commodity or equipment intrinsic want
Element.In the absence of more restrictions, the key element limited by sentence "including a ...", it is not excluded that wanted including described
Also there is other identical element in process, method, commodity or the equipment of element.
It will be understood by those skilled in the art that embodiments herein can be provided as method, system or computer program product.
Therefore, the application can be using the embodiment in terms of complete hardware embodiment, complete software embodiment or combination software and hardware
Form.Deposited moreover, the application can use to can use in one or more computers for wherein including computer usable program code
The shape for the computer program product that storage media is implemented on (including but is not limited to magnetic disk storage, CD-ROM, optical memory etc.)
Formula.
Claims (20)
1. a kind of preference Forecasting Methodology of cold start-up user, it is characterised in that it includes：
Obtain in N number of frame of reference, scoring of the cold start-up user in M categories of preferences；N is the natural number more than 1, and M is
Natural number；The frame of reference is the other systems with the historical record for analyzing the cold start-up user preference；
According to some categories of preferences in the scoring of N number of frame of reference, the cold start-up user is obtained in the pre- of the categories of preferences
Survey score value.
2. according to the method described in claim 1, it is characterised in that the scoring of M categories of preferences of each frame of reference
For the score value after normalized.
3. according to the method described in claim 1, it is characterised in that it is described according to some categories of preferences N number of reference scoring,
Prediction score value of the cold start-up user in the categories of preferences is obtained, including：To some categories of preferences in N number of frame of reference
Scoring carries out linear weighted function, obtains prediction score value of the cold start-up user in the categories of preferences.
4. method according to claim 3, it is characterised in that the scoring of N number of frame of reference is adopted in linear weighted function
N number of weight and for 1.
5. method according to claim 3, it is characterised in that the scoring of N number of frame of reference is adopted in linear weighted function
N number of weight is different and different because of categories of preferences.
6. the method according to claim 1 or 3, it is characterised in that it is described according to some categories of preferences in N number of frame of reference
Scoring, obtain prediction score value of the cold start-up user in the categories of preferences, including：According to some categories of preferences N number of
The scoring of frame of reference and the degree of correlation of frame of reference and the categories of preferences, obtain the cold start-up user in the preference
The prediction score value of classification；The higher frame of reference with the categories of preferences degree of correlation, has bigger to the prediction score value
Influence.
7. according to the method described in claim 1, it is characterised in that described to obtain in N number of frame of reference, the cold start-up user
In the scoring of M categories of preferences, including：The cold start-up is asked to frame of reference using the personal information of the cold start-up user
The preference categories scoring of user, receives the system that N number of frame of reference returns to the cold start-up user in M categories of preferences
Scoring.
8. a kind of preference Forecasting Methodology of cold start-up user, is applied with the historical record for analyzing the cold start-up user preference
System in, it is characterised in that it includes：
The request that cold starting system obtains the classification scoring of cold start-up user preference is received, the request includes cold start-up user's
Personal information；The cold starting system is the system for lacking the historical record for analyzing the cold start-up user preference；
The cold start-up user is generated in the scoring of M categories of preferences according to the personal information of cold start-up user, M scoring is returned
Back to cold starting system, cooling starts the scoring in the system and other frames of reference according to some categories of preferences, obtains described
Prediction score value of the cold start-up user in the categories of preferences；M is natural number.
9. method according to claim 8, it is characterised in that the scoring of the M categories of preferences is after normalized
Score value.
10. method according to claim 8, it is characterised in that described that institute is generated according to the personal information of cold start-up user
Scoring of the cold start-up user in M categories of preferences is stated, including：The cold start-up is obtained according to the personal information of cold start-up user to use
The historical record at family, carries out after data cleansing and filtering to historical record, obtains described cold using predetermined preference categories algorithm
Start scoring of the user in M categories of preferences.
11. a kind of preference prediction meanss of cold start-up user, it is characterised in that including：
Classification scoring acquiring unit, for obtaining in N number of frame of reference, scoring of the cold start-up user in M categories of preferences；
N is the natural number more than 1, and M is natural number；The frame of reference is with the historical record for analyzing the cold start-up user preference
Other systems；
Score value computing unit is predicted, for, in the scoring of N number of frame of reference, obtaining the cold start-up according to some categories of preferences and using
Prediction score value of the family in the categories of preferences.
12. device according to claim 11, it is characterised in that M categories of preferences of each frame of reference is commented
It is divided into the score value after normalized.
13. device according to claim 11, it is characterised in that the prediction score value computing unit specifically for：To certain
Individual categories of preferences carries out linear weighted function in the scoring of N number of frame of reference, obtains the cold start-up user in the categories of preferences
Predict score value.
14. device according to claim 13, it is characterised in that the scoring of N number of frame of reference is in linear weighted function
Use N number of weight and for 1.
15. device according to claim 13, it is characterised in that the scoring of N number of frame of reference is in linear weighted function
The N number of weight used is different and different because of categories of preferences.
16. the device according to claim 11 or 13, it is characterised in that the prediction score value computing unit specifically for：
According to some categories of preferences in the scoring of N number of frame of reference and the degree of correlation of frame of reference and the categories of preferences, institute is obtained
State prediction score value of the cold start-up user in the categories of preferences；Higher frame of reference, right with the categories of preferences degree of correlation
The prediction score value has bigger influence.
17. device according to claim 11, it is characterised in that the classification scoring acquiring unit specifically for：Using
The personal information of the cold start-up user asks the preference categories of the cold start-up user to score to frame of reference, receives N number of ginseng
Test system return the system to the cold start-up user M categories of preferences scoring.
