CN108647064A - The method and device of courses of action navigation - Google Patents
The method and device of courses of action navigation Download PDFInfo
- Publication number
- CN108647064A CN108647064A CN201810220308.5A CN201810220308A CN108647064A CN 108647064 A CN108647064 A CN 108647064A CN 201810220308 A CN201810220308 A CN 201810220308A CN 108647064 A CN108647064 A CN 108647064A
- Authority
- CN
- China
- Prior art keywords
- user
- path
- object run
- active user
- page
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/451—Execution arrangements for user interfaces
- G06F9/453—Help systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/30—Creation or generation of source code
- G06F8/31—Programming languages or programming paradigms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/445—Program loading or initiating
- G06F9/44505—Configuring for program initiating, e.g. using registry, configuration files
- G06F9/4451—User profiles; Roaming
Abstract
This specification embodiment provides a kind of method and apparatus of user's operation path navigation, method includes obtaining the sequence of operation and user characteristic data of active user first, then uses path navigation model, based on the sequence of operation and user characteristic data, the object run page of active user is predicted.Then, recommend the object run page to active user.Whether receive the recommended object run page depending on active user, determines the modification to path navigation model.In this way, user is helped to be rapidly achieved the object run page.
Description
Technical field
The method that this specification one or more embodiment is related to field of computer technology more particularly to courses of action navigation
And device.
Background technology
With the development of Internet technology, people usually use various client softwares, such as mobile phone terminal APP to carry out
Various issued transactions, including shopping, payment, social activity, financing etc..This allows for the interaction design of client software and user very
It is important.Complicated client software function is various, if interaction design is not reasonable, user cannot pass through simple and fast mode
It arrives at and performance objective operates, it will the usage experience of user is reduced, but also the service ability of product declines.In particular, right
Require higher operation in some timeliness, for example, in some Wealth management software products fund, stock exchange class operation,
If quickly the property loss for possibly even causing user cannot be operated performance objective.Also, on the other hand, different user has
Different operating habits and demand.Just for the user interaction design of client software, more stringent requirements are proposed for this.
Accordingly, it would be desirable to there is improved plan, by operating effectively path navigation, user is helped fast and efficiently to hold
Row object run.
Invention content
This specification one or more embodiment describes a kind of method and apparatus, is navigated by courses of action, helps to use
The quick performance objective operation in family.
According in a first aspect, provide a kind of method of user's operation path navigation, including：Obtain the operation of active user
Sequence and user characteristic data；Using path navigation model, it is based on the sequence of operation and user characteristic data, is worked as described in prediction
The object run page of preceding user；Recommend the object run page to the active user；Whether received according to active user
The object run page recommended determines the modification to the path navigation model.
In one embodiment, above-mentioned path navigation model is constructed to, the historical operating data based on multiple users
Collection and user characteristic data collection, establish the correspondence of multiple user groups and operation mode.
According to a kind of possible design, the object run page of active user is predicted in the following manner：According to the use
Family characteristic determines the corresponding particular group of the active user；According to multiple user groups of path navigation model foundation
With the correspondence of operation mode, the corresponding certain operational modes of particular group are determined；According to the certain operational modes, really
The fixed matched specific operation path of the sequence of operation with the active user；According to the specific operation path, work as described in determination
The object run page of preceding user.
According to a kind of possible design, the object run page is provided one of in the following manner：
Floating layer icon is provided on original interface, the floating layer icon is directed toward the object run page；
Increase icon newly in interface, which is directed toward the object run page；
The object run page is supplied to the active user as next operation pages.
According to a kind of embodiment, determination may include to the modification of path navigation model：Whether connect according to active user
By the object run page recommended, the recommendation accuracy of the path navigation model is determined；It is less than in the recommendation accuracy
In the case of predetermined threshold value, modify to the path navigation model.
Further, in one embodiment, to path navigation model modify including：Record the active user's
Subsequent operation；According to the sequence of operation and the subsequent operation, object run path is formed；According to the object run road
Diameter updates the correspondence of multiple user groups and operation mode that the path navigation model is established.
According to second aspect, a kind of device for user's operation path navigation is provided, including：Acquiring unit is configured to
Obtain the sequence of operation and user characteristic data of active user；Predicting unit is configured to use path navigation model, based on described
The sequence of operation and user characteristic data predict the object run page of the active user；Recommendation unit is configured to work as to described
Preceding user recommends the object run page；And determination unit, it is configured to whether receive recommended mesh according to active user
Operation pages are marked, determine the modification to the path navigation model.
According to the third aspect, a kind of computer readable storage medium is provided, computer program is stored thereon with, when described
When computer program executes in a computer, enable computer execute first aspect method.
According to fourth aspect, a kind of computing device, including memory and processor are provided, which is characterized in that described to deposit
It is stored with executable code in reservoir, when the processor executes the executable code, the method for realizing first aspect.
The method and device provided by this specification embodiment recommends object run using path navigation model to user
The page, while according to user feedback constantly training and improving the model so that the recommendation accuracy of model is continuously improved, to more
Big degree user is helped to be rapidly achieved target pages, simplifies user's operation, promote user experience.
Description of the drawings
In order to illustrate the technical solution of the embodiments of the present invention more clearly, required use in being described below to embodiment
Attached drawing be briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for this
For the those of ordinary skill of field, without creative efforts, others are can also be obtained according to these attached drawings
Attached drawing.
Fig. 1 is the implement scene schematic diagram of one embodiment that this specification discloses；
Fig. 2 shows the schematic diagrames of the courses of action of user in one example；
Fig. 3 shows the flow chart of the method for the user's operation path navigation according to one embodiment；
Fig. 4 shows the flow chart of the prediction object run page according to one embodiment；
Fig. 5 shows the schematic diagram of the newly-increased icon according to one embodiment；
Fig. 6 shows the schematic diagram of the floating layer icon according to one embodiment.
Fig. 7 shows the schematic block diagram of the courses of action navigation device according to one embodiment.
Specific implementation mode
Below in conjunction with the accompanying drawings, the scheme provided this specification is described.
Fig. 1 is the implement scene schematic diagram of one embodiment that this specification discloses.As shown, each user is by making
With client software, such as Alipay App, ant wealth App etc. can form user's history operation information.Computing platform, example
Such as the server of above-mentioned client software, these user's history operation informations and the characteristic information of user, structure can be based on
One path navigation model.The path navigation model is obtained each by analyzing above-mentioned historical operation information and user's characteristic information
The use habit of class user.It then, can be while carrying out user's operation path navigation using the path navigation model, to this
Model is tested and is trained.When getting the sequence of operation that user A is operated into the client software, will capture
The sequence of operation and the user's characteristic information of the user A input above-mentioned path navigation model.Before path navigation model is based on
The next object run pages of the user A are predicted in analysis to all types of user use habit.Then, computing platform is by the mesh
Mark operation pages recommend user.Then, whether the object run recommended has been selected according to user A, to judge object run
Prediction accuracy, using the prediction accuracy as feedback, change path navigation model.Process in this way, uses on one side
The model carries out the recommendation of object run to user, changes model as feedback result according to the user's choice on one side, to
The accuracy that model is recommended is continuously improved, preferably recommends object run for user, user is helped to efficiently reach object run.Under
Face describes the specific implementation step of above procedure.
A kind of embodiment for illustrating description according to this, in the first stage, for user's history operation data collection and user
Characteristic information data collection establishes path navigation model.In the modelling phase, the information at least two parties face is obtained：User's history operates
Message data set and user's characteristic information data set.Path navigation model is established for the information of these two aspects, which is used for
Above-mentioned user's history operation information and user's characteristic information are analyzed and learnt, show that the operation of all types of user is practised with this
It is used.
User's history operation information includes user's history operation trace, or is historical operation path.In order to obtain in this way
Information usually in user when being operated on terminal interface, capture the operation behavior and related pages of user.
The operation behavior of user may include the various operations that user carries out the page, such as clicks, sliding, pulls, character key entry, puts
Big or diminution etc..The above-mentioned page can be Webpage, Web five application pages, the App pages or other can be handed over user
The mutual page.It may be used and bury mode a little in terminal applies to capture the operation of user.For example, in advance to each in the page
Can interactive module, such as button, scroll bar, banner (web advertisement) etc., be numbered, in order to record user's operation.
Fig. 2 shows the schematic diagrames of the courses of action of user in one example.In fig. 2, it by taking Alipay as an example, shows to use
Operation of the family to the page.As shown in Fig. 2, user since Alipay homepage A1, clicks " ant wealth " (executing operation P1),
Into page A2.In the A2 pages, " financing " (executing operation P2) is clicked, into financing page A3.In the A3 pages further
Show various financing projects.It is assumed that user clicks " fund " (executing operation P3) in the A3 pages, hence into page A4, wherein
Various fund selections are shown.In page A4, user clicks " fund ranking " (executing operation P4) again, then enters the page
A5 shows page A6 in page A5 by click " valuation seniority among brothers and sisters " (executing operation P5).The content of user's browsing pages A6
Later, client is closed.In this way, A1.P1-A2.P2-A3.P3-A4.P4-A5.P5-A6 shows the operation road of a user
Diameter OP1.
In one embodiment, while capturing user's operation, relevant temporal information is also recorded.In this way, user goes through
History operation information further includes the time parameters information such as operation moment, residence time.For example, for courses of action shown in Fig. 3
OP1：A1.P1-A2.P2-A3.P3-A4.P4-A5.P5-A6 can also record user and execute each operation P1-P2-P3-P4-P5
At the time of, such as T1-T2-T3-T4-T5 and user are in the residence time of each page, such as ST1-ST2-ST3-ST4-
ST5-ST6.It is also based on operation moment and/or residence time, determines the information such as the time interval of operation.
In one embodiment, above user's history operation information can be recorded as a variety of shapes such as sequence, vector, matrix
Formula.For example, in one example, a courses of action are recorded as a vector, each vector element corresponds to a user behaviour
Make it is targeted can interactive module number.In one example, courses of action and the courses of action corresponding time are believed
Breath arranges the sequence for equal length, and it is matrix form that these sequences, which are arranged,.It is appreciated that other lattice can also be used
Formula or form record user's history operation information.
In addition to user's history operation information, on the other hand, the characteristic information of user is also obtained.In one embodiment, it uses
Family characteristic information includes the essential attribute information of user, such as attribute information included in log-on message, such as age, property
Not, region etc..In one embodiment, user's characteristic information further includes, with the relevant user of client software function other
Attribute information.For example, for the client software of financial class, such as Alipay, ant wealth can also obtain the remaining of user
The information such as the total net value of volume, fund, the transaction frequency are as its characteristic information.
In one embodiment, user's characteristic information includes user's portrait information.User's portrait information is according to user
Basic attribute data and the user model of a labeling taken out.Certain methods are had existed, the basic category based on user
Property data (such as log-on message, consumer record etc.), assign some labels for user, the set of these labels constitutes user's portrait
Information.For example, in a specific example, user's portrait information includes：" registration new person ", " financing little Bai ", " savings expert ",
" medium anti-risk grade ", label informations such as " full-time mother ", wherein the label of " registration new person " is according in user's registration information
Registion time and determine, about financing, savings, the ability to ward off risks label according to such as remaining sum of user, financing income, storage
The informixes such as grade, consumer record are stored to determine.
In one embodiment, the basic attribute data of user is obtained, the above method is then used, basic based on these
Attribute data determines user's portrait information.In another embodiment, client software is for other functional purposes, for
User depicts user's portrait.In this case, it directly acquires such user and draws a portrait information as user's characteristic information.
It is appreciated that in order to analyze the operating habit of all types of user group, need to obtain the respective history behaviour of a large number of users
Make information and user's characteristic information.Path navigation model is established for the information of these two aspects, which is used for above-mentioned user
Historical operation information and user's characteristic information are analyzed and are learnt, and the operating habit of all types of user is obtained with this.
According to one embodiment, in above-mentioned path navigation model use clustering algorithm, to a large number of users information into
Row classification analysis, and the quick recurrence of lift scheme.For example, can be clustered to user's characteristic information, to will largely use
Family is divided into multiple user groups according to certain particle size, to build its operation behavior model for each user group.Can also to
Family historical operation information carries out clustering, to be some operation behavior modes by complicated user's operation behavior processing.
In cluster process, the direct partition clustering method of such as K-means can be taken, the cluster based on level is calculated
Method, such as BIRCH algorithms, CURE algorithms etc., density-based algorithms, such as DBSCAN algorithms, OPTICS algorithms etc.
Deng and other possible clustering algorithms.
According to one embodiment, the algorithm and structure of neural network, such as convolution may be used in above-mentioned path navigation model
Neural network CNN, Recognition with Recurrent Neural Network RNN etc..The algorithm and structure of above-mentioned neural network may be used in path navigation model, right
Input data, i.e. the user's history operation information and user's characteristic information of a large number of users carry out deep learning.
By using various algorithms, path navigation model is based on a large number of users historical operation information and user's characteristic information,
The custom that operates with for analyzing and learning each user group, that is, establish the correspondence of multiple user groups and operation mode.This
The process of analysis and study can specifically include, and count the frequency in various historical operation paths, operating time, according to historical operation
Path and operating time parameter form multiple modes of operation, count the correspondence of each user group and operation mode, to for
Each user group builds operation behavior model.More specifically, above-mentioned user group can be the group set according to various granularities
Body, for example, in one example, the dividing of user group can draw a portrait with user in include crowd's label it is corresponding to the same；
In another example, the division of user group can correspond to the combination of various crowd's labels in user's portrait, for example user group 1 corresponds to
In having " financing little Bai " simultaneously, the user of the labels such as " savings expert ", user group 2 corresponds to while having " financing master-hand ",
The user of labels such as " stock market's introductions ".In another example, the division of user group can be based on independent division rule.User group
Granularity of division can be arranged according to business demand, in one example, user group can include even only a use
Family.
On the other hand, operation mode may include the historical operation path that the frequency is more than certain threshold value.Further, it operates
Pattern can also include the time dependent parameter corresponding to historical operation path, such as operate the moment, residence time etc..
In a specific example, user user1 continually (such as more than a threshold value, 10 times) can execute Fig. 2 institutes daily
The courses of action OP1 shown：A1.P1-A2.P2-A3.P3-A4.P4-A5.P5-A6 and another courses of action OP2：A1.P1-
A2.P2-A3.P7-A8.P8-A9.P9-A10.Such information is recorded as historical operation information.Path navigation model
The characteristic information of analysis user user1 first, user group G1 is divided by user user1.Also, path navigation model is by dividing
Historical operation information is analysed, the frequency of occurrence of various courses of action is especially counted, it can be found that the high frequency behaviour of above-mentioned user user1
Make path OP1 and OP2.The courses of action of such high frequency time can form a kind of operation mode M1.More specifically, path navigation
Model is by analyzing the operating time of the user user1 it can be found that daily 10 points or so of user user1 is averagely executed on 5 times
The courses of action OP1 stated, daily 12 noon or so averagely execute 6 above-mentioned courses of action OP2,11 points or so of every night
Averagely execute 7 above-mentioned courses of action OP1.Based on such information, path navigation model can be established belonging to user user1
User group G1 and operation mode M1 correspondence, in operation mode M1 comprising above-mentioned high-frequency operation path OP1 and
OP2.Further, in one embodiment, operation mode M1 can also include the time dependent parameter of each courses of action, example
As high frequency executes period (OP1：10：00,23:00；OP2:12:00), residence time etc..
In this way, historical operating data collection and user characteristic data collection of the path navigation model based on a large number of users, are established more
The correspondence of a user group and operation mode.The path navigation model built in this way can be used for carrying out path to user's operation
Prediction and navigation.Fig. 3 shows the flow chart of the method for the user's operation path navigation according to one embodiment.As shown in figure 3, should
Method flow includes the following steps：Step 31, the sequence of operation and user characteristic data of active user are obtained；Step 32, it uses
Path navigation model is based on the sequence of operation and user characteristic data, predicts the object run page of the active user；Step
Rapid 33, recommend the object run page to the active user；Step 34, whether receive recommended mesh according to active user
Operation pages are marked, determine the modification to the path navigation model.The specific implementation procedure of above each step is described below.
First, in step 31, the sequence of operation and user characteristic data of active user are obtained.In order to be different from aforementioned modeling
The description in stage is described with specific user, user A as active user in example below.It will be understood, however, that
Active user A can be that the historical data being based in the path navigation model modeling stage concentrates the historical user for including, can also
It is the new user for being not included in historical data concentration.
Similar with historical operation information is obtained, the modes such as a little that can be buried by the page capture the sequence of operation SA of user A.
Sequence of operation SA can only include an operation, can also include multinomial continuous operation.These operation can be include pair
In the page can the click of interactive module, sliding, dragging, character key in, zoom in or out etc..In a specific example,
Step 31 captures, and active user A clicks " ant wealth " (executing operation P1), into page since Alipay homepage A1
Face A2 clicks " financing " (executing operation P2), into financing page A3 then in the A2 pages.It is thusly-formed sequence of operation SA：
A1.P1-A2.P2-A3.In another example, user A continues to operate, and sequence of operation SA is updated to：A1.P1-A2.P2-
A3.P3-A4。
On the other hand, the user characteristic data of active user A is also obtained.
In one embodiment, the attribute data of active user A is obtained as user characteristic data, and above-mentioned attribute data can
To include, data in log-on message, such as age, gender, registion time etc., and with client major function it is relevant its
The data such as remaining sum, income in his attribute data, such as financing client.
In another embodiment, the attribute data of the user A of acquisition is converted to user's representation data as user characteristics
Data.Specifically, the attribute value (such as age, registion time, remaining sum, income etc.) of each attribute of user A is obtained first,
Crowd's label (such as " registration new person ", " financing little Bai ", " savings expert " etc. belonging to user A are determined according to these attribute values
Deng), determine the representation data of user A as the user characteristic data according to crowd's label.
In another embodiment, client software depicts user for each user and draws for other functional purposes
Picture.In this case, the representation data of active user A is directly acquired as its user's characteristic information.
Then, in step 32, using path navigation model, the sequence of operation and user characteristics number of the user A based on capture
According to predicting the object run page of active user.As previously mentioned, having had been built up path navigation model, which is based on multiple
The historical operating data collection and user characteristic data collection of user, establishes the correspondence of multiple user groups and operation mode.In this way,
Such path navigation model may be used, based on the correspondence established, the object run page of prediction active user A.
Fig. 4 shows the flow chart of the prediction object run page according to one embodiment, that is, the sub-step of step 32
Suddenly.As shown in figure 4, in step 321, according to the user characteristic data of active user, the corresponding specific use of the active user is determined
Family group.
Specifically, for active user A, if the historical data that the path navigation model modeling stage is based on is concentrated
Including user's (such as user A practically equals to user user1 above-mentioned), then can be directly according to the user of modelling phase
Group divides, and determines the user group belonging to user A, i.e. user group G1.
If historical data concentrate do not include the user A, can based on the user characteristic data of user A, and
The user group division rule in path navigation model modeling stage, determines the user group belonging to user A.For example, the use of modelling phase
The dividing of family group can draw a portrait with user in include crowd's label it is corresponding to the same, or can correspond to each in user's portrait
The combination of kind crowd's label.In this case, the user characteristic data for being primarily based on user A obtains its user portrait number
According to obtain crowd's label of user A, according to these crowd's labels and user group division rule, determining the use belonging to user A
Family group.In other examples, the modelling phase is also based on other user group division rules.Here it is assumed that according to user A's
User characteristic data and user group division rule determine that user A belongs to user group G1.
On the basis of the particular group belonging to user A is determined, in step 322, built according to path navigation model
The correspondence of vertical multiple user groups and operation mode determines the corresponding certain operational modes of the particular group.
As previously mentioned, path navigation model is by analyzing and learning the historical operating data collection and user characteristics of a large number of users
Data set establishes the correspondence of multiple user groups and operation mode.The user group belonging to user A is determined in step 321
On the basis of G1, the operation mode M1 corresponding to user group G1 can be determined by the correspondence having built up.
Then, in step 323 the matched spy of the sequence of operation with active user is determined according to above-mentioned certain operational modes
Determine courses of action.Usually, operation mode may include the historical operation path that the frequency is more than certain threshold value.Further, it grasps
Operation mode can also include the time dependent parameter corresponding to historical operation path, such as operate the moment, residence time etc..
In one embodiment, in step 323, the historical operation path that above-mentioned certain operational modes include is obtained, therefrom determines and works as
The matched courses of action of the sequence of operation of preceding user are as above-mentioned specific operation path.The above-mentioned sequence of operation with active user
Matched courses of action can contain the courses of action of the sequence of operation.In one embodiment, it can also further examine
Consider the time dependent parameter in each historical operation path for including in operation mode.For example, having in certain operational modes multiple
In the case that historical operation path includes current operation sequence, it may be determined that the current operating temporal of current operation sequence, with
And the historic execution time in each historical operation path recorded in operation mode, it will wherein historic execution time and current operation
Time, immediate historical operation path was determined as the specific operation path with current operation sequences match.
For the user group G1 in previous example, corresponding operation mode M1 includes at least history high-frequency operation road
Diameter OP1：A1.P1-A2.P2-A3.P3-A4.P4-A5.P5-A6 and OP2： A1.P1-A2.P2-A3.P7-A8.P8-
A9.P9-A10。
In one example, it is assumed that the current operation sequence SA of user A includes A1.P1-A2.P2-A3.P3-A4.Pass through
The element of current operation sequence is compared with historical operation path, it can be seen that historical operation path OP1 is contained currently
Sequence of operation SA, thus OP1 can as with the matched specific operation paths SA.
In another example, it is assumed that the current operation sequence SA of user A includes A1.P1-A2.P2-A3.It can be seen that behaviour
The historical operation path OP1 and OP2 that operation mode M1 includes contain current operation sequence SA, therefore OP1 and OP2 can be with
Alternately courses of action.It is possible to further the time dependent parameter in reference operation pattern M1, such as when high frequency execution
Section.It is assumed that the operating time of current operation sequence SA is 10:The history of 05, operation mode M1 record executes the period and includes：OP1:
10:00 and 23:00, OP2:12:00.Current operating temporal 10:05 closer to historical operation path OP1 high frequency execute when
Between, therefore, OP1 is determined as and the matched specific operation paths current operation sequence SA.
Then, the object run page of active user is determined according to specific operation path determined above in step 324.
Specifically, by the end page in specific operation path, the object run page as active user.In the preceding example, false
Surely it is determined that OP1 is specific operation path, then can be by the end page A6 of OP1 (corresponding to " valuation seniority among brothers and sisters " in Fig. 2
The page), the object run page as user A.
In this way, by mode shown in Fig. 4, in step 32, using path navigation model, the target of prediction active user A
Operation pages.Fig. 3 is returned to, next, in step 33, recommends the identified object run page to active user.
In one embodiment, the identified object run page is supplied to active user as next operation pages.
For example it is assumed that having determined that specific operation path is OP1, the page A6 that terminates of OP1 is the object run page of user A, and uses
The current operation page of family A is A3.If the operation that user A is carried out on page A3, such as click " fund " in page A3 and (hold
Row operation P3), continue to meet specific operation path OP1, is directly user's A display targets then skipping intermediate page A4 and A5
Operation pages A6.
In one embodiment, in order to avoid interfering the operation of user, icon is increased newly in client end interface, this is newly-increased
Icon is directed toward the object run page.Fig. 5 shows the schematic diagram of the newly-increased icon according to one embodiment.As shown in figure 5, in " branch
Pay treasured " icon " quick ", the icon direction object run page are increased newly in the handle icon of the bottom row of client.It is assumed that
Page A6 is confirmed as the object run page of user A, then in the case of fig. 5, user A is in current operation sequence process
In, by clicking the icon " quick ", page A6 can be jumped directly to, the intermediate page, such as A4 and A5 are save.
In another embodiment, in order to be further reduced the modification to client end interface, floating layer is provided on original interface
Icon, the floating layer icon are directed toward the object run page.Fig. 6 shows the schematic diagram of the floating layer icon according to one embodiment.Such as
Shown in Fig. 6, on the basis of " Alipay " original interface, floating layer icon " understanding that you are thought " is increased, which is directed toward target behaviour
Make the page.In one example, floating layer icon can be designed as translucent, to influence the display at original interface as little as possible.
In the preceding example, it is assumed that page A6 is confirmed as the object run page of user A, then in the case of fig. 7, user A exists
Under any page of current operation sequence, page A6 can be jumped directly to by clicking the floating layer icon " understanding that you are thought ",
Save intermediate page.
It is appreciated that by other means, such as mode can also be pushed, recommend identified object run page to user
Face.
By way of recommending the object run page to user, can helping user, fast and efficiently performance objective operates,
Target pages are reached, user experience is promoted.
At the same time it can also be used as feedback to the situation that receives of the target pages of recommendation according to user, further training, modification
Path navigation model makes it more meet the use habit of user.It is, in step 34, whether institute is received according to active user
The object run page of recommendation determines the modification to path navigation model.
In one embodiment, it is first determined whether active user receives the recommended object run page.Aforementioned
In example, object run page A6 is recommended to user A by various modes.If user A is entering the object run recommended
It after page A6, is performed within the very short predetermined time (such as 1s) and exits operation, such as jumped by clicking other icons
Go out the page, or operation return key exits the page, or close entire client software, then can consider user A not
There is selection or receives the object run page of the recommendation.If user A after entering the object run page recommended, is stopped
The rational time either performs the operations such as browsing, further interaction so it is considered that user A is selected or received this
The object run page of recommendation.
Whether receive the recommended object run page according to active user, it may be determined that the recommendation of path navigation model is accurate
Exactness.Recommend accuracy that can be fed back to the receiving of the object run page of recommendation based on multiple active users and counts and obtain.
For example, after recommending the object run page to 100 users, it is assumed that wherein 40 users receive recommendation, then can recognize
It is 40% to recommend accuracy.For each active user, situation can be received to recommendation according to it, to recommend accuracy into
Row is cumulative, to which accuracy is recommended in dynamic update.For example, assuming that active user is user A, and 20 users of warp-wise before
The recommendation of the object run page is carried out, wherein 9 users receive recommendation, it is 9/20 to recommend accuracy.If current use
Family A receives the object run page of recommendation, accuracy can will be recommended to be updated to 10/21；If user A does not receive recommendation
The object run page, can by recommend accuracy be updated to 9/21.
In the case where the recommendation accuracy is less than predetermined threshold value, can modify to path navigation model.More
Body, can record the subsequent operation of active user, i.e., from the object run page of recommendation exit after operation.It is appreciated that
Such subsequent operation is to reflect the operation of active user's realistic objective.According to the sequence of operation before the active user and upper
Subsequent operation is stated, object run path is formed, then according to the object run path, update path navigation model is established more
The correspondence of a user group and operation mode.More specifically, in one embodiment, above-mentioned object run path is added to
Historical operating data is concentrated, and forms newer historical operating data collection, and path navigation model is grasped for newer history
Analytic learning is re-started as data set, to re-establish the correspondence of user group and operation mode.In another embodiment
In, according to above-mentioned object run path, update the user group (such as user group G1) belonging to active user (such as user A)
Operation mode (such as operation mode M1).Specifically, using above-mentioned object run path as newly-increased historical operation path, for
User group belonging to active user updates the statistics to historical operation path, to update the corresponding operation mode of the user group
Content.
It is appreciated that some conventional algorithms during model training can also be used, such as gradient transmission etc., change mould
The other parameters of type so that model preferably returns, and is promoted according to user feedback forward direction and recommend accuracy.
In this way, passage path navigation model realizes the positive feedback of closed loop.Recommending the same of the object run page to user
When, constantly train and improve the model so that the recommendation accuracy of model is continuously improved, to help user fast to a greater degree
Speed reaches target pages, simplifies user's operation, promotes user experience.
According to the embodiment of another aspect, a kind of device for user's operation path navigation is also provided.Fig. 7 shows root
According to the schematic block diagram of the courses of action navigation device of one embodiment.As shown in fig. 7, the navigation device 700 includes：It obtains single
Member 71 is configured to obtain the sequence of operation and user characteristic data of active user；Predicting unit 72 is configured to use path navigation
Model is based on the sequence of operation and user characteristic data, predicts the object run page of the active user；Recommendation unit
73, it is configured to recommend the object run page to the active user；And determination unit 74, it is configured to according to active user
Whether receive the recommended object run page, determines the modification to the path navigation model.
In one embodiment, above-mentioned acquiring unit 71 is configured to, and obtains at least one attribute of the active user
Attribute value determines at least one crowd's label belonging to the active user, root according to the attribute value of at least one attribute
Determine the representation data of the active user as the user characteristic data according at least one crowd's label.
According to one embodiment, above-mentioned path navigation model is constructed to, the historical operating data based on multiple users
Collection and user characteristic data collection, establish the correspondence of multiple user groups and operation mode.
In one embodiment, above-mentioned predicting unit 72 includes (not shown)：First determining module is configured to according to
User characteristic data determines the corresponding particular group of the active user；Second determining module, is configured to according to the multiple
The correspondence of user group and operation mode determines the corresponding certain operational modes of the particular group；Third determining module,
It is configured to, according to the certain operational modes, determine the matched specific operation path of the sequence of operation with the active user；With
And the 4th determining module, it is configured to determine the object run page of the active user according to the specific operation path.
Further, in one embodiment, above-mentioned third determining module is configured to：From the certain operational modes,
Determine the multiple historical operation paths for including the sequence of operation；Determine the current operating temporal of the sequence of operation, Yi Jisuo
State multiple respective historic execution times in historical operation path；Historic execution time and the current operating temporal is immediate
Historical operation path is determined as the specific operation path.
According to a kind of embodiment, above-mentioned recommendation unit 73 is configured to, and provides the target behaviour one of in the following manner
Make the page：Floating layer icon is provided on original interface, the floating layer icon is directed toward the object run page；It is increased newly in interface
Icon, the newly-increased icon are directed toward the object run page；The object run page is provided as next operation pages
To the active user.
In one embodiment, above-mentioned determination unit 74 includes (not shown)：Accuracy determining module is configured to basis and works as
Whether preceding user receives the recommended object run page, determines the recommendation accuracy of the path navigation model；Modified module,
It is configured to, in the case where the recommendation accuracy is less than predetermined threshold value, modify to the path navigation model.
Further, in one embodiment, the modified module is configured to：Record the follow-up behaviour of the active user
Make；According to the sequence of operation and the subsequent operation, object run path is formed；According to the object run path, update
The correspondence for the multiple user groups and operation mode that the path navigation model is established.Further, implement at one
In example, above-mentioned modified module is additionally configured to：Using the object run path as newly-increased historical operation path, work as described
The statistics in the particular group update historical operation path belonging to preceding user；It is counted according to newer historical operation path, update
The corresponding operation mode of the particular group.
By above device, the object run page is recommended to user using path navigation model, while anti-according to user
The model is constantly trained and is improved in feedback so that the recommendation accuracy of model is continuously improved, to help user fast to a greater degree
Speed reaches target pages, simplifies user's operation, promotes user experience.
According to the embodiment of another aspect, a kind of computer readable storage medium is also provided, is stored thereon with computer journey
Sequence enables computer execute and combines method described in Fig. 3 when the computer program executes in a computer.
According to the embodiment of another further aspect, a kind of computing device, including memory and processor, the memory are also provided
In be stored with executable code, when the processor executes the executable code, realize the method in conjunction with described in Fig. 3.
Those skilled in the art are it will be appreciated that in said one or multiple examples, work(described in the invention
It can be realized with hardware, software, firmware or their arbitrary combination.It when implemented in software, can be by these functions
Storage in computer-readable medium or as on computer-readable medium one or more instructions or code be transmitted.
Above-described specific implementation mode has carried out further the purpose of the present invention, technical solution and advantageous effect
It is described in detail, it should be understood that the foregoing is merely the specific implementation mode of the present invention, is not intended to limit the present invention
Protection domain, all any modification, equivalent substitution, improvement and etc. on the basis of technical scheme of the present invention, done should all
Including within protection scope of the present invention.
Claims (20)
1. a kind of method of user's operation path navigation, including：
Obtain the sequence of operation and user characteristic data of active user；
Using path navigation model, it is based on the sequence of operation and user characteristic data, predicts the target behaviour of the active user
Make the page；
Recommend the object run page to the active user；
Whether receive the recommended object run page according to active user, determines the modification to the path navigation model.
2. according to the method described in claim 1, it includes obtaining the active user's wherein to obtain the user characteristic data
The attribute value of at least one attribute determines at least one belonging to the active user according to the attribute value of at least one attribute
A crowd's label determines the representation data of the active user as the user characteristics according at least one crowd's label
Data.
3. according to the method described in claim 1, the wherein described path navigation model is constructed to, based on going through for multiple users
History operation data collection and user characteristic data collection, establish the correspondence of multiple user groups and operation mode.
4. according to the method described in claim 3, the object run page of the wherein described prediction active user includes：
According to the user characteristic data, the corresponding particular group of the active user is determined；
According to the correspondence of the multiple user group and operation mode, the corresponding specific operation mould of the particular group is determined
Formula；
According to the certain operational modes, the matched specific operation path of the sequence of operation with the active user is determined；
According to the specific operation path, the object run page of the active user is determined.
5. according to the method described in claim 4, wherein determining the behaviour with the active user according to the certain operational modes
Make the matched specific operation path of sequence, including：
From the certain operational modes, the multiple historical operation paths for including the sequence of operation are determined；
When determining that the respective history of current operating temporal and the multiple historical operation path of the sequence of operation executes
Between；
Historic execution time and the immediate historical operation path of the current operating temporal are determined as the specific operation road
Diameter.
6. according to the method described in claim 1, it includes passing through wherein to recommend the object run page to the active user
One of following manner provides the object run page：
Floating layer icon is provided on original interface, the floating layer icon is directed toward the object run page；
Increase icon newly in interface, which is directed toward the object run page；
The object run page is supplied to the active user as next operation pages.
7. according to the method described in claim 3, wherein whether receiving the recommended object run page according to active user, really
It is fixed to include to the modification of the path navigation model：
Whether receive the recommended object run page according to active user, determines that the recommendation of the path navigation model is accurate
Degree；
In the case where the recommendation accuracy is less than predetermined threshold value, modify to the path navigation model.
8. according to the method described in claim 7, wherein to the path navigation model modify including：
Record the subsequent operation of the active user；
According to the sequence of operation and the subsequent operation, object run path is formed；
According to the object run path, pair of multiple user groups and operation mode that the path navigation model is established is updated
It should be related to.
9. according to the method described in claim 8, wherein updating the multiple user groups and behaviour that the path navigation model is established
The correspondence of operation mode, including：
Using the object run path as newly-increased historical operation path, for the particular group belonging to the active user
Update the statistics in historical operation path；
It is counted according to newer historical operation path, updates the corresponding operation mode of the particular group.
10. a kind of device for user's operation path navigation, including：
Acquiring unit is configured to obtain the sequence of operation and user characteristic data of active user；
Predicting unit is configured to use path navigation model, is based on the sequence of operation and user characteristic data, works as described in prediction
The object run page of preceding user；
Recommendation unit is configured to recommend the object run page to the active user；
Determination unit is configured to whether receive the recommended object run page according to active user, and the path is led in determination
The modification of model plane type.
11. device according to claim 10 obtains the active user at least wherein the acquiring unit is configured to
The attribute value of one attribute determines at least one people belonging to the active user according to the attribute value of at least one attribute
Group's label, determines the representation data of the active user as the user characteristics number according at least one crowd's label
According to.
12. device according to claim 10, wherein the path navigation model is constructed to, based on multiple users'
Historical operating data collection and user characteristic data collection, establish the correspondence of multiple user groups and operation mode.
13. device according to claim 12, wherein the predicting unit includes：
First determining module is configured to, according to the user characteristic data, determine the corresponding particular group of the active user；
Second determining module is configured to the correspondence according to the multiple user group and operation mode, determines the specific use
The corresponding certain operational modes of family group；
Third determining module is configured to according to the certain operational modes, and determination is matched with the sequence of operation of the active user
Specific operation path；
4th determining module is configured to determine the object run page of the active user according to the specific operation path.
14. device according to claim 13, wherein the third determining module is configured to：
From the certain operational modes, the multiple historical operation paths for including the sequence of operation are determined；
When determining that the respective history of current operating temporal and the multiple historical operation path of the sequence of operation executes
Between；
Historic execution time and the immediate historical operation path of the current operating temporal are determined as the specific operation road
Diameter.
15. device according to claim 10, wherein the recommendation unit is configured to, institute is provided one of in the following manner
State the object run page：
Floating layer icon is provided on original interface, the floating layer icon is directed toward the object run page；
Increase icon newly in interface, which is directed toward the object run page；
The object run page is supplied to the active user as next operation pages.
16. device according to claim 12, wherein the determination unit includes：
Accuracy determining module is configured to whether receive the recommended object run page according to active user, determines the road
The recommendation accuracy of diameter navigation model；
Modified module, be configured to the recommendation accuracy be less than predetermined threshold value in the case of, to the path navigation model into
Row modification.
17. device according to claim 16, wherein the modified module is configured to：
Record the subsequent operation of the active user；
According to the sequence of operation and the subsequent operation, object run path is formed；
According to the object run path, pair of multiple user groups and operation mode that the path navigation model is established is updated
It should be related to.
18. device according to claim 17, wherein the modified module is further configured to：
Using the object run path as newly-increased historical operation path, for the particular group belonging to the active user
Update the statistics in historical operation path；
It is counted according to newer historical operation path, updates the corresponding operation mode of the particular group.
19. a kind of computer readable storage medium, is stored thereon with computer program, when the computer program in a computer
When execution, computer perform claim is enabled to require the method for any one of 1-9.
20. a kind of computing device, including memory and processor, which is characterized in that be stored with executable generation in the memory
Code when the processor executes the executable code, realizes the method described in any one of claim 1-9.
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810220308.5A CN108647064A (en) | 2018-03-16 | 2018-03-16 | The method and device of courses of action navigation |
| TW108100199A TW201938993A (en) | 2018-03-16 | 2019-01-03 | Operation path navigation |
| PCT/CN2019/070617 WO2019174379A1 (en) | 2018-03-16 | 2019-01-07 | Operation path navigation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810220308.5A CN108647064A (en) | 2018-03-16 | 2018-03-16 | The method and device of courses of action navigation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108647064A true CN108647064A (en) | 2018-10-12 |
Family
ID=63744217
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810220308.5A Pending CN108647064A (en) | 2018-03-16 | 2018-03-16 | The method and device of courses of action navigation |
Country Status (3)
| Country | Link |
|---|---|
| CN (1) | CN108647064A (en) |
| TW (1) | TW201938993A (en) |
| WO (1) | WO2019174379A1 (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109492831A (en) * | 2018-12-20 | 2019-03-19 | 平安科技（深圳）有限公司 | Recommended method, device, medium and the electronic equipment in marketing path |
| CN110032597A (en) * | 2018-11-30 | 2019-07-19 | 阿里巴巴集团控股有限公司 | The visible processing method and device of application program operation behavior |
| WO2019174379A1 (en) * | 2018-03-16 | 2019-09-19 | 阿里巴巴集团控股有限公司 | Operation path navigation |
| CN110750718A (en) * | 2019-10-10 | 2020-02-04 | 深圳Tcl新技术有限公司 | User guiding method, device, equipment and storage medium based on recommendation system |
| CN111767465A (en) * | 2020-07-09 | 2020-10-13 | 中国工商银行股份有限公司 | Information recommendation method and device, computing equipment and medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107728874A (en) * | 2017-09-06 | 2018-02-23 | 阿里巴巴集团控股有限公司 | The method, apparatus and equipment of user prompt operation are provided |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070121653A1 (en) * | 2005-11-04 | 2007-05-31 | Reckamp Steven R | Protocol independent application layer for an automation network |
| CN106294614B (en) * | 2016-08-01 | 2019-09-20 | 华为技术有限公司 | Method and apparatus for accessing business |
| CN108647064A (en) * | 2018-03-16 | 2018-10-12 | 阿里巴巴集团控股有限公司 | The method and device of courses of action navigation |
-
2018
- 2018-03-16 CN CN201810220308.5A patent/CN108647064A/en active Pending
-
2019
- 2019-01-03 TW TW108100199A patent/TW201938993A/en unknown
- 2019-01-07 WO PCT/CN2019/070617 patent/WO2019174379A1/en active Application Filing
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107728874A (en) * | 2017-09-06 | 2018-02-23 | 阿里巴巴集团控股有限公司 | The method, apparatus and equipment of user prompt operation are provided |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019174379A1 (en) * | 2018-03-16 | 2019-09-19 | 阿里巴巴集团控股有限公司 | Operation path navigation |
| CN110032597A (en) * | 2018-11-30 | 2019-07-19 | 阿里巴巴集团控股有限公司 | The visible processing method and device of application program operation behavior |
| CN109492831A (en) * | 2018-12-20 | 2019-03-19 | 平安科技（深圳）有限公司 | Recommended method, device, medium and the electronic equipment in marketing path |
| CN109492831B (en) * | 2018-12-20 | 2023-12-29 | 平安科技（深圳）有限公司 | Marketing path recommendation method and device, medium and electronic equipment |
| CN110750718A (en) * | 2019-10-10 | 2020-02-04 | 深圳Tcl新技术有限公司 | User guiding method, device, equipment and storage medium based on recommendation system |
| WO2021068676A1 (en) * | 2019-10-10 | 2021-04-15 | 深圳Tcl新技术有限公司 | Recommendation system-based user guiding method, apparatus, device and storage medium |
| CN111767465A (en) * | 2020-07-09 | 2020-10-13 | 中国工商银行股份有限公司 | Information recommendation method and device, computing equipment and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2019174379A1 (en) | 2019-09-19 |
| TW201938993A (en) | 2019-10-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20220198289A1 (en) | Recommendation model training method, selection probability prediction method, and apparatus | |
| US10599449B1 (en) | Predictive action modeling to streamline user interface | |
| CN108647064A (en) | The method and device of courses of action navigation | |
| CN110046304A (en) | A kind of user's recommended method and device | |
| WO2021238722A1 (en) | Resource pushing method and apparatus, device, and storage medium | |
| EP4181026A1 (en) | Recommendation model training method and apparatus, recommendation method and apparatus, and computer-readable medium | |
| CN111259222B (en) | Article recommendation method, system, electronic equipment and storage medium | |
| CN107463701B (en) | Method and device for pushing information stream based on artificial intelligence | |
| US20190228105A1 (en) | Dynamic website content optimization | |
| CN110807150A (en) | Information processing method and device, electronic equipment and computer readable storage medium | |
| CN112632403B (en) | Training method, recommendation method, device, equipment and medium for recommendation model | |
| Li et al. | A CTR prediction model based on user interest via attention mechanism | |
| CN112115363A (en) | Recommendation method, computing device and storage medium | |
| WO2022016556A1 (en) | Neural network distillation method and apparatus | |
| CN111767466A (en) | Recommendation information recommendation method and device based on artificial intelligence and electronic equipment | |
| CN111859149A (en) | Information recommendation method and device, electronic equipment and storage medium | |
| Wang et al. | A spatiotemporal graph neural network for session-based recommendation | |
| CN110347781A (en) | Article falls discharge method, article recommended method, device, equipment and storage medium | |
| CN114417174B (en) | Content recommendation method, device, equipment and computer storage medium | |
| CN116823410B (en) | Data processing method, object processing method, recommending method and computing device | |
| Ballandies et al. | Mobile link prediction: Automated creation and crowdsourced validation of knowledge graphs | |
| Yan et al. | Dynamic clustering based contextual combinatorial multi-armed bandit for online recommendation | |
| CN116308640A (en) | Recommendation method and related device | |
| CN112749335B (en) | Lifecycle state prediction method, lifecycle state prediction apparatus, computer device, and storage medium | |
| KR102484522B1 (en) | Method and system for providing usability test platform service based on metaverse |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181012 |
|
| RJ01 | Rejection of invention patent application after publication |