CN109710876A - A kind of information recommendation method and device, computer readable storage medium - Google Patents
A kind of information recommendation method and device, computer readable storage medium Download PDFInfo
- Publication number
- CN109710876A CN109710876A CN201811605413.7A CN201811605413A CN109710876A CN 109710876 A CN109710876 A CN 109710876A CN 201811605413 A CN201811605413 A CN 201811605413A CN 109710876 A CN109710876 A CN 109710876A
- Authority
- CN
- China
- Prior art keywords
- access
- access side
- accessed
- accessed object
- retention amount
- 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
Landscapes
- Information Transfer Between Computers (AREA)
Abstract
The embodiment of the invention discloses a kind of information recommendation method and device, computer readable storage medium, the information recommendation method includes: acquisition real-time streaming data;The real-time streaming data includes: mapping relations of the mark of each accessed object with each access side for the login moment of each accessed object;According to the mapping relations, determine each access side for the access retention amount of each accessed object;According to the access retention amount, each access side is clustered, cluster result is obtained;According to the cluster result, recommend accessed object corresponding with cluster centre to each access side.
Description
Technical field
The present invention relates to big data analysis and calculating field more particularly to a kind of information recommendation methods and device, computer
Readable storage medium storing program for executing.
Background technique
The booming of internet brings the outburst of massive information, targeted in massive information in order to improve user
Ground obtains the efficiency of information needed, can obtain user demand by retaining analysis user interest, and then targetedly recommend
The interested information of user.
The retention analysis calculation method of mainstream is acquired to data at present, is uploaded to the distributed post of high-throughput
Message system is subscribed to, for handling the everything flow data in website, such as web page browsing, search or other access objects
Data are accessed, retention analysis uses Hive scripting language, according to user's active identification, user's concern relation or analysis user
Feature obtains recommendation information.
However, it is existing retain analysis calculation method be a kind of T+1 mode off-line calculation model, to offline data into
Row retains analysis, there is a problem of that accuracy is low, and existing retentions analysis is based only on retention and analyzes result and directly acquires and push away
Information is recommended, there are problems that information recommendation lacks flexibility.
Summary of the invention
In order to solve the above technical problems, an embodiment of the present invention is intended to provide a kind of information recommendation method and devices, computer
Readable storage medium storing program for executing can targetedly be recommended based on access object, improve the flexibility of information recommendation.
The technical scheme of the present invention is realized as follows:
In a first aspect, the embodiment of the present invention provides a kind of information recommendation method, which comprises
Obtain real-time streaming data;The real-time streaming data include: each accessed object mark and each access side for institute
State the mapping relations at the login moment of each accessed object;
According to the mapping relations, determine each access side for the access retention amount of each accessed object;
According to the access retention amount, each access side is clustered, cluster result is obtained;
According to the cluster result, recommend accessed object corresponding with cluster centre to each access side.
Second aspect, the embodiment of the present invention provide a kind of information recommending apparatus, and described device includes:
Acquiring unit, for obtaining real-time streaming data;The real-time streaming data includes: the mark of each accessed object and each
Mapping relations of the access side for the login moment of each accessed object;
Determination unit, for determining each access side for each accessed object according to the mapping relations
Access retention amount;
Cluster cell, for the access retention amount according to each access side for each accessed object, to described
Each access side clusters;
Recommendation unit, for recommending to each access side corresponding with cluster centre accessed according to cluster result
Object.
The third aspect, the embodiment of the present invention provide a kind of information recommending apparatus, wherein the information recommending apparatus at least wraps
Include processor, the memory for being stored with the processor-executable instruction, communication interface, and for connect the processor,
The bus of the memory and the communication interface, when executed, the processor realize above-mentioned reality when executing
The information recommendation method of example offer is provided.
Fourth aspect, the embodiment of the present invention provide a kind of computer readable storage medium, store computer program thereon, institute
It states and realizes information recommendation method provided by the above embodiment when computer program is executed by processor.
The embodiment of the invention provides a kind of information recommendation method and device, computer readable storage medium, which is pushed away
The method of recommending includes: acquisition real-time streaming data;Real-time streaming data include: each accessed object mark and each access side for each quilt
Access the mapping relations at the login moment of object;According to mapping relations, access of each access side for each accessed object is determined
Retention amount;According to access retention amount, each access side is clustered, cluster result is obtained;According to cluster result, to each access side
Recommend accessed object corresponding with cluster centre.That is, on the one hand, the embodiment of the present invention obtains real-time streaming data,
Analysis can effectively be retained in real time based on real-time streaming data progress, another aspect is also sharp after obtaining access and retaining information
Each access side is clustered with information is retained, and carries out information recommendation for cluster result, so, it is possible based on each access side
Targetedly recommended, improves the flexibility and accuracy of information recommendation.
Detailed description of the invention
Fig. 1 is a kind of implementation process schematic diagram one of information recommendation method provided in an embodiment of the present invention;
Fig. 2 is a kind of implementation process schematic diagram two of information recommendation method provided in an embodiment of the present invention;
Fig. 3 is a kind of implementation process schematic diagram three of information recommendation method provided in an embodiment of the present invention;
Fig. 4 is a kind of implementation process schematic diagram four of information recommendation method provided in an embodiment of the present invention;
Fig. 5 is a kind of implementation process schematic diagram five of information recommendation method provided in an embodiment of the present invention;
Fig. 6 is a kind of implementation process schematic diagram six of information recommendation method provided in an embodiment of the present invention;
Fig. 7 is a kind of implementation process schematic diagram seven of information recommendation method provided in an embodiment of the present invention;
Fig. 8 is a kind of information recommending apparatus composed structure schematic diagram one provided in an embodiment of the present invention;
Fig. 9 is a kind of information recommending apparatus composed structure schematic diagram two provided in an embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description.
Embodiment one
The embodiment of the present invention provides a kind of information recommendation method, is applied in information recommending apparatus, and Fig. 1 is that the present invention is implemented
The implementation process schematic diagram one for a kind of information recommendation method that example provides, as shown in Figure 1, in one kind provided in an embodiment of the present invention
Information recommending apparatus in information recommendation method realizes that information recommendation method may include step 101 to step 104.It is as follows:
Step 101 obtains real-time streaming data.
In the embodiment of the present invention, in order to which the retention analysis for solving current calculates data volume using existing for off-line calculation model
Greatly and it is unable to get the problem of effectively retention is analyzed in real time, the embodiment of the present invention obtains real-time fluxion by information recommending apparatus
According to, wherein real-time streaming data includes: login moment of the mark of each accessed object with each access side for each accessed object
Mapping relations.
It should be noted that information recommending apparatus includes real-time streaming system and operation system, which is by industry
The data that business system generates carry out real-time collecting, transfer to stream process frame to carry out data processing, statistics, are put in storage, and by visual
The mode of change carries out real-time exhibition to statistical result;Retain analysis is to judge whether each access side accesses each accessed object lasting
A kind of statistical method, most can intuitively monitor whether each access side transforms into stable access side, to allow each quilt
The operator and exploitation for accessing object understand access side to the attraction degree of each accessed object.
In the embodiment of the present invention, the mark of above-mentioned each accessed object accesses each accessed object for characterizing access side
The page or element.
It should be noted that access type includes registration access type and login during access side accesses
Access type, according to different access types, corresponding processing mode is different, and therefore, the embodiment of the present invention is obtaining real-time streams
Need first to determine the access type of access side before.
Further, information recommending apparatus acquisition real-time stream may include: in the embodiment of the present invention
Obtain the access information that each access side accesses each access object;When determining access type is sign-on access, obtain
The login moment of each accessed object;According to the mark for logging in moment and each accessed object, real-time streaming data is obtained.
It should be noted that access information includes the mark and access type of each access object, access type is for characterizing
The access behavior of each access side is that sign-on access or registration access obtain access identities and step on when determination is to log in type
It records the moment, the real-time stream of generation can be the mark for being accessed object at Login-access side-login moment-;Work as determination
When being registration type, show that the user just registers, needs record to log in the moment next time, and then obtain real-time streaming data.
Illustratively, real-time stream can be Login-a-2018-06-26-A page.
Step 102, according to mapping relations, determine each access side for the access retention amount of each accessed object.
In the embodiment of the present invention, information recommending apparatus, according to mapping relations, determines each visit after obtaining real-time streaming data
Access retention amount of the side of asking for each accessed object.
It should be noted that access retention amount can be expressed as access side-first time login moment-login time it is poor-
Mark-number of accessed object, wherein login time difference be according to each access side for each accessed object login when
Determination is carved, number is to determine according to each access side in the number that the login moment accesses each access object.
Illustratively, access retention amount can be 10-A of a-2018-06-24-page-n, pass through the access retention amount
For available access side a after first time logs in accessed object as 2018-06-24, interval accesses this accessed pair for 10 days
The number of the corresponding A page of mark of elephant is n.
It should be noted that information recommending apparatus determines each access side for each accessed object according to mapping relations
Access retention amount may include: to carry out duplicate removal processing to the corresponding mapping relations of real-time stream in preset time period, obtain
Mapping relations after duplicate removal;Retention analysis is carried out to the mapping relations after duplicate removal, obtains each access side for each accessed object
Access retention amount, wherein access retention amount includes the retention amount and access side's retention amount of accessed object.
Step 103, according to access retention amount, each access side is clustered, cluster result is obtained.
In the embodiment of the present invention, information recommending apparatus determines each access side for each interviewed according to the mapping relations
After the access retention amount for asking object, according to each access side for the access retention amount of each accessed object, to each access side into
Row cluster, obtains cluster result.
It should be noted that information recommending apparatus can be interviewed each access side for each by K-means clustering algorithm
Ask that the access retention amount of object is analyzed, realization clusters each access side, wherein
K-means algorithm is that central point, that is, mass center of class is moved to Qi Bao by the process of the mobile class central point of a repetition
Then mean place containing member repartitions its internal members.Wherein, k is the calculated hyper parameter of algorithm, indicates the number of class
Amount;K-means can distribute sample to different classes automatically, but cannot determine to divide several classes actually;K must be a ratio
The small positive integer of training set sample number;
The quantity of class can be is specified by problem content;The parameter of K-means is centroid position and its internal sight of class
The position of measured value, similar with generalized linear model and decision tree, the optimal solution of K-means parameter is also with cost function minimum
Turn to target.
Cost function is the sum of each class distortion degree, and the distortion degree of each class is equal to such center and its internal members
The quadratic sum of positional distance.The distortion degree of class is smaller if the member inside class the compact to each other, conversely, if inside class
Member is bigger more the distortion degree for dispersing then class to each other.The parameter for solving cost function minimization is exactly that a repetition configures often
The observation that a class includes, offset does not move the process of class mass center.In fact, the position of observation can be randomly selected in centroid position
It sets.When each iteration,
K-means algorithm can be assigned to observation the class nearest from them, then mass center be moved to such all at
The average value of member position, the K-means cost function such as formula (1) are as follows:
Wherein, cost function V, k cluster number, XjRepresent data set, SiRepresent XjData set dimension, uiRepresent cluster
Center.
Step 104, according to cluster result, recommend accessed object corresponding with cluster centre to each access side.
In the embodiment of the present invention, information recommending apparatus is after clustering each access side, according to cluster result, to each
Access side recommends accessed object corresponding with cluster centre.
It should be noted that cluster result, which can be, obtains k cluster centre by K-means clustering algorithm, then pass through
The corresponding accessed object of the available each access side of the k cluster centre, it can obtain the interested quilt of each access side
Object is accessed, and is recommended corresponding access side.
Illustratively, access side a obtains the access side a by K-means algorithm and belongs to cluster centre A, then shows to cluster
The corresponding page of center A or element are the interested page of access side a or element, by the corresponding page of cluster centre A or
Person's element recommends access side as a as information to be recommended.
Fig. 2 is a kind of implementation process schematic diagram two of information recommendation method provided in an embodiment of the present invention, as shown in Fig. 2,
Information recommending apparatus in a kind of information recommendation method provided in an embodiment of the present invention determines each access side according to mapping relations
For the access retention amount of each accessed object, i.e. the process of step 102 may include step 102a and step 102b.It is as follows:
Step 102a, according to mapping relations, determine that each access side accesses the access times of each accessed object.
In the embodiment of the present invention, information recommending apparatus is according to the mark and each access side of each accessed object for each interviewed
The mapping relations for asking the login moment of object, determine that each access side accesses the access times of each accessed object.
It should be noted that showing that access side accesses quilt if access side has accessed accessed object on some date
The access times for accessing object are 1 time, if access side has accessed the accessed formation on the date again, show the access
The access times that side accesses accessed object are 2 times, are successively added up to access times, i.e., statistics available to obtain each access side's visit
Ask the access times of each access object.
Illustratively, real-time stream is Login-a-2018-06-30-A page, and access side a is in 2018-06-30
Having accessed the A page again is n times, then is n+1 in the access times of this day access side a access A page of 2018-06-30.
Step 102b, at the login moment that each accessed object is accessed according to access times and each access side, each visit is determined
Access retention amount of the side of asking for each accessed object.
In the embodiment of the present invention, information recommending apparatus obtain each access side access by the access times of each access object and
After logging in the moment, access of each access side for each accessed object can be obtained according to the access times and login moment
Retention amount.
Illustratively, real time access data flow is Login-a-2018-06-30-A page, when corresponding first time accesses
Between be 2018-06-24, access time difference be 6, access times 3, then available access retention amount be a -2018-06-
24-6-page-3 A indicate that access side a accesses the access of the A page behind the interval access time first time 2018-06-24 6 days
Number is 3.
Fig. 3 is a kind of implementation process schematic diagram three of information recommendation method provided in an embodiment of the present invention, as shown in figure 3,
Information recommending apparatus in a kind of information recommendation method provided in an embodiment of the present invention is according to access times and each access side
The login moment for accessing each accessed object determines each access side for the access retention amount of each accessed object, i.e. step
The process of 102b includes step 102b1, step 102b2 and step 102b3.It is as follows:
Step 102b1, according to each access side, each access side accesses each accessed object is searched from presetting database
It is primary to log in the moment.
What presetting database stored in the embodiment of the present invention, in information recommending apparatus is that each access side accesses each visit for the first time
At the time of asking object or at the time of each access side registers each access object.
Illustratively, the information of presetting database storage can be with are as follows: Login-a-2018-06-24-A page,
Login-b-2018-05-24-A page, Login-c-2018-04-24-A page.
Further, it is each interviewed to search each access side's access according to each access side from presetting database for information recommending apparatus
Ask that the first time of object logs in the moment, comprising:
According to each access side, each access side is obtained from presetting database and accesses the history identification of each accessed object and goes through
History logs in the moment;When determining that history identification is consistent with the mark of each accessed object, using the historical log time as each access
The first time that side accesses each accessed object logs in the moment.
It should be noted that when if the real-time stream that information recommending apparatus obtains is Login-access side-login
Quarter-each accessed object identity it is each then to need to find out access side access from default access database by each access side
The corresponding historical data stream of object is accessed, historical data stream includes the history identification that the access side accesses each access object for the first time
The corresponding historical log moment;If obtained real-time stream is Register-access side-registration moment, record
The primary access moment, be then based on next time access the moment handled, obtain corresponding retention amount.
If it is determined that history identification is consistent with the mark of accessed object, then show that the access side has accessed identical history
Mark can find the historical log moment for accessing the history identification for the first time from database;If it is determined that present count
According to not with the consistent history identification of the mark of accessed object, then showing that the access side is that access this for the first time accessed in library
The access identities of object are stored in default in this regard, the information for needing to access the access side the accessed object records
In database, it is convenient for next calculating.
Illustratively, real-time stream is Login-a-2018-06-30-A page, searches and corresponds to from preset data
Historical data stream be Login-a-2018-06-24-A page, it is determined that the mark one of history identification and accessed object
It causes, accordingly the first time of the available access A page logs in the moment as 2018-06-24.
When real-time stream is Login-a-2018-06-30-A page, access side a does not have in corresponding presetting database
The corresponding A page of access identities was accessed, then was recorded the information that the access side accesses the accessed object, i.e.,
Login-a-2018-06-30-A page, storage is in the preset database.
Step 102b2, the login moment of each accessed object is accessed according to each access side and first time logs in the moment, obtained
Log in time difference.
In the embodiment of the present invention, information recommending apparatus logs in the moment at the login moment for obtaining real-time stream and for the first time
Later, it by logging in moment progress difference processing to the login moment and first time, can obtain logging in time difference.
Illustratively, searching login moment first time that presetting database obtains is 2018-06-24, each access side's access
The login moment of each accessed object is 2018-06-30, then corresponding login time difference is 6.
Step 102b3, according to time difference and access times are logged in, determine that each access side stays the access of each access object
Storage.
In the embodiment of the present invention, information recommending apparatus, can be according to stepping on after obtaining and logging in time difference and access times
Time difference and access times are recorded, determine each access side for the access retention amount of each access object.
Illustratively, if real-time stream is Login-a-2018-06-30-A page, on the day of 2018-06-30
The A page was accessed again, and corresponding historical data stream is Login-a-2018-06-24-A page in the preset database
Face, then available corresponding login time difference is 6, access times 2, and then according to time difference and access times are logged in, really
The retention amount of the fixed access side a access A page is the page-2 2-A of a-2018-06-24-.
Fig. 4 is a kind of implementation process schematic diagram four of information recommendation method provided in an embodiment of the present invention, as shown in figure 4,
Information recommending apparatus in a kind of information recommendation method provided in an embodiment of the present invention is according to access retention amount, to each access
Before Fang Jinhang cluster, i.e., before step 103, information recommending apparatus can also include step 105.It is as follows:
Step 105, according to each access side in predetermined time length for the access retention amount of each accessed object whether
Lower than preset value, the type of each access side is determined.
In the embodiment of the present invention, information recommending apparatus can exist before clustering to each access side according to access side
Whether preset value is lower than for the access retention amount of each accessed object in predetermined time length, determines the type of each access side.
It should be noted that access side's type may include loyal access side and non-loyalty access side, wherein loyalty access
Side is higher than preset value for the access retention amount of each accessed object for being characterized in predetermined time length;Non- loyalty access side
Preset value is lower than for the access retention amount of each accessed object in predetermined time length for characterizing the access side.
In the embodiment of the present invention, information recommending apparatus is by access side for each accessed object in predetermined time period
Access retention amount, available each access side accesses the number of each accessed object, and then can be by access times and pre-
If value is compared, to determine the type of each access side, when access times are higher than preset value, determine that each access side is loyal visit
The side of asking determines that each access side is non-loyal access side when access times are lower than preset value.
It should be noted that predetermined time period can be configured according to actual needs, such as when can be set default
Between length be one month.
Illustratively, access side a is in predetermined time period for the access retention amount of each accessed object are as follows: a-
2018-06-24-n-A the page-n, a-2018-06-24-n-B page-q, a-2018-06-24-n-C page-
X can be obtained, and the number that access side a accesses each accessed object is n+q+x.
Information recommending apparatus is compared by access times and preset value, before the type to determine each access side,
It also needs to set and enlivens access side, such as one month login number of days of setting was to enliven access side greater than 20 days, passed through active visit
The number mean value that the access side in preset time period accesses each accessed object is obtained in the retention amount for the side of asking.
Illustratively, predetermined time period can be set as one month, corresponding preset value is 20.
Accordingly, the access retention amount of each accessed object clusters each access side according to each access side, wraps
It includes:
When the type for determining each access side is preset kind, preset kind accordingly each accessed object is monitored;When true
Determine preset kind accordingly each accessed object variation when, according to preset kind accordingly each accessed object access retain
Amount, to preset kind, accordingly each access side is clustered.
In the embodiment of the present invention, preset kind is for characterizing access side in predetermined time length for each accessed object
Access retention amount be higher than preset value.
Based on above-mentioned it is found that access side is higher than in advance the access retention amount of each accessed object in predetermined time length
If value can correspond to loyal access side, that is to say, that the loyal access side of monitoring of the embodiment of the present invention is corresponding each accessed
Whether object changes.
It should be noted that the embodiment of the present invention can be through access times the type for determining each access side, one
After the type of denier access side changes, then show that the number of the accessed object of access side's access changes, and then can be true
Surely the information for recommending the access side, which may not meet its interest, leads to its change, therefore, for the preset kind pair after variation
Each accessed object is answered to need to redefine the information to be recommended recommended to it.
Fig. 5 is a kind of implementation process schematic diagram five of information recommendation method provided in an embodiment of the present invention, as shown in figure 5,
Information recommending apparatus in a kind of information recommendation method provided in an embodiment of the present invention is according to each access side for each accessed
The access retention amount of object, clusters each access side, i.e. the process of step 103 may include step 103a and step
103b.It is as follows:
Step 103a, according to each accessed object, initial mass center is set.
In the embodiment of the present invention, information recommending apparatus is retaining the access of each accessed object according to each access side
Amount, during clustering to each access side, needs to set initial mass center, could be based on the initial mass center and be clustered.
Step 103b, according to access retention amount and initial mass center, each access side is clustered, cluster result is obtained.
In the embodiment of the present invention, information recommending apparatus can pass through K-means clustering algorithm after setting initial mass center
Each access side of access retention amount and initial mass center to to(for) each accessed object, cluster each access side.
It should be noted that K-means clustering algorithm may comprise steps of:
1) an initial mass center is determined for each cluster, i.e., select K data from data set, as K initial mass centers.
2) sample is assigned in closest cluster according to minimal distance principle.
It should be noted that the distance between two points size is measured using Euclidean distance formula, such as two point T0
(x1, y2) and T1 (x2, y2), the Euclidean distance d such as formula (2) between T0 and T1 are as follows:
D=sqrt ((x1-x2)2+(y1-y2)2) (2)
3) use the sample average in each cluster as new cluster centre.
4) it repeats step 2) and 3) until cluster centre no longer changes, obtains k and cluster.
It should be noted that K-means clustering algorithm is unsupervised learning mode, for big data without carrying out in advance
Training and study, can directly carry out statistic of classification.
Fig. 6 is a kind of implementation process schematic diagram six of information recommendation method provided in an embodiment of the present invention, as shown in fig. 6,
Information recommending apparatus in a kind of information recommendation method provided in an embodiment of the present invention is pushed away according to cluster result to each access side
Accessed object corresponding with cluster centre is recommended, i.e. step 104 includes step 104a, 104b and step 104c.It is as follows:
Step 104a, according to cluster result, cluster centre belonging to each access side is determined.
In the embodiment of the present invention, information recommending apparatus needs root before obtaining the corresponding accessed object of cluster centre
According to cluster result, cluster centre belonging to each access side is determined.
Illustratively, cluster result can be access side a [10,10,5,10] corresponding cluster centre A;Access side b [20,5,
10,20] corresponding cluster centre B, wherein a [10,10,5,10] is for characterizing the retention that different access side accesses each accessed object
Amount.
Step 104b, from each access object, accessed object corresponding with cluster centre is obtained.
Step 104c, recommend each accessed object to each access side.
In the embodiment of the present invention, the corresponding each access pair of information recommending apparatus available cluster centre from cluster result
As recommending each access side using the corresponding each access object of the cluster centre as recommendation information.
It should be noted that the embodiment of the present invention is to access loyal access side after each access object changes, i.e. loyalty
When sincere access side is changing into non-loyal access side, is further classified to loyalty access by clustering algorithm, in this way can
The interested each access object of loyal access side after obtaining variation, and then realize and recommend for the access side after variation.
It is understood that the embodiment of the present invention distinguishes the different type of access side by access retention amount, then based on visit
The type for the side of asking carries out information recommendation, the information of recommendation can be so adjusted in time for the type of different access object, is gone forward side by side
The targeted recommendation information of row.
A kind of information recommendation method provided through the embodiment of the present invention, on the one hand, according to access time and accession page
The retention amount of different objects is analyzed, treatment process can be simplified, reduces data processing amount, and clearly interested target object
On the specific page;On the other hand, classify to access object, targetedly recommended based on access object type,
It can be improved the accuracy of information recommendation.
Embodiment two
Same inventive concept based on embodiment one, the information recommendation side that embodiment provides in order to better illustrate the present invention
Method corresponds to accession page with the mark of each accessed object and illustrates, and the method that above- mentioned information are recommended is applied and is pushed away in information
It recommending in device, Fig. 7 is a kind of implementation process schematic diagram seven of information recommendation method provided in an embodiment of the present invention, as described in Figure 7,
Information recommending apparatus realizes that information recommendation method may comprise steps of:
Step 201 establishes database according to access side's behavioral data.
In the embodiment of the present invention, access side's behavior includes sign-on access behavior and registration access behavior, accordingly, access row
It include first time logon data and log-on data for data, information recommending apparatus establishes database according to access side's behavioral data,
Including establishing log form and establishing registration table.
Illustratively, above-mentioned database can be non-relational database redis.
It should be noted that information recommending apparatus establish database include: firstly, to the registion time of different access side into
Row storage, stores the exact date of access side and access side's registration, and setting key assignments key is that " access side-access side's registration is specific
Date ", and then obtain corresponding registration table.Secondly, the first time login time to different access side stores, need to store
Three column datas are the specific page of access side, the exact date that access side logs in for the first time and access side's access, setting respectively
Key is " the specific page for the exact date-access side access that access side-access side logs in for the first time ", obtains corresponding step on
Record table.
Illustratively, registration table such as table 1 and log form such as table 2.
Table 1
Register—a—2018-06-10 |
Register—b—2018-05-10 |
Register—c—2018-04-10 |
Register—d—2018-04-10 |
Table 2
Login-a-2018-06-24-A page |
Login-b-2018-05-24-A page |
Login-c-2018-04-24-A page |
Step 202 obtains real time access data flow.
In the embodiment of the present invention, information recommending apparatus collects the behavioral data of access side in real-time stream, access
Square behavioral data includes access time and access information according to access side's behavioral data determining access side's behavior type;Work as access
When square behavior type is sign-on access type, corresponding obtained real-time stream is Login-access side-login moment-visit
The specific page asked;When determination is registration type, at the record registration moment, it is convenient for subsequent login.
Illustratively, real-time stream can be Login-a-2018-06-26-A page.
Step 203, according to database and real-time stream, obtain the retention amount of access side's accession page.
In the embodiment of the present invention, if receiving real-time stream in information recommending apparatus are as follows: Login-a -2018-06-
26-A the pages, by the log form of the access side a finding step 201, the login moment for obtaining access side's a first time is 2018-
06-24, calculating and this time logging in moment 2018-06-26 with the difference for logging in moment 2018-06-24 for the first time is 2, according to difference
The retention amount of the obtained access side a access A page are as follows: the page-1 2-A of a-2018-06-24-;If on the day of access side a again
The A page is had accessed, then is added up, obtains the retention amount of the access side a access A page are as follows: a-2018-06-24-2-A page
The retention amount in face-2, the available access side a access A page that successively adds up is 2-A of a-2018-06-24-page-n.
If receiving real-time stream in information recommending apparatus are as follows: Login-a-2018-06-30-A page, accordingly
Access side a has logged in the A page in login moment 2018-06-30, is by the login moment that the access side a finds first time
2018-06-24 calculates the difference for this time logging in moment 2018-06-30 and logging in moment 2018-06-24 for the first time are as follows: and 6, it obtains
To the retention amount of the access side a access A page are as follows: the page-1 6-A of a-2018-06-24-.
If receiving data flow in information recommending apparatus are as follows: Register-d -2018-04-10 then illustrates this access side
D is just registered, and is also not logged in, and needs to record login time next time, the superposition that can be just worth.
If receiving real-time stream in information recommending apparatus are as follows: Login-a-2018-07-24-B page, by looking into
It looks for the log form of step 201 that can obtain the access side a before this and has not visited the B page, then login is recorded in the real-time stream
In table.
The embodiment of the present invention, information recommending apparatus calculate real-time stream by above-mentioned, can obtain access side at certain
The number of accession page, such as obtains the retention amount of access side's accession page are as follows: 10-A of a-2018-06-24-in the section time difference
The page-n then shows that the 4 login moment for logging in the A page were 2018-06-2 to access side a for the first time, and the same day access after 10 days should
The number of the A page is n.
Step 204 handles the retention amount of access side's accession page in preset time period, and it is default to obtain access side
The retention amount of accession page in the section time.
Illustratively, the retention amount of access side's accession page in preset time period are as follows: 10-A of a-2018-06-24-
The page-2018-05-24-10 the n, b-- A page-2018-04-24-10 the m, c-- A page-p, a-2018-06-
24-10-B page-2018-05-24-10 the q, b-- B page-2018-04-24-10 the r, c-- B pages-s, a-
10-C of 2018-06-24-page-2018-05-24-10 x, b-- C page-2018-04-24-10 y, c-- C pages
Face-z.
By the retention amount of above-mentioned access side's accession page it is found that access side's first time login page, is visited after 10 days
The total retention amount for asking the A page is n+m+p, and total retention amount of the B page is q+r+s, and total retention amount of the C page is x+y+z, this is interviewed
Ask that object assumes there was only A, tri- pages of B, C, then total retention amount of this accessed object is all:n+m+p+q+r+s+x+y+z.A
The retention ratio of the page is (n+m+p)/all, and the retention ratio of the B page is (q+r+s)/all, the retention ratio of the C page be (x+y+z)/
all。
User's retention amount after corresponding first time registration also can be obtained by the retention amount of above-mentioned access side's accession page
And retention ratio: access side is counted, staying after 10 days after having registered this accessed object for the first time using duplicate removal
Storage be tri- access sides of a, b, c, then access side's retention amount be 3, the access side's quantity registered at the beginning as a+b+c+d, always
User's retention ratio is 3/ (a+b+c+d).Step 205 is carried out by the retention amount for presetting accession page in the section time to access side
Analysis determines loyal access side
In the embodiment of the present invention, the retention amount that information recommending apparatus presets accession page in the section time to access side is divided
Analysis includes: statistics access side access day within a preset period of time;All pages accessed object on are accessed access side daily
Number carry out cumulative statistics, successively can handy access side time of all pages on accessed object is accessed in section at any time
It counts, the frequency table of all pages on an available access target object.
Illustratively, if real time access statistical data stream is a-2018-06-24-n-A page-n, a-2018-
06-24-n-B the page-q, a-2018-06-24-n-C page-x to a-2018-06-24-n-N page-z, then
The number for obtaining all pages on the daily access target object of user equipment is n+q+x+ ...+z.
Assuming that preset time period is one month, it is to enliven access side that a month access day, which is arranged, greater than 20 days, simultaneously
It is enlivened in the corresponding frequency table of access side from what is obtained, takes out its number for accessing this accessed object daily in one month
Mean value, it is corresponding, by access side's access times and the number mean value, and the access side on accessed object can be divided into
Loyal access side and non-loyalty access side specifically when access side's access times are greater than or equal to number mean value, determine the visit
The side of asking is loyal access side, when access side's access times are less than number mean value, determines that the access side is non-loyal access side.
Illustratively, setting the number mean value accessing this accessed object daily in one month is 20, then if accessed
The daily number for accessing accessed object in square a mono- month is more than or equal to 20, determines corresponding access side a for loyal access
Otherwise side is determined as non-loyal access side.
For loyal access side, as long as whether the such loyal access side of real-time oversight becomes non-loyal access side, once become
At non-loyal access side, then the non-loyal access side is clustered.
Step 206, when monitoring that loyal access side is changing into non-loyal access side, K- is carried out to non-loyal access side
Means algorithm obtains information to be recommended.
In the embodiment of the present invention, information recommending apparatus does K-means clustering algorithm for non-loyal access side, chooses non-
The corresponding accession page of non-loyalty access side of loyal access side, the selection of timesharing phase carry out personalization to non-loyal access side
Recommend, and retention calculating is carried out to the data after personalized recommendation, and whether the access day checked in fixed time period increases
It is more, compare do personalized recommendation to the impact effect of access side's quantity using the different pages respectively, thus obtain can most attract it is non-
The page of loyal access side, just using this page as the homepage of personalized recommendation.
Illustratively, as the data set of table 3, the table 3 have counted the Z class page, the data of every row represent each specific user
Within some time difference, the retention amount of each page is accessed.
Table 3
The A page | The B page | The C page | …… | The Z page |
10 | 10 | 5 | …… | 10 |
20 | 5 | 10 | …… | 20 |
…… | …… | …… | …… | …… |
5 | 10 | 20 | …… | 1 |
Assuming that the data of 2000 people have been counted, using the data of above-mentioned form as data set.Setting k is Z, accordingly, should
Data set one shares Z mass center, has then selected and has taken 1500 samples as test set, and accordingly, the content one of output is shared
1500 rows.
By carrying out class statistic to above-mentioned data set, obtain exporting Z center-of-mass coordinate data in experimentation are as follows: the 0th
Type cluster centre coordinate is [a, b, c ...];1st type cluster centre coordinate is [d, e, f ...];2nd type cluster centre
Coordinate is [i, j, k ...];……;Z-1 type cluster centre coordinate is [x, y, z ...], wherein 0-Z cluster centre difference
Represent be divided into Z class.
The experimental result of obtained cluster output are as follows: access side a [10,10,5 ... ..10] belongs to the 1st type cluster centre;
Access side b [20,5,10 ... ..20] belongs to the 1st type cluster centre;……;Access side z [5,10,20 ... ..1] belongs to the 1st class
Type cluster centre.
By above-mentioned experimental result it is found that it is each cluster centre coordinate that this accessed object, which can most attract the page of access side,
The corresponding page, thus the recommendation to being customized of access side.In addition, in a practical situation, some access side is usually not
Only there is an interest, if expecting the front three of the page of this access side concern, can often be obtained in code calculating process
The center clustered to one, calculate the coordinate points for saving this access side's accession page between cluster centre point at a distance from, and every
During an iteration, this distance is iterated.In last output result, by this coordinate points at a distance from each central point
It is compared, exports the front three page.
Meanwhile having any problem if realizing and carrying out individualized content recommendation to the corresponding page of each access side, also it can use
Each access side is gathered into limited class by K-means algorithm, realizes the classification of each access side, is carried out for each access side of limited class a
Property content it is basic matching and recommend, reduce difficulty while, can also match the hobby content of user.
Step 207 gives information recommendation to be recommended to non-loyal access side.
Through the embodiment of the present invention, the moment is logged in access side and login times calculates access side and access accessed object
Retention amount can either simplify treatment process, while can excavate the corresponding specific interested page in different access side;It is another
Aspect, the embodiment of the present invention first classify to access side, then are carried out targetedly based on the non-loyal access side after variation
Recommend, recommendation information can be adjusted in real time, improve the accuracy of information recommendation.
Embodiment three
Based on the same inventive concept of embodiment one to embodiment two, the embodiment of the present invention provides a kind of information recommendation dress
It sets, Fig. 8 is a kind of information recommending apparatus composed structure schematic diagram one provided in an embodiment of the present invention, as shown in figure 8, in the present invention
In embodiment, information recommending apparatus 1000 includes acquiring unit 1001, determination unit 1002, cluster cell 1003 and recommendation unit
1004, wherein
The acquiring unit 1001, for obtaining real-time streaming data;The real-time streaming data includes: each accessed object
Mark is with each access side for the mapping relations at the login moment of each accessed object;
The determination unit 1002, for determining each access side for described each interviewed according to the mapping relations
Ask the access retention amount of object;
The cluster cell 1003, for clustering, being gathered to each access side according to the access retention amount
Class result;
The recommendation unit 1004, for recommending and cluster centre phase to each access side according to the cluster result
Corresponding accessed object.
In other embodiments, the determination unit 1002 may include:
First determination unit 1002a, for determining that each access side's access is described each interviewed according to the mapping relations
Ask the access times of object;When accessing the login of each accessed object according to the access times and each access side
It carves, determines each access side for the access retention amount of each accessed object.
In other embodiments, described device 1000 can also include:
Judging unit 1005, for according to the access side in predetermined time length for each accessed object
Whether access retention amount is lower than preset value, determines the type of each access side.
In other embodiments, the cluster cell 1003 may include:
First cluster cell 1003a, for setting initial mass center according to each accessed object;According to the access
Retention amount and the initial mass center, cluster each access side.
In other embodiments, the recommendation unit 1004 may include:
Second determination unit 1004a, for determining cluster centre belonging to each access side according to cluster result;
First acquisition unit 1004b, it is used for from each access object, obtains corresponding with the cluster centre interviewed
Ask object;
First recommendation unit 1004c, for recommending each accessed object to each access side.
In other embodiments, the first determination unit 1002a may include:
Searching unit 1002a1, for searching each access side from presetting database and visiting according to each access side
Ask that the first time of each accessed object logs in the moment;
Third determination unit 1002a2, for accessing the login moment of each accessed object according to each access side
The moment is logged in the first time, obtains and logs in time difference;
The third determination unit 1002a3, described in determining according to the login time difference and the access times
Access retention amount of each access side for each access object.
Through the embodiment of the present invention, the moment is logged in access side and login times calculates access side and access accessed object
Retention amount can either simplify treatment process, while can excavate the corresponding specific interested page in different access side;It is another
Aspect, the embodiment of the present invention first classify to access side, then are carried out targetedly based on the non-loyal access side after variation
Recommend, recommendation information can be adjusted in real time, improve the accuracy of information recommendation.
Example IV
Based on the same inventive concept of embodiment one to embodiment three, a kind of information recommendation dress provided in an embodiment of the present invention
It sets, Fig. 9 is a kind of information recommending apparatus composed structure schematic diagram two provided in an embodiment of the present invention, as shown in figure 9, above- mentioned information
Recommendation apparatus includes at least processor 01, memory 02, communication interface 03 and communication bus 04, wherein communication bus 04 is used for
Realize the connection communication between processor and memory;Communication interface 03 is sended and received for information;Processor 01 is used for
The information recommendation program that stores in memory 02 is executed, to realize information recommendation side that above-described embodiment one is provided to embodiment two
Step in method.
Through the embodiment of the present invention, the moment is logged in access side and login times calculates access side and access accessed object
Retention amount can either simplify treatment process, while can excavate the corresponding specific interested page in different access side;It is another
Aspect, the embodiment of the present invention first classify to access side, then are carried out targetedly based on the non-loyal access side after variation
Recommend, recommendation information can be adjusted in real time, improve the accuracy of information recommendation.
Based on previous embodiment, the embodiment of the invention provides a kind of computer readable storage mediums, are stored thereon with letter
Recommended program is ceased, the information recommendation in said one or multiple embodiments is realized when above- mentioned information recommended program is executed by processor
Method.
It should be understood by those skilled in the art that, the embodiment of the present invention can provide as method, system or computer program
Product.Therefore, the shape of hardware embodiment, software implementation or embodiment combining software and hardware aspects can be used in the present invention
Formula.Moreover, the present invention, which can be used, can use storage in the computer that one or more wherein includes computer usable program code
The form for the computer program product implemented on medium (including but not limited to magnetic disk storage and optical memory etc.).
The present invention be referring to according to the method for the embodiment of the present invention, the process of equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
The foregoing is only a preferred embodiment of the present invention, is not intended to limit the scope of the present invention.
Claims (10)
1. a kind of information recommendation method, which is characterized in that the described method includes:
Obtain real-time streaming data;Mark that the real-time streaming data includes: each accessed object and each access side are for described each
The mapping relations at the login moment of accessed object;
According to the mapping relations, determine each access side for the access retention amount of each accessed object;
According to the access retention amount, each access side is clustered, cluster result is obtained;
According to the cluster result, recommend accessed object corresponding with cluster centre to each access side.
2. determining each access the method according to claim 1, wherein described according to the mapping relations
Access retention amount of the side for each accessed object, comprising:
According to the mapping relations, determine that each access side accesses the access times of each accessed object;
The login moment that each accessed object is accessed according to the access times and each access side determines described each
Access retention amount of the access side for each accessed object.
3. the method according to claim 1, wherein described according to the access retention amount, to each access
Fang Jinhang cluster, before obtaining cluster result, further includes:
Whether the access retention amount of each accessed object is lower than in predetermined time length according to the access side pre-
If value determines the type of each access side.
4. the method according to claim 1, wherein described according to the access retention amount, to each access
Fang Jinhang cluster, obtains cluster result, comprising:
According to each accessed object, initial mass center is set;
According to the access retention amount and the initial mass center, each access side is clustered, the cluster result is obtained.
5. the method according to claim 1, wherein described according to the cluster result, to each access side
Recommend accessed object corresponding with cluster centre, comprising:
According to the cluster result, cluster centre belonging to each access side is determined;
From each access object, accessed object corresponding with the cluster centre is obtained;
Recommend the accessed object to each access side.
6. according to the method described in claim 2, it is characterized in that, described according to the access times and each access side
The login moment for accessing each accessed object determines that each access side retains the access of each accessed object
Amount, comprising:
According to each access side, each access side accesses each accessed object first is searched from presetting database
The secondary login moment;
The login moment of each accessed object is accessed according to each access side and the first time logs in the moment, and acquisition is stepped on
Record time difference;
According to the login time difference and the access times, access of each access side for each access object is determined
Retention amount.
7. according to the method described in claim 3, it is characterized in that, described according to the access retention amount, to each access
Fang Jinhang cluster, obtains cluster result, comprising:
When the type for determining each access side is preset kind, the preset kind accordingly each accessed object is monitored,
The preset kind is used to characterize the access side and the access of each accessed object is retained in predetermined time length
Amount is higher than the preset value;
It is accordingly each accessed according to the preset kind when determining preset kind accordingly each accessed object variation
The access retention amount of object, to the preset kind, accordingly each access side is clustered, and obtains cluster result.
8. a kind of information recommending apparatus, which is characterized in that described device includes:
Acquiring unit, for obtaining real-time streaming data;The real-time streaming data includes: the mark and each access of each accessed object
Mapping relations of the side for the login moment of each accessed object;
Determination unit, for determining access of each access side for each accessed object according to the mapping relations
Retention amount;
Cluster cell, for being clustered to each access side, obtaining cluster result according to the access retention amount;
Recommendation unit, for recommending to each access side corresponding with cluster centre accessed according to the cluster result
Object.
9. a kind of information recommending apparatus, which is characterized in that the information recommending apparatus includes at least processor, is stored with the place
Memory, the communication interface of device executable instruction are managed, and is connect for connecting the processor, the memory and the communication
The bus of mouth, when executed, the processor realize side as described in any one of claim 1 to 7 when executing
Method.
10. a kind of computer readable storage medium, stores computer program thereon, which is characterized in that the computer program quilt
Processor realizes method as described in any one of claim 1 to 7 when executing.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811605413.7A CN109710876B (en) | 2018-12-26 | 2018-12-26 | Information recommendation method and device and computer readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811605413.7A CN109710876B (en) | 2018-12-26 | 2018-12-26 | Information recommendation method and device and computer readable storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109710876A true CN109710876A (en) | 2019-05-03 |
CN109710876B CN109710876B (en) | 2021-08-06 |
Family
ID=66257781
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811605413.7A Active CN109710876B (en) | 2018-12-26 | 2018-12-26 | Information recommendation method and device and computer readable storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109710876B (en) |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1804844A (en) * | 2006-01-10 | 2006-07-19 | 西安交通大学 | Web page metadata based formalized description method for user access behaviors |
US20080243815A1 (en) * | 2007-03-30 | 2008-10-02 | Chan James D | Cluster-based assessment of user interests |
CN101685521A (en) * | 2008-09-23 | 2010-03-31 | 北京搜狗科技发展有限公司 | Method for showing advertisements in webpage and system |
CN102158365A (en) * | 2011-05-20 | 2011-08-17 | 北京邮电大学 | User clustering method and system in weblog mining |
CN102637178A (en) * | 2011-02-14 | 2012-08-15 | 北京瑞信在线系统技术有限公司 | Music recommending method, music recommending device and music recommending system |
CN104063801A (en) * | 2014-06-23 | 2014-09-24 | 广州优蜜信息科技有限公司 | Mobile advertisement recommendation method based on cluster |
CN106528778A (en) * | 2016-11-04 | 2017-03-22 | 广州华多网络科技有限公司 | Method and device for obtaining user retention data |
CN107483613A (en) * | 2017-08-31 | 2017-12-15 | 江西博瑞彤芸科技有限公司 | A kind of information-pushing method |
CN107944059A (en) * | 2017-12-29 | 2018-04-20 | 深圳市中润四方信息技术有限公司西安分公司 | A kind of user behavior analysis method and system based on stream calculation |
-
2018
- 2018-12-26 CN CN201811605413.7A patent/CN109710876B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1804844A (en) * | 2006-01-10 | 2006-07-19 | 西安交通大学 | Web page metadata based formalized description method for user access behaviors |
US20080243815A1 (en) * | 2007-03-30 | 2008-10-02 | Chan James D | Cluster-based assessment of user interests |
CN101685521A (en) * | 2008-09-23 | 2010-03-31 | 北京搜狗科技发展有限公司 | Method for showing advertisements in webpage and system |
CN102637178A (en) * | 2011-02-14 | 2012-08-15 | 北京瑞信在线系统技术有限公司 | Music recommending method, music recommending device and music recommending system |
CN102158365A (en) * | 2011-05-20 | 2011-08-17 | 北京邮电大学 | User clustering method and system in weblog mining |
CN104063801A (en) * | 2014-06-23 | 2014-09-24 | 广州优蜜信息科技有限公司 | Mobile advertisement recommendation method based on cluster |
CN106528778A (en) * | 2016-11-04 | 2017-03-22 | 广州华多网络科技有限公司 | Method and device for obtaining user retention data |
CN107483613A (en) * | 2017-08-31 | 2017-12-15 | 江西博瑞彤芸科技有限公司 | A kind of information-pushing method |
CN107944059A (en) * | 2017-12-29 | 2018-04-20 | 深圳市中润四方信息技术有限公司西安分公司 | A kind of user behavior analysis method and system based on stream calculation |
Non-Patent Citations (1)
Title |
---|
唐胡鑫: "《网站运营与数据分析 双色版》", 31 December 2016, 航空工业出版社 * |
Also Published As
Publication number | Publication date |
---|---|
CN109710876B (en) | 2021-08-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106506705B (en) | Crowd classification method and device based on location service | |
CN106021364B (en) | Foundation, image searching method and the device of picture searching dependency prediction model | |
Paindaveine et al. | Computing multiple-output regression quantile regions | |
CN104298679A (en) | Application service recommendation method and device | |
El Mabrouk et al. | Towards an intelligent hybrid recommendation system for e-learning platforms using data mining | |
CN108595461A (en) | Interest heuristic approach, storage medium, electronic equipment and system | |
CN111489201A (en) | Method, device and storage medium for analyzing customer value | |
CN110110237A (en) | User interest information recommended method, storage medium | |
CN113222170B (en) | Intelligent algorithm and model for AI collaborative service platform of Internet of things | |
CN106874943A (en) | Business object sorting technique and system | |
CN106251260A (en) | A kind of candidates' aspiration makes a report on analog systems and method | |
CN110020152A (en) | Using recommended method and device | |
CN102141988B (en) | Method, system and device for clustering data in data mining system | |
CN106599291A (en) | Method and device for grouping data | |
CN109657695A (en) | A kind of fuzzy division clustering method and device based on definitive operation | |
CN109710876A (en) | A kind of information recommendation method and device, computer readable storage medium | |
CN107844536A (en) | The methods, devices and systems of application program selection | |
CN104462480B (en) | Comment big data method for digging based on typicalness | |
CN110413926A (en) | A kind of questionnaire method and device | |
FI130045B (en) | Analyzing measurement results of a communications network or other target system | |
CN114065038A (en) | Big data-based head information recommendation method and device | |
CN109271303B (en) | Software configuration recommendation method | |
CN108062395A (en) | A kind of track traffic big data analysis method and system | |
Guo et al. | A type-aware workload prediction strategy for non-stationary cloud service | |
CN114186168A (en) | Correlation analysis method and device for intelligent city network resources |
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 |