18. a kind of preference prediction meanss of cold start-up user, are applied with the history note for analyzing the cold start-up user preference
In the system of record, it is characterised in that it includes：
Scored request reception unit, and the request of cold start-up user preference classification scoring is obtained for receiving cold starting system, described
Request includes the personal information of cold start-up user；The cold starting system is to lack to analyze going through for the cold start-up user preference
The system of Records of the Historian record；
Score request-response unit, for generating the cold start-up user in M preference according to the personal information of cold start-up user
The scoring of classification, cold starting system is returned to by M scoring, and cooling starts according to some categories of preferences in the system and other ginsengs
The scoring of test system, obtains prediction score value of the cold start-up user in the categories of preferences；M is natural number.
19. device according to claim 18, it is characterised in that the scoring of the M categories of preferences is normalized
Score value afterwards.
20. device according to claim 18, it is characterised in that the scoring request-response unit specifically for：According to
The personal information of cold start-up user obtains the historical record of the cold start-up user, and data cleansing and filtering are carried out to historical record
Afterwards, scoring of the cold start-up user in M categories of preferences is obtained using predetermined preference categories algorithm, M scoring is returned
To cold starting system.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710170343.6A CN107103033B (en) | 2017-03-21 | 2017-03-21 | Preference prediction method and device for cold-start user |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710170343.6A CN107103033B (en) | 2017-03-21 | 2017-03-21 | Preference prediction method and device for cold-start user |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107103033A true CN107103033A (en) | 2017-08-29 |
| CN107103033B CN107103033B (en) | 2021-04-27 |
Family
ID=59675081
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710170343.6A Active CN107103033B (en) | 2017-03-21 | 2017-03-21 | Preference prediction method and device for cold-start user |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107103033B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070136753A1 (en) * | 2005-12-13 | 2007-06-14 | United Video Properties, Inc. | Cross-platform predictive popularity ratings for use in interactive television applications |
| CN103198418A (en) * | 2013-03-15 | 2013-07-10 | 北京亿赞普网络技术有限公司 | Application recommendation method and application recommendation system |
| CN103605656A (en) * | 2013-09-30 | 2014-02-26 | 小米科技有限责任公司 | Music recommendation method and device and mobile terminal |
| CN104317959A (en) * | 2014-11-10 | 2015-01-28 | 北京字节跳动网络技术有限公司 | Data mining method and device based on social platform |
-
2017
- 2017-03-21 CN CN201710170343.6A patent/CN107103033B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070136753A1 (en) * | 2005-12-13 | 2007-06-14 | United Video Properties, Inc. | Cross-platform predictive popularity ratings for use in interactive television applications |
| CN103198418A (en) * | 2013-03-15 | 2013-07-10 | 北京亿赞普网络技术有限公司 | Application recommendation method and application recommendation system |
| CN103605656A (en) * | 2013-09-30 | 2014-02-26 | 小米科技有限责任公司 | Music recommendation method and device and mobile terminal |
| CN104317959A (en) * | 2014-11-10 | 2015-01-28 | 北京字节跳动网络技术有限公司 | Data mining method and device based on social platform |
Non-Patent Citations (1)
| Title |
|---|
| 李超 等: "基于用户相似性传递的跨平台交叉推荐算法", 《中文信息学报》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107103033B (en) | 2021-04-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20210141814A1 (en) | Concept-level user intent profile extraction and applications | |
| US20120174006A1 (en) | System, method, apparatus and computer program for generating and modeling a scene | |
| CN103514204B (en) | Information recommendation method and device | |
| Golbeck | Personalizing applications through integration of inferred trust values in semantic web-based social networks | |
| Woerndl et al. | Utilizing physical and social context to improve recommender systems | |
| Cleger-Tamayo et al. | Top-N news recommendations in digital newspapers | |
| Xu et al. | Integrated collaborative filtering recommendation in social cyber-physical systems | |
| Zheng et al. | Mapping the field of communication technology research in Asia: Content analysis and text mining of SSCI journal articles 1995–2014 | |
| US9256692B2 (en) | Clickstreams and website classification | |
| KR20160057475A (en) | System and method for actively obtaining social data | |
| US10496720B2 (en) | Systems and methods for providing feed preference surveys in a social networking system | |
| TW201503019A (en) | Method and system for discovery of user unknown interests | |
| Al-Yazidi et al. | Measuring reputation and influence in online social networks: a systematic literature review | |
| Gao et al. | SeCo-LDA: Mining service co-occurrence topics for composition recommendation | |
| McKenzie et al. | A thematic approach to user similarity built on geosocial check-ins | |
| Shin et al. | Stexnmf: Spatio-temporally exclusive topic discovery for anomalous event detection | |
| US10896213B2 (en) | Interface for a distributed network system | |
| Bolduc et al. | Examining the impact of social identification with music on music streaming behavior | |
| Liang et al. | SMS: A framework for service discovery by incorporating social media information | |
| Khan et al. | Assessing the long-term fragmentation of information systems research with a longitudinal multi-network analysis | |
| Gasparetti et al. | Community Detection and Recommender Systems. | |
| KR101423690B1 (en) | Contents recommendation system and method based on social network | |
| Rosa et al. | A weight-aware recommendation algorithm for mobile multimedia systems | |
| Inaam ul Haq et al. | Analyzing the research trends of IoT using topic modeling | |
| Kanavos et al. | Fuzzy information diffusion in twitter by considering user’s influence |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20191209 Address after: P.O. Box 31119, grand exhibition hall, hibiscus street, 802 West Bay Road, Grand Cayman, Cayman Islands Applicant after: Innovative advanced technology Co., Ltd Address before: A four-storey 847 mailbox in Grand Cayman Capital Building, British Cayman Islands Applicant before: Alibaba Group Holding Co., Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |