CN110442789A - Method, apparatus and electronic equipment are determined based on the association results of user behavior - Google Patents
Method, apparatus and electronic equipment are determined based on the association results of user behavior Download PDFInfo
- Publication number
- CN110442789A CN110442789A CN201910712047.3A CN201910712047A CN110442789A CN 110442789 A CN110442789 A CN 110442789A CN 201910712047 A CN201910712047 A CN 201910712047A CN 110442789 A CN110442789 A CN 110442789A
- Authority
- CN
- China
- Prior art keywords
- user
- target object
- behavior
- data
- association results
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
Abstract
The embodiment of the invention discloses a kind of association results based on user behavior to determine method, apparatus and electronic equipment, this method comprises: determining the behavior sequence of each user to sort according to time order and function;The behavior sequence of each user is divided into subsequence based on time interval;Object ID in object ID and adjacent latter item data in each subsequence in every item data is formed into object ID pair, and records each object ID to frequency of occurrence;By the object ID pair comprising target object ID, as ID pairs of target object, and using other object IDs of the target object ID centering as the affiliated partner ID of the target object ID;The association results of the target object ID are determined based on the frequency of occurrence of ID pairs of the affiliated partner ID and corresponding target object, relevance that can to the greatest extent between object of reservation, the correlation of the association results of object is set to be guaranteed, the accurate association results of each object are obtained, provide accurate foundation for object recommendation.
Description
Technical field
The present embodiments relate to data processing technique more particularly to a kind of association results determination sides based on user behavior
Method, device and electronic equipment.
Background technique
With the numerous and complicated of the fast development of development of Mobile Internet technology, the explosive growth of internet information and type,
User is caused usually to feel to feel at a loss when facing information selection.This selection diversity is not only no to generate economic benefit,
User satisfaction is reduced instead, finds that interested object is more and more difficult from bulk information.
In the related technology, the user behavior number in the past period is usually automatically analyzed and counted using recommender system
According to, and the relevance between different objects is calculated according to user behavior data, by the high object recommendation of relevance to user.But
User interest be not it is constant always, can constantly change with the recommendation of time, while user interest is to connect whithin a period of time
It passes through, when analyzing user behavior data, in the related technology due to not accounting for the continuity and gradual change of user interest
Property, cause the relevance between different objects to there is a problem of inaccuracy, so that the accurate association of each object cannot be obtained
As a result, good foundation cannot be provided for object recommendation.
Summary of the invention
The embodiment of the present invention provides a kind of association results based on user behavior and determines method, apparatus and electronic equipment, can
With the relevance between maximum object of reservation, the correlation of the association results of object is made to be guaranteed, it is each right to obtain
As accurate association results, accurate foundation is provided for object recommendation.
In a first aspect, the embodiment of the invention provides a kind of association results based on user behavior to determine method, comprising:
Determine the behavior sequence of each user to sort according to time order and function;Wherein, every in the behavior sequence of the user
Item data includes object identity ID and user behavior time of origin;
The behavior sequence of each user is divided into subsequence by the time interval based on user behavior time of origin;
Object ID in each subsequence in every item data is formed with the object ID in adjacent latter item data
Object ID pair, and each object ID is recorded to frequency of occurrence;
By the object ID pair comprising target object ID, as ID pairs of target object, and by the target object ID centering
Affiliated partner ID of other object IDs as the target object ID;
The target object ID is determined based on the frequency of occurrence of ID pairs of the affiliated partner ID and corresponding target object
Association results.
Second aspect, the association results determining device based on user behavior that the embodiment of the invention also provides a kind of, comprising:
Behavior sequence determining module, for determining the behavior sequence of each user to sort according to time order and function;Wherein, institute
Stating each data in the behavior sequence of user includes object identity ID and user behavior time of origin;
The behavior sequence of each user is divided by division module for the time interval based on user behavior time of origin
Subsequence;
Object ID to forming module, for by each subsequence in every item data object ID with it is adjacent latter
Object ID in item data forms object ID pair, and records each object ID to frequency of occurrence;
Affiliated partner ID determining module, for that will include the object ID pair of target object ID, as ID pairs of target object, and
Using other object IDs of the target object ID centering as the affiliated partner ID of the target object ID;
Association results determining module, for the occurrence out based on ID pairs of the affiliated partner ID and corresponding target object
Number determines the association results of the target object ID.
The third aspect, the embodiment of the invention also provides a kind of electronic equipment, comprising:
One or more processors;
Storage device, for storing one or more programs,
When one or more of programs are executed by one or more of processors, so that one or more of processing
Device realizes that a kind of association results based on user behavior provided in an embodiment of the present invention determine method.
Fourth aspect, the embodiment of the invention provides a kind of computer readable storage mediums, are stored thereon with computer journey
Sequence, the program realize a kind of association results determination side based on user behavior provided in an embodiment of the present invention when being executed by processor
Method.
Technical solution provided in an embodiment of the present invention, by the time interval according to user behavior time of origin by user's
Behavior sequence is divided into subsequence, and the object ID in the object ID and latter item data in item data every in subsequence is formed pair
As ID pairs;By traversing each object ID pair, the affiliated partner ID of target object ID is determined, so that it is determined that the pass of target object ID
Connection makes the relevance of association results be guaranteed as a result, the gradually changeable and continuity of user interest can be fully taken into account, and makes pair
Relevance as between is retained to the greatest extent, is obtained the accurate association results of each object, is provided for object recommendation
Accurate foundation.
Detailed description of the invention
Fig. 1 is that a kind of association results based on user behavior provided in an embodiment of the present invention determine method flow diagram;
Fig. 2 a is that a kind of association results based on user behavior provided in an embodiment of the present invention determine method flow diagram;
Fig. 2 b is that a kind of association results based on user behavior provided in an embodiment of the present invention determine method flow diagram;
Fig. 3 is a kind of association results determining device structural block diagram based on user behavior provided in an embodiment of the present invention;
Fig. 4 is a kind of electronic equipment structural schematic diagram provided in an embodiment of the present invention.
Specific embodiment
The present invention is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched
The specific embodiment stated is used only for explaining the present invention rather than limiting the invention.It also should be noted that in order to just
Only the parts related to the present invention are shown in description, attached drawing rather than entire infrastructure.
Fig. 1 is that a kind of association results based on user behavior provided in an embodiment of the present invention determine method flow diagram, described
Method can by being executed based on the association results determining device of user behavior, described device can by software and/or hardware Lai
It realizes, described device can be only fitted to the electronic equipments such as server, terminal.The method can be applied to electric business, finance, view
Frequently, in the recommender system of information, recruitment, tourism etc..
As shown in Figure 1, technical solution provided in an embodiment of the present invention includes:
S110: the behavior sequence of each user to sort according to time order and function is determined;Wherein, every in the behavior sequence of user
Item data includes object identity ID and user behavior time of origin.
In embodiments of the present invention, determine that the method for the behavior sequence of each user to sort according to time order and function can be with
It is: obtains original user behavioral data, the abnormal data in original user behavioral data is filtered, and obtain each user
Behavioral data.Wherein, original user behavioral data can be original user and click behavioral data, can be user and clicks quotient
The behavioral data of the generations such as product, webpage, video, information, object can be commodity, article, webpage, video, information etc..Original use
Family behavioral data may include the primitive behavior data of multiple users.Abnormal data can be User ID as sky, and object ID is sky,
Or user behavior time of origin is the data of non-time-stamped.
Wherein, after the behavioral data for obtaining each user, by object ID popular in the behavioral data of each user or
The corresponding data item of person's unexpected winner object ID is filtered, by the every item data obtained after filtering according to user behavior time of origin
Sequencing is ranked up, and obtains the behavior sequence of the user to sort according to time order and function.As a result, by by unexpected winner object ID or
The corresponding data item of person's hot topic object ID is filtered, and unexpected winner object and popular object can be removed from association results, is made whole
Body association results resist dry property to greatly enhance.
In the other embodiments of the embodiment of the present invention, it is also possible to that mode is implemented as follows: is obtaining each user's
After behavioral data, directly the behavioral data of user can be ranked up according to the sequencing of user behavior time of origin,
Obtain the behavior sequence of the user to sort according to chronological order.
S120: the behavior sequence of each user is divided into subsequence by the time interval based on user behavior time of origin.
In embodiments of the present invention, for the behavior sequence of each user, when the behavior sequence to user is divided into sub- sequence
When column, in the same subsequence, the interval of user's time of origin in adjacent two item data is less than setting time interval, not
In same subsequence, the interval of user's time of origin between two adjacent subsequences is greater than setting time interval.Lead to as a result,
It crosses the time interval based on user behavior time of origin and the behavior sequence of each user is divided into subsequence, in different subsequences
The time interval of user's time of origin is greater than setting time interval, can consider user interest gradually from time interval angle
Denaturation makes the relevance of association results be guaranteed to obtain the association results for being suitble to user interest.
S130: the object ID in item data every in each subsequence is formed with the object ID in adjacent latter item data
Object ID pair, and each object ID is recorded to frequency of occurrence.
In embodiments of the present invention, each data includes object ID and user behavior time of origin in subsequence.By sub- sequence
The object ID in object ID and adjacent latter item data in column in every item data forms object ID pair, and records each object
ID is to frequency of occurrence.For example, if the subsequence that the behavior sequence of user is divided into be [[Aid, T1], [Bid, T2],
[Cid, T3]], then the object ID that the subsequence is formed can recorde each to may is that [Aid, Bid] and [Bid, Cid]
The frequency of occurrence of object ID pair records each object ID to realize as soon as accumulated counts are primary when object ID is to appearance time
To the purpose of frequency of occurrence.By the way that the object ID in the object ID and latter item data in item data every in subsequence is formed pair
As ID pairs, it may be considered that the relevance between adjacent object, it is contemplated that the continuity of user interest makes the relevance between object
Maximum program is obtained to retain.
S140:,, will be described by target object ID as ID couples of target object by the object ID pair comprising target object ID
Affiliated partner ID of other object IDs of target object ID centering as the target object ID.
It in embodiments of the present invention, can be by the object ID pair comprising target object ID, as ID pairs of target object.It is optional
, it can will include target object ID, and target object ID is the object ID pair of setting sequence, as ID pairs of target object.
It can be appreciated that will include target object ID, and the serial number of target object ID is the object ID pair for setting serial number, as
ID pairs of target object.For example, [Aid, Bid], [Aid, Cid] and [Aid, Did] three object ID centerings, the object ID for including
It is Aid, then sequence number of the Aid in three object IDs pair is 1, then can be by three object IDs to as including Aid
ID pairs of target object.
In embodiments of the present invention, the object ID by each target object ID centering in addition to target object ID, as mesh
Mark the affiliated partner ID of object ID.Summarize object of all target object ID centerings in addition to target object ID, obtains target pair
As all affiliated partner ID of ID.For example, if [Aid, Bid], [Aid, Cid] and [Aid, Did] is as comprising Aid
ID pairs of target object, then Bid, Cid and Did, are the affiliated partner ID of Aid.
S150: the target pair is determined based on the frequency of occurrence of ID pairs of the affiliated partner ID and corresponding target object
As the association results of ID.
In embodiments of the present invention, the affiliated partner ID based on target object ID and mesh corresponding with affiliated partner ID
The frequency of occurrence for marking object ID pair, determines the association results of target object ID.Wherein target pair corresponding with affiliated partner ID
As ID is to referring to: ID pairs of target object comprising target object ID and affiliated partner ID.
It, can be by affiliated partner ID according to ID pairs of corresponding target object in an embodiment of the embodiment of the present invention
Frequency of occurrence be ranked up from big to small, by frequency of occurrence be less than setting frequency threshold value target object ID to corresponding association
Object ID is filtered;Take top n affiliated partner ID as the association results of target object ID;Or take top n affiliated partner
Association results of the corresponding target object ID of ID and top n affiliated partner ID to the number of appearance as target object ID.
It, optionally, can be by affiliated partner ID according to corresponding target pair in an embodiment of the embodiment of the present invention
As ID pairs of frequency of occurrence is ranked up from big to small, using the affiliated partner ID after sequence as the association knot of target object ID
Fruit.Or affiliated partner ID can also be ranked up from big to small according to the frequency of occurrence of corresponding ID pairs of target object, it will go out
The target object ID that occurrence number is less than setting frequency threshold value is filtered corresponding affiliated partner ID, the pass that will be obtained after filtering
Join association results of the object ID as target object ID.
It should be noted that method provided in an embodiment of the present invention does not depend on each attribute and text word of object itself
Section is suitable for various recommendation scenes, can obtain wider application.
In the related technology, the user behavior number in the past period is usually automatically analyzed and counted using recommender system
According to, and the relevance between different objects is calculated according to user behavior data, by the high object recommendation of relevance to user.But
User interest be not it is constant always, can constantly change with the recommendation of time, while user interest is to connect whithin a period of time
It passes through, when analyzing user behavior data, in the related technology due to not accounting for the continuity and gradual change of user interest
Property, cause the relevance between different objects to there is a problem of inaccuracy, so that the accurate association of each object cannot be obtained
As a result, good foundation cannot be provided for object recommendation.Method provided in an embodiment of the present invention, by being occurred according to user behavior
The behavior sequence of user is divided into subsequence by the time interval of time, can be fully taken into account the gradually changeable of user interest, be made
The relevance of association results is guaranteed;By by item data every in subsequence object ID and latter item data in object
ID forms object ID pair;By traversing each object ID pair, the affiliated partner ID of target object ID is determined, so that it is determined that target pair
As the association results of ID, the continuity of user interest can be fully considered, obtain the relevance between object to the greatest extent
Retain, obtains the accurate association results of each object, provide accurate foundation for object recommendation.
In the related technology, with the fast development of development of Mobile Internet technology, the explosive growth of internet information and type
Numerous and complicated, cause user usually face information selection when feel to feel at a loss.This selection diversity does not produce not only
Raw economic benefit, reduces user satisfaction instead.Meanwhile there is long-tail phenomenon in the various objects on internet, refer to big portion again
Object is divided to belong to chance of the unexpected winner without displaying.In recent years, recommender system can be used to solve the problems, such as long-tail, go out not in face of layer
Poor information and service, personalized recommendation emerges rapidly, and yields unusually brilliant results, in finance, electric business, video, information, live streaming, trick
The every field such as engage, travel can see the presence of recommender system.
In the related technology, for user in recommender system interactive process, positive feedback data are relatively fewer.User is constantly clear
It lookes at many objects, the operation such as has click, thumb up, collect, sharing, but many times positive feedback of the user for recommendation results
" thumbing up " or scoring behavior be it is seldom, the yet very small of behavior " is not liked " for negative-feedback, this is just to some supervision
The application of learning model brings inconvenience, and method provided in an embodiment of the present invention is obtained using analyzing user behavior data
It, can be to avoid the defect of supervised learning model to the association results of each object.
In the related technology, using content-based recommendation method, the accuracy of article self attributes is relied on, such as classification, mark
The text informations such as label, but there is also many problems in accuracy during actual implementation, in addition also have in headline very much
Useless text, the correlation for eventually leading to recommendation results is inadequate, affects final recommendation experience.It is provided in an embodiment of the present invention
Attribute and the text field of the method independent of object itself, can also accurate association results, with for user recommend provide
Foundation.
In the related technology, the method for object-based collaborative filtering, the behavior sequence interior for a period of time dependent on user, into
And calculate the correlation between different objects.But it is multifactor to be susceptible to user behavior sequence length, time interval size etc.
Influence, cause object dependencies are very high not excavate, the embodiment of the present invention by according to time interval by user
Behavior sequence be divided into subsequence, the gradually changeable of user interest can be fully taken into account, so that the correlation of association results obtains
It is sufficiently examined to guarantee by the way that the object ID in the object ID and latter item data in item data every in subsequence is formed ID pairs
Consider the continuity of user interest, it is contemplated that the association between adjacent object, so that the relevance between object obtains maximum journey
Sequence retains, and by each object pair of traversal, determines the affiliated partner ID of target object ID, thus the association results of target object,
To guarantee the high association results of available correlation.
Technical solution provided in an embodiment of the present invention, by the time interval according to user behavior time of origin by user's
Behavior sequence is divided into subsequence, and the object ID in the object ID and latter item data in item data every in subsequence is formed pair
As ID pairs;By traversing each object ID pair, the affiliated partner ID of target object ID is determined, so that it is determined that the pass of target object ID
Connection makes the relevance of association results be guaranteed as a result, the gradually changeable and continuity of user interest can be fully taken into account, and makes pair
Relevance as between is retained to the greatest extent, is obtained the accurate association results of each object, is provided for object recommendation
Accurate foundation.
Fig. 2 a is that a kind of association results based on user behavior provided in an embodiment of the present invention determine method flow diagram, this reality
Applying the scheme in example can be combined with the optinal plan of said one or multiple embodiments, in embodiments of the present invention,
Optionally, the behavior sequence of each user is divided into subsequence by the time interval based on user behavior time of origin, packet
It includes:
In the behavior sequence of each user, if being used in user behavior time of origin and the next item up data in current item data
The time interval of family behavior time of origin is more than setting time interval, current item data is put into new subsequence;Otherwise,
Current item data is put into subsequence identical with the next item up data.
Optionally, the behavior sequence for each user that the determination is sorted according to time order and function, comprising:
The behavior sequence of each user is generated based on user behavior data;
The triggering times of each object ID in counting user behavioral data;
In the behavior sequence of each user, if the triggering times of object ID are greater than the first setting number or less than second
Number is set, the corresponding data item of the object ID is filtered;
The every item data obtained after filtering is ranked up according to the sequencing of user behavior time of origin, obtain according to
The behavior sequence of the user of time order and function sequence.
As shown in Figure 2 a, technical solution provided in an embodiment of the present invention includes:
S210: original user behavioral data is obtained, and abnormal data is filtered, obtains the behavioral data of user.
In embodiments of the present invention, wherein original user behavioral data can be original user and click behavioral data, can be with
It is the behavioral data that user clicks the generations such as commodity, webpage, video, information.Original user behavioral data may include multiple use
The primitive behavior data at family.Abnormal data can be User ID as sky, and object ID is that empty or user behavior time of origin is non-
The data of timestamp.Abnormal data is filtered to obtain the behavioral data of user.Every behavioral data include user identifier ID,
Object ID and user behavior time of origin.
S220: the behavior sequence of each user is generated based on user behavior data.
In embodiments of the present invention, traverse user behavioral data generates the behavior sequence of each user, wherein Mei Geyong
Each data in the behavior sequence at family include object ID and user behavior time of origin.
S230: the triggering times of each object ID in counting user behavioral data.
In embodiments of the present invention, the triggering times of each object ID can be the number of clicks of each object ID.
S240: in the behavior sequence of each user, if the triggering times of object ID are greater than the first setting number or small
Number is set in second, the corresponding data item of the object ID is filtered.
In embodiments of the present invention, the first setting number is greater than the second setting number.First setting number and the second setting
Number, which can according to need, to be configured.In the behavior sequence of each user, by triggering times be greater than first setting number or
Person is filtered less than the corresponding data item of object ID of the second setting number, can will be popular in the behavioral data of each user
Object ID or the corresponding data item of unexpected winner object ID are filtered, and association results removal unexpected winner object and hot topic can be made right
As greatly enhance association results resists dry property.
S250: the every item data obtained after filtering is ranked up according to the sequencing of user behavior time of origin, is obtained
To the behavior sequence of the user to sort according to time order and function.
S260: in the behavior sequence of each user, if user behavior time of origin and the next item up number in current item data
It is more than setting time interval according to the time interval of middle user behavior time of origin, current item data is put into new subsequence
In;Otherwise, current item data is put into subsequence identical with the next item up data.
In embodiments of the present invention, judge user behavior time of origin and user's row in the next item up data in current item data
It whether is more than setting time interval for the interval of time of origin, if so, current item data is put into new subsequence, if
It is no, current item data is put into subsequence identical with the next item up data, it is ensured that in same subsequence, adjacent two
The interval of user's time of origin in item data is less than setting time interval, in different subsequences, two adjacent sub- sequences
The interval of user's time of origin between column is greater than setting time interval, and user interest can be considered from time interval angle
Gradually changeable, thus obtain be suitble to user interest association results, so that the relevance of association results is guaranteed.
S270: by the object ID in the object ID and adjacent latter item data in each subsequence in every item data
Object ID pair is formed, and records each object ID to frequency of occurrence.
S280: by the object ID pair comprising target object ID, as ID couples of target object, and by ID couples of the target object
In affiliated partner ID of other object IDs as the target object ID.
S290: the target pair is determined based on the frequency of occurrence of ID pairs of the affiliated partner ID and corresponding target object
As the association results of ID.
It is optional described based on the affiliated partner ID and corresponding in an embodiment of the embodiment of the present invention
The frequency of occurrence of ID pairs of the target object determines the association results of the target object ID, comprising:
Affiliated partner ID is ranked up from big to small according to ID pairs of corresponding target object of frequency of occurrence;By occurrence out
The target object ID that number is less than setting frequency threshold value is filtered corresponding affiliated partner ID;Top n affiliated partner ID is taken to make
For the association results of the target object ID;Or take top n affiliated partner ID and the top n affiliated partner ID corresponding
Association results of the target object ID to the number of appearance as target object ID.For example, [Aid, Bid], [Aid, Cid] and
[Aid, Did] is ID pairs of Aid corresponding target object respectively, and the number occurred respectively is 100,10,1.Bid, Cid and Did points
Be not the affiliated partner ID of Aid, by the affiliated partner ID of Aid according to corresponding target object pair frequency of occurrence from big to small into
Row sequence is: Bid, Cid and Did.If setting frequency threshold value is 5 times, Did is filtered, Bid and Cid are obtained, can incited somebody to action
Association results of the Bid and Cid as Aid, or can also association results by [Bid, 100] and [Cid, 10] as Aid, or
N can also be taken 1 by person, then using Bid as the association results of Aid.
On the basis of the above embodiments, technical solution provided in an embodiment of the present invention can also include: when user triggers
When the target object, the association results of target object are determined based on the association results of the target object ID, and by the mesh
The association results of mark object recommend user.As a result, by determining target object ID's based on providing method of the embodiment of the present invention
The association results of target object are recommended user by association results, can recommend possible interested object for user, and raising pushes away
Recommend the clicking rate of object.
Specific method provided in an embodiment of the present invention can also refer to method shown in Fig. 2 b, and specific steps may include:
Step 1: the original user behavioral data behavior_ in nearest a period of time BEHAVIOUR_DAYS is obtained
Data, first progress data cleansing and screening, user id be sky, article id is that sky, behavior time of origin non-time-stamped etc. are different
Regular data filters out, while only retaining click behavior, the data cleaning_behavior_data after obtaining final process.Often
Behavioral data mainly includes three fields: userid, itemid, action_time, respectively indicates user id, article id, point
Hit behavior time of origin.
Step 2: the traversal every data line of cleaning_behavior_data, generates the row of each user userid
For sequence list behavior_seq, the format of each item_value is [itemid, action_time], entire list
Format is [item_value_1, item_value_2 ..., item_value_n], while counting the click frequency of each itemid
Secondary data freq_count.
Step 3: traversing the behavior_seq of each user userid, judge each item_value's
The freq_count of itemid if it is greater than threshold value MAX_FREQ_THRESHOLD or is less than MIN_FREQ_THRESHOLD just
It filters out, then sorts from front to back according to action_time, guarantee the itemid first clicked sequence in front, reaction is
One user userid constantly clicks the order of different itemid over time, obtains orderly behavior sequence
Order_behavior_seq, format are [[itemid_1, action_time_1], [itemid_2, action_time_
2]…,[itemid_n,action_time_n]]。
Step 4: each item_value of traversal order_behavior_seq, if with next item_
The difference of the action_time of value is no more than specified threshold TIME_DIFF_THRESHOLD and is just put into the same subsequence
Otherwise sub_order_behavior_seq is just put into new subsequence, the item_value in each subsequence is still
According to action_time ascending order.After handling in this way, the order_behavior_seq of a user has just split different
Subsequence sub_order_behavior_seq, the interval action_time of the adjacent item_value of the same subsequence is not
More than TIME_DIFF_THRESHOLD, the interval action_time of any two item_value of different subsequences is super
Cross TIME_DIFF_THRESHOLD.
Step 5: each item_value and the latter of traversal subsequence sub_order_behavior_seq
Item_value forms item pairs, and accumulated counts are primary.
Step 6: repeating the above steps two-five, different item couples of all co-occurrence number co_freq are finally obtained, are denoted as
Item_co_seq, each single item are denoted as co_value, and format is [itemid_i, itemid_j, co_freq_ij], item_co_
The format of seq is [co_value_1, co_value_2 ..., co_value_n].
Step 7: each single item of traversal item_co_seq, obtains all association results lists of each itemid_i
Co_result_list, format be [[itemid_1, co_freq_1], [itemid_2, co_freq_2] ..., [itemid_n,
Co_freq_n]], then according to co_freq, descending is arranged from big to small, filters out the FREQ_ that co_freq is less than specified threshold
Then THRESHOLD takes final association results of the TopN as current itemid_i, and arranges according to correlation degree descending.
Fig. 3 is a kind of association results determining device structural block diagram based on user behavior provided in an embodiment of the present invention, such as
Shown in Fig. 3, device provided in an embodiment of the present invention includes: behavior sequence determining module 310, division module 320, object ID to shape
At module 330, affiliated partner ID determining module 340 and association results determining module 350.
Behavior sequence determining module 310, for determining the behavior sequence of each user to sort according to time order and function;Its
In, each data includes object identity ID and user behavior time of origin in the behavior sequence of the user;
Division module 320 draws the behavior sequence of each user for the time interval based on user behavior time of origin
It is divided into subsequence;
Object ID is to forming module 330, for by object ID in each subsequence in every item data and adjacent
Object ID in latter item data forms object ID pair, and records each object ID to frequency of occurrence;
Affiliated partner ID determining module 340, for that will include the object ID pair of target object ID, as target object ID
It is right, and using other object IDs of the target object ID centering as the affiliated partner ID of the target object ID;
Association results determining module 350, for based on the affiliated partner ID and corresponding target object ID to going out
Occurrence number determines the association results of the target object ID.
Optionally, division module 320, in the behavior sequence of each user, if user behavior in current item data
The time interval of user behavior time of origin is more than setting time interval in time of origin and the next item up data, by current item data
It is put into new subsequence;Otherwise, current item data is put into subsequence identical with the next item up data.
Optionally, behavior sequence determining module 310, is used for:
The behavior sequence of each user is generated based on user behavior data;
The triggering times of each object ID in counting user behavioral data;
In the behavior sequence of each user, if the triggering times of object ID are greater than the first setting number or less than second
Number is set, the corresponding data item of the object ID is filtered;
The every item data obtained after filtering is ranked up according to the sequencing of user behavior time of origin, obtain according to
The behavior sequence of the user of time order and function sequence.
Optionally, described device behavioral data obtains module, for obtaining original user behavioral data, and by abnormal data
It is filtered, obtains the behavioral data of user;
Wherein, every behavioral data includes user identifier ID, object ID and user behavior time of origin.
Optionally, described device further includes recommending module, for being based on the mesh when user triggers the target object
The association results of mark object ID determine the association results of target object, and the association results of the target object push away user
It recommends.
Optionally, association results determining module 350, is used for:
Affiliated partner ID is ranked up from big to small according to ID pairs of corresponding target object of frequency of occurrence;
The target object ID that frequency of occurrence is less than setting frequency threshold value is filtered corresponding affiliated partner ID;
Take top n affiliated partner ID as the association results of the target object ID;Or take top n affiliated partner ID with
And association results of the corresponding target object ID of the top n affiliated partner ID to the number of appearance as target object ID.
Optionally, described by the object ID pair comprising target object ID, as ID pairs of target object, comprising:
Will include target object ID, and the target object ID be ordered as setting sequence object ID pair, as target
Object ID pair.
Method provided by any embodiment of the invention can be performed in above-mentioned apparatus, has the corresponding functional module of execution method
And beneficial effect.
Fig. 4 is a kind of electronic equipment structural schematic diagram provided in an embodiment of the present invention, as shown in figure 4, the equipment includes:
One or more processors 410, in Fig. 4 by taking a processor 410 as an example;
Memory 420;
The equipment can also include: input unit 430 and output device 440.
Processor 410, memory 420, input unit 430 and output device 440 in the equipment can pass through bus
Or other modes connect, in Fig. 4 for being connected by bus.
Memory 420 be used as a kind of non-transient computer readable storage medium, can be used for storing software program, computer can
Program and module are executed, as one of embodiment of the present invention determines the corresponding journey of method based on the association results of user behavior
Sequence instruction/module is (for example, attached behavior sequence determining module 310 shown in Fig. 3, division module 320, object ID are to formation module
330, affiliated partner ID determining module 340 and association results determining module 350).Processor 410 is stored in memory by operation
Software program, instruction and module in 420, thereby executing the various function application and data processing of computer equipment, i.e.,
Realize that a kind of association results based on user behavior of above method embodiment determine method, it may be assumed that
Determine the behavior sequence of each user to sort according to time order and function;Wherein, every in the behavior sequence of the user
Item data includes object identity ID and user behavior time of origin;
The behavior sequence of each user is divided into subsequence by the time interval based on user behavior time of origin;
Object ID in each subsequence in every item data is formed with the object ID in adjacent latter item data
Object ID pair, and each object ID is recorded to frequency of occurrence;
By the object ID pair comprising target object ID, as ID pairs of target object, and by the target object ID centering
Affiliated partner ID of other object IDs as the target object ID;
The target object ID is determined based on the frequency of occurrence of ID pairs of the affiliated partner ID and corresponding target object
Association results.
Memory 420 may include storing program area and storage data area, wherein storing program area can store operation system
Application program required for system, at least one function;Storage data area can be stored to be created according to using for computer equipment
Data etc..In addition, memory 420 may include high-speed random access memory, it can also include non-transitory memory, such as
At least one disk memory, flush memory device or other non-transitory solid-state memories.In some embodiments, it stores
Optional device 420 includes the memory remotely located relative to processor 410, these remote memories can be by being connected to the network extremely
Terminal device.The example of above-mentioned network includes but is not limited to internet, intranet, local area network, mobile radio communication and its group
It closes.
Input unit 430 can be used for receiving the number or character information of input, and generate the user with computer equipment
Setting and the related key signals input of function control.Output device 440 may include output interface etc..
The embodiment of the invention provides a kind of computer readable storage mediums, are stored thereon with computer program, the program
Realize that a kind of association results based on user behavior such as provided in an embodiment of the present invention determine method when being executed by processor:
Determine the behavior sequence of each user to sort according to time order and function;Wherein, every in the behavior sequence of the user
Item data includes object identity ID and user behavior time of origin;
The behavior sequence of each user is divided into subsequence by the time interval based on user behavior time of origin;
Object ID in each subsequence in every item data is formed with the object ID in adjacent latter item data
Object ID pair, and each object ID is recorded to frequency of occurrence;
By the object ID pair comprising target object ID, as ID pairs of target object, and by the target object ID centering
Affiliated partner ID of other object IDs as the target object ID;
The target object ID is determined based on the frequency of occurrence of ID pairs of the affiliated partner ID and corresponding target object
Association results.
It can be using any combination of one or more computer-readable media.Computer-readable medium can be calculating
Machine readable signal medium or computer readable storage medium.Computer readable storage medium for example can be --- but it is unlimited
In system, device or the device of --- electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor, or any above combination.It calculates
The more specific example (non exhaustive list) of machine readable storage medium storing program for executing includes: electrical connection with one or more conducting wires, just
Taking formula computer disk, hard disk, random access memory (RAM), read-only memory (ROM), erasable type may be programmed read-only storage
Device (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), light storage device, magnetic memory device,
Or above-mentioned any appropriate combination.In this document, computer readable storage medium can be it is any include or storage journey
The tangible medium of sequence, the program can be commanded execution system, device or device use or in connection.
Computer-readable signal media may include in a base band or as carrier wave a part propagate data-signal,
Wherein carry computer-readable program code.The data-signal of this propagation can take various forms, including --- but
It is not limited to --- electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be
Any computer-readable medium other than computer readable storage medium, which can send, propagate or
Transmission is for by the use of instruction execution system, device or device or program in connection.
The program code for including on computer-readable medium can transmit with any suitable medium, including --- but it is unlimited
In --- wireless, electric wire, optical cable, RF etc. or above-mentioned any appropriate combination.
The computer for executing operation of the present invention can be write with one or more programming languages or combinations thereof
Program code, described program design language include object oriented program language-such as Java, Smalltalk, C++,
Further include conventional procedural programming language-such as " C " language or similar programming language.Program code can be with
It fully executes, partly execute on the user computer on the user computer, being executed as an independent software package, portion
Divide and partially executes or executed on a remote computer or server completely on the remote computer on the user computer.In
Be related in the situation of remote computer, remote computer can pass through the network of any kind --- including local area network (LAN) or
Wide area network (WAN)-be connected to subscriber computer, or, it may be connected to outer computer (such as mentioned using Internet service
It is connected for quotient by internet).
Note that the above is only a better embodiment of the present invention and the applied technical principle.It will be appreciated by those skilled in the art that
The invention is not limited to the specific embodiments described herein, be able to carry out for a person skilled in the art it is various it is apparent variation,
It readjusts and substitutes without departing from protection scope of the present invention.Therefore, although being carried out by above embodiments to the present invention
It is described in further detail, but the present invention is not limited to the above embodiments only, without departing from the inventive concept, also
It may include more other equivalent embodiments, and the scope of the invention is determined by the scope of the appended claims.
Claims (10)
1. a kind of association results based on user behavior determine method characterized by comprising
Determine the behavior sequence of each user to sort according to time order and function;Wherein, every item number in the behavior sequence of the user
According to including object identity ID and user behavior time of origin;
The behavior sequence of each user is divided into subsequence by the time interval based on user behavior time of origin;
Object ID in object ID and adjacent latter item data in each subsequence in every item data is formed into object
ID pairs, and each object ID is recorded to frequency of occurrence;
By the object ID pair comprising target object ID, as ID pairs of target object, and by the target object ID centering other
Affiliated partner ID of the object ID as the target object ID;
The pass of the target object ID is determined based on the frequency of occurrence of ID pairs of the affiliated partner ID and corresponding target object
It is coupled fruit.
2. the method according to claim 1, wherein the time interval based on user behavior time of origin will
The behavior sequence of each user is divided into subsequence, comprising:
In the behavior sequence of each user, if user behavior time of origin and user's row in the next item up data in current item data
It is more than setting time interval for the time interval of time of origin, current item data is put into new subsequence;Otherwise, will work as
Preceding item data is put into subsequence identical with the next item up data.
3. the method according to claim 1, wherein
The behavior sequence for each user that the determination is sorted according to time order and function, comprising:
The behavior sequence of each user is generated based on user behavior data;
The triggering times of each object ID in counting user behavioral data;
In the behavior sequence of each user, if the triggering times of object ID are greater than the first setting number or less than the second settings
The corresponding data item of the object ID is filtered by number;
The every item data obtained after filtering is ranked up according to the sequencing of user behavior time of origin, is obtained according to the time
The behavior sequence of the user successively to sort.
4. according to the method described in claim 3, it is characterized by further comprising:
Original user behavioral data is obtained, and abnormal data is filtered, obtains the behavioral data of user;
Wherein, every behavioral data includes user identifier ID, object ID and user behavior time of origin.
5. the method according to claim 1, wherein further include:
When user triggers the target object, the association knot of target object is determined based on the association results of the target object ID
Fruit, and the association results of the target object are recommended into the user.
6. the method according to claim 1, wherein described based on the affiliated partner ID and corresponding described
ID pairs of target object of frequency of occurrence determines the association results of the target object ID, comprising:
Affiliated partner ID is ranked up from big to small according to ID pairs of corresponding target object of frequency of occurrence;
The target object ID that frequency of occurrence is less than setting frequency threshold value is filtered corresponding affiliated partner ID;
Take top n affiliated partner ID as the association results of the target object ID;Or take top n affiliated partner ID and institute
State association results of the corresponding target object ID of top n affiliated partner ID to the number of appearance as target object ID.
7. the method according to claim 1, wherein described by the object ID pair comprising target object ID, as
ID pairs of target object, comprising:
Will include target object ID, and the target object ID be ordered as setting sequence object ID pair, as target object
ID pairs.
8. a kind of association results determining device based on user behavior characterized by comprising
Behavior sequence determining module, for determining the behavior sequence of each user to sort according to time order and function;Wherein, the use
Each data includes object identity ID and user behavior time of origin in the behavior sequence at family;
The behavior sequence of each user is divided into sub- sequence for the time interval based on user behavior time of origin by division module
Column;
Object ID is to forming module, for by the object ID and adjacent latter item number in each subsequence in every item data
Object ID in forms object ID pair, and records each object ID to frequency of occurrence;
Affiliated partner ID determining module, for that will include the object ID pair of target object ID, as ID pairs of target object, and by institute
State affiliated partner ID of other object IDs of target object ID centering as the target object ID;
Association results determining module, it is true for the frequency of occurrence based on ID pairs of the affiliated partner ID and corresponding target object
The association results of the fixed target object ID.
9. a kind of electronic equipment characterized by comprising
One or more processors;
Storage device, for storing one or more programs,
When one or more of programs are executed by one or more of processors, so that one or more of processors are real
A kind of association results based on user behavior now as claimed in claim 1 determine method.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is by processor
Realize that a kind of association results based on user behavior as claimed in claim 1 determine method when execution.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910712047.3A CN110442789B (en) | 2019-08-02 | 2019-08-02 | User behavior based association result determination method and device and electronic equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910712047.3A CN110442789B (en) | 2019-08-02 | 2019-08-02 | User behavior based association result determination method and device and electronic equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110442789A true CN110442789A (en) | 2019-11-12 |
CN110442789B CN110442789B (en) | 2022-12-09 |
Family
ID=68433036
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910712047.3A Active CN110442789B (en) | 2019-08-02 | 2019-08-02 | User behavior based association result determination method and device and electronic equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110442789B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111475720A (en) * | 2020-03-31 | 2020-07-31 | 北京三快在线科技有限公司 | Recommendation method, recommendation device, server and storage medium |
CN113763110A (en) * | 2021-02-08 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Article recommendation method and device |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20160196305A1 (en) * | 2015-01-04 | 2016-07-07 | Emc Corporation | Analyzing user behaviors |
CN107330050A (en) * | 2017-06-28 | 2017-11-07 | 达而观信息科技(上海)有限公司 | A kind of personalized recommendation method and system |
US20180307544A1 (en) * | 2017-04-21 | 2018-10-25 | International Business Machines Corporation | Event sequence management |
CN109711948A (en) * | 2019-01-09 | 2019-05-03 | 广州视源电子科技股份有限公司 | Item recommendation method, device, equipment and storage medium |
-
2019
- 2019-08-02 CN CN201910712047.3A patent/CN110442789B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20160196305A1 (en) * | 2015-01-04 | 2016-07-07 | Emc Corporation | Analyzing user behaviors |
US20180307544A1 (en) * | 2017-04-21 | 2018-10-25 | International Business Machines Corporation | Event sequence management |
CN107330050A (en) * | 2017-06-28 | 2017-11-07 | 达而观信息科技(上海)有限公司 | A kind of personalized recommendation method and system |
CN109711948A (en) * | 2019-01-09 | 2019-05-03 | 广州视源电子科技股份有限公司 | Item recommendation method, device, equipment and storage medium |
Non-Patent Citations (1)
Title |
---|
黎丹雨等: "一种多层多维的关联规则挖掘算法在推荐系统中的应用", 《计算机与现代化》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111475720A (en) * | 2020-03-31 | 2020-07-31 | 北京三快在线科技有限公司 | Recommendation method, recommendation device, server and storage medium |
CN113763110A (en) * | 2021-02-08 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Article recommendation method and device |
Also Published As
Publication number | Publication date |
---|---|
CN110442789B (en) | 2022-12-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107888950B (en) | A kind of method and system for recommending video | |
CN110457581A (en) | A kind of information recommended method, device, electronic equipment and storage medium | |
US9972055B2 (en) | Fact checking method and system utilizing social networking information | |
CN109511015B (en) | Multimedia resource recommendation method, device, storage medium and equipment | |
CN112364202B (en) | Video recommendation method and device and electronic equipment | |
KR20190011829A (en) | Estimating and displaying social interest in time-based media | |
CN108540826A (en) | Barrage method for pushing, device, electronic equipment and storage medium | |
CN111680254B (en) | Content recommendation method and device | |
CN108959319B (en) | Information pushing method and device | |
CN108519991A (en) | A kind of method and apparatus of main broadcaster's account recommendation | |
CN109408639A (en) | A kind of barrage classification method, device, equipment and storage medium | |
CN104782138A (en) | Identifying a thumbnail image to represent a video | |
CN110941738B (en) | Recommendation method and device, electronic equipment and computer-readable storage medium | |
CN112765400B (en) | Weight updating method, content recommending method, device and equipment for interest labels | |
CN109743589B (en) | Article generation method and device | |
CN109275047A (en) | Video information processing method and device, electronic equipment, storage medium | |
CN111091006A (en) | Entity intention system establishing method, device, equipment and medium | |
CN111062527A (en) | Video collection flow prediction method and device | |
CN113779381A (en) | Resource recommendation method and device, electronic equipment and storage medium | |
CN110442789A (en) | Method, apparatus and electronic equipment are determined based on the association results of user behavior | |
CN109245989A (en) | A kind of processing method, device and computer readable storage medium shared based on information | |
CN112287165A (en) | File processing method and device | |
US20090113466A1 (en) | System, Method and Computer Program Product for Evaluating Media Streams | |
CN108319628A (en) | A kind of user interest determines method and device | |
CN110569447B (en) | Network resource recommendation method and device and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |