CN109741098A - Broadband off-network prediction technique, equipment and storage medium - Google Patents
Broadband off-network prediction technique, equipment and storage medium Download PDFInfo
- Publication number
- CN109741098A CN109741098A CN201811615555.1A CN201811615555A CN109741098A CN 109741098 A CN109741098 A CN 109741098A CN 201811615555 A CN201811615555 A CN 201811615555A CN 109741098 A CN109741098 A CN 109741098A
- Authority
- CN
- China
- Prior art keywords
- network
- user
- broadband
- probability
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention provides a kind of broadband off-network prediction technique, equipment and storage medium, passes through the user characteristics for obtaining any broadband user in existing net；And in the broadband off-network prediction model for obtaining user characteristics input in advance, the off-network probability and off-network for obtaining the existing net middle width strip user may reasons；Then determine that the off-network of existing net middle width strip user is inclined to according to the off-network probability and the possible reason of the off-network.The broadband off-network prediction model obtained in the present invention by the user characteristics of existing net middle width strip user and in advance, it can accurately predict the off-network tendency and the possible reason of off-network of existing net middle width strip user, and then it can instruct operator that targetedly existing network users are maintained and serviced, realize the precision marketing to user and maintain.
Description
Technical field
The present invention relates to field of communication technology more particularly to a kind of broadband off-network prediction techniques, equipment and storage medium.
Background technique
Home broadband business is the main operation business that operator manages communication network, and broadband user's maintains and develop straight
Connect the market development scale for being related to operator.Therefore, the work that prediction broadband off-network user maintains user and strategy is marketed
With very great.
Currently, for predicting broadband off-network user, it can be according to the corresponding broadband associated data of broadband account be taken, instead
Terminal is reflected the degree of dependence in broadband is come to carry out off-network prediction to the target broadband account in network.But existing prediction
Method usually only by decision tree, can not compare superiority and inferiority, cause forecasting accuracy lower.
Summary of the invention
The present invention provides a kind of broadband off-network prediction technique, equipment and storage medium, to improve existing net middle width strip user's
Off-network is inclined to forecasting accuracy, while can also predict the possible reason of the off-network of broadband user, and operator can be instructed targeted
Ground is maintained and is serviced to existing network users, is realized the precision marketing to user and is maintained.
The first aspect of the present invention is to provide a kind of broadband off-network prediction technique, comprising:
Obtain the user characteristics of any broadband user in existing net；
In the broadband off-network prediction model that user characteristics input is obtained in advance, the existing net middle width strip user is obtained
Off-network probability and off-network may reason；
The off-network tendency of existing net middle width strip user is determined according to the off-network probability and the possible reason of the off-network.
Further, the method also includes:
The a variety of user properties for obtaining off-network user, as training data and test data；
According to the training data at least two machine learning algorithms of training, alternative model is obtained respectively；
The predictablity rate of each alternative model is tested according to the test data；
Select the highest alternative model of predictablity rate as optimal alternative model, by neural network algorithm to it is described most
Excellent alternative model optimizes, and obtains the broadband off-network prediction model, and the broadband off-network prediction model is used for according to user
Feature exports off-network probability and influences the different degree ranking of the user characteristics of off-network probability.
Further, after a variety of user properties for obtaining off-network user, further includes:
There are the data of absent field for acquisition, and fill up to absent field.
It is further, described that absent field is filled up, comprising:
Judge whether the absent field belongs to Gaussian Profile；
If the absent field belongs to Gaussian Profile, the missing word is filled up using gauss of distribution function random value
Section；
If the absent field is not belonging to Gaussian Profile, gauss of distribution function is constructed using method for normalizing, is recycled
Gauss of distribution function random value is to fill up the absent field.
Further, after a variety of user properties for obtaining off-network user, further includes:
Obtain the variance expansion factor in a variety of user properties of off-network user between any two users' attribute；
The user property that variance expansion factor is less than preset threshold is chosen, is used as off-network in training data and test data
The user property at family.
Further, a variety of user properties of the off-network user comprise at least one of the following:
B numeric field data that the user of off-network user enters an item of expenditure in the accounts, the domain O network data, user behavior data, customer complaint data.
Further, it is described according to the off-network probability and the off-network may reason determine existing net middle width strip user from
After net tendency, further includes:
The possible reason of the user list, off-network probability and off-network with off-network tendency is exported, according to the user list
Progress user maintains and user service.
The second aspect of the present invention is to provide a kind of pre- measurement equipment of broadband off-network, comprising:
Memory, for storing computer program；
Processor, for running the computer program stored in the memory to realize: obtaining any broadband in existing net
The user characteristics of user；In the broadband off-network prediction model that user characteristics input is obtained in advance, obtain in the existing net
The off-network probability and off-network of broadband user may reason；It is determined in existing net according to the off-network probability and the possible reason of the off-network
The off-network of broadband user is inclined to.
Further, the processor is also configured to
The a variety of user properties for obtaining off-network user, as training data and test data；
According to the training data at least two machine learning algorithms of training, alternative model is obtained respectively；
The predictablity rate of each alternative model is tested according to the test data；
Select the highest alternative model of predictablity rate as optimal alternative model, by neural network algorithm to it is described most
Excellent alternative model optimizes, and obtains the broadband off-network prediction model, and the broadband off-network prediction model is used for according to user
Feature exports off-network probability and influences the different degree ranking of the user characteristics of off-network probability.
The third aspect of the present invention is to provide a kind of computer readable storage medium, is stored thereon with computer program；
Method as described in relation to the first aspect is realized when the computer program is executed by processor.
Broadband off-network prediction technique, equipment and storage medium provided by the invention are used by obtaining any broadband in existing net
The user characteristics at family；And in the broadband off-network prediction model for obtaining user characteristics input in advance, obtain in the existing net
The off-network probability and off-network of broadband user may reason；Then it is determined according to the off-network probability and the possible reason of the off-network existing
The off-network of net middle width strip user is inclined to.The broadband obtained in the present invention by the user characteristics of existing net middle width strip user and in advance
Off-network prediction model, can accurately predict existing net middle width strip user off-network tendency and off-network may reason, and then can be with
It instructs operator that targetedly existing network users are maintained and serviced, realize the precision marketing to user and maintains.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention without any creative labor, may be used also for those of ordinary skill in the art
To obtain other drawings based on these drawings.
Fig. 1 is off-network prediction technique flow chart in broadband provided in an embodiment of the present invention；
Fig. 2 be another embodiment of the present invention provides broadband off-network prediction technique flow chart；
Fig. 3 is the structure chart of off-network prediction meanss in broadband provided in an embodiment of the present invention；
Fig. 4 is the structure chart of off-network pre- measurement equipment in broadband 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, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Fig. 1 is off-network prediction technique flow chart in broadband provided in an embodiment of the present invention.Present embodiments provide a kind of broadband
Off-network prediction technique, specific step is as follows for this method:
S101, the user characteristics for obtaining any broadband user in existing net.
In the present embodiment, the user characteristics for now netting broadband user can the user according to involved in the off-network prediction model of broadband
Feature determines, B numeric field data that wherein user characteristics can enter an item of expenditure in the accounts from the user of existing network users, the domain O network data, user behavior number
According to being obtained in the data such as, customer complaint data.
In S102, the broadband off-network prediction model for obtaining user characteristics input in advance, obtain wide in the existing net
Off-network probability with user and off-network may reasons.
In the present embodiment, broadband off-network prediction model can be obtained in advance, middle width strip off-network prediction model can be from a variety of
Optimal algorithm building model is chosen in machine learning algorithm, and the model is optimized by neural network algorithm, under
State the acquisition process that will be described in detail broadband off-network prediction model in embodiment.Broadband off-network in certain the present embodiment predicts mould
Type may be other models, and details are not described herein again.The user characteristics of existing net middle width strip user are inputted into width in the present embodiment
In band off-network prediction model, the off-network probability and the possible reason of off-network of the broadband user can be obtained, wherein off-network can be active
Because poor such as video-aware, webpage opening is slow, double faults rate is low, complaint fix-rate is low.
S103, determine that the off-network of existing net middle width strip user is inclined to according to the off-network probability and the possible reason of the off-network.
In the present embodiment, it can determine whether existing net middle width strip user has according to off-network probability and the possible reason of off-network
Off-network tendency, such as the threshold value of off-network probability can be set in the present embodiment, it is somebody's turn to do when the off-network probability of existing net middle width strip user is greater than
Threshold value, it is determined that the broadband user is inclined to off-network.
Further, determine existing net middle width strip user's according to the off-network probability and the possible reason of the off-network described
After off-network tendency, it may also include that
The possible reason of the user list, off-network probability and off-network with off-network tendency is exported, according to the user list
Progress user maintains and user service.
More specifically, the broadband user of preset threshold can be greater than in the present embodiment to the off-network probability of existing net middle width strip user
The label of " off-network " is set, and the broadband user of " off-network " label is the broadband user for being predicted to be off-network tendency, further
The statistics available broadband user with " off-network " label, thus the user list for obtaining having off-network to be inclined to, and by corresponding off-network
Probability and off-network may reason export together, in order to operator according to user list, off-network may reason etc. targetedly into
Row is maintained and is serviced.
Off-network prediction technique in broadband provided in this embodiment, by the user characteristics for obtaining any broadband user in existing net；
And in the broadband off-network prediction model for obtaining user characteristics input in advance, the off-network of the existing net middle width strip user is obtained
Probability and off-network may reasons；Then determine existing net middle width strip user's according to the off-network probability and the possible reason of the off-network
Off-network tendency.The broadband off-network prediction mould obtained in the present embodiment by the user characteristics of existing net middle width strip user and in advance
Type, can accurately predict the off-network tendency and the possible reason of off-network of existing net middle width strip user, and then can instruct operator
Targetedly existing network users are maintained and serviced, realize the precision marketing to user and are maintained.
On the basis of the above embodiments, as shown in Fig. 2, the broadband off-network prediction technique of the present embodiment further include:
S201, a variety of user properties for obtaining off-network user, as training data and test data.
In the present embodiment, big data Mining Thought is introduced, the user that data source expansion includes at least off-network user enters an item of expenditure in the accounts
B numeric field data, the domain O network data, user behavior data, one of customer complaint data etc. or a variety of, wherein user enters an item of expenditure in the accounts
B numeric field data may include signing broadband, signing duration, arrearage duration, supplement promptness rate with money etc., the domain O network data may include network
Network speed, time delay, flow etc., user behavior data may include the number etc. that user accesses other operator's portals, customer complaint
Data may include complaining type, the complaint frequency, fix-rate etc., may also include outside ventures, fault message etc..
All data can all be organized into a user-level data summary table, include that attribute field surpasses 200 in table, such as user
The B numeric field datas such as network entry time, user's gender birthday, package name, marketing channel, user's line duration, broadband bearing mode, width
Tape jam number, failure averagely last, whether open the O numeric field datas such as IPTV, IPTV rating duration, the perception of IPTV rating, user
It accesses time delay that other website numbers, the popular application used, popular application use, flow, use the behavioral datas such as duration, use
Family complains number, complaint handling duration etc. to complain data, has been truly realized the big data analysis of interconnection networking.In the present embodiment,
80% is used as training data in the data that can be will acquire, and 20% is used as test data.
S202, at least two machine learning algorithms are trained according to the training data, obtains alternative model respectively.
In the present embodiment, machine learning algorithm may include random forests algorithm, Gradient Boosting algorithm,
XGboosting etc. is trained study to each machine learning algorithm according to training data, each machine learning after the completion of study
Algorithm respectively obtains a kind of alternative model, and each alternative model eachs relate to the user property of a set of off-network user as user spy
Sign.
S203, the predictablity rate that each alternative model is tested according to the test data.
In the present embodiment, the off-network user according to needed for obtaining each alternative model in the user property of test data
User characteristics, and alternative model is verified according to the user characteristics of the off-network user of acquisition, obtain the prediction of alternative model
Accuracy rate.
S204, it selects the highest alternative model of predictablity rate as optimal alternative model, passes through neural network algorithm pair
The optimal alternative model optimizes, and obtains the broadband off-network prediction model, and the broadband off-network prediction model is used for root
Off-network probability is exported according to user characteristics and influences the different degree ranking of the user characteristics of off-network probability.
In the present embodiment, optimal alternative model is selected according to predictablity rate, then uses artificial neural network algorithm
Optimal alternative model is further supplemented and perfect, so that broadband off-network prediction model, which can export, influences broadband off-network
The different degree ranking of the user characteristics of the off-network probability of prediction model prediction, can determine whether the off-network of broadband user according to off-network probability
The off-network of tendency, the user that according to the different degree ranking of user characteristics can determine whether that there is off-network to be inclined to may reason.
In the present embodiment, the method for artificial neural network built using independent design, the Design on Artificial Neural Networks is hidden
Hiding layer is 38 layers, and activation primitive selects RELU and Sigmoid cross-reference.The design obtained by multiple repetition test, 38 layers
Hidden layer and RELU, the cross-reference of Sigmoid can make the model of wideband data obtain higher predictablity rate.
On the basis of the above embodiments, after a variety of user properties that off-network user is obtained described in S201, may also include that
There are the data of absent field for acquisition, and fill up to absent field.
In the present embodiment, in due to existing communications industry data, many data be it is incomplete, there are data
The main problems such as missing and data field are more.For shortage of data, can be counted especially by following process in the present embodiment
According to filling up for absent field:
Judge whether the absent field belongs to Gaussian Profile；
If the absent field belongs to Gaussian Profile, the missing word is filled up using gauss of distribution function random value
Section；
If the absent field is not belonging to Gaussian Profile, gauss of distribution function is constructed using method for normalizing, is recycled
Gauss of distribution function random value is to fill up the absent field.
The complementing method of shortage of data field in the present embodiment mean value enthesis more usually used than in industry is more sticked on
The nearly field regularity of distribution, catches the essence of broadband user's data, is conducive to the accuracy for improving model.
Further, after a variety of user properties that off-network user is obtained described in S201, may also include that
Obtain the variance expansion factor in a variety of user properties of off-network user between any two users' attribute；
The user property that variance expansion factor is less than preset threshold is chosen, is used as off-network in training data and test data
The user property at family.
In the present embodiment, by variance expansion factor (variance inflation factor, abbreviation VIF) to mould
User property needed for type is screened, can preferably Controlling model quality.Wherein variance expansion factor is characterization independent variable
The numerical value of multi-collinearity degree between observed value, size can reflect out between the observed value of independent variable with the presence or absence of multiple conllinear
How are property and its degree, and variance expansion factor is bigger, and multi-collinearity is more serious.The field of VIF < 1.5 is chosen in the present embodiment
Into model, can better Controlling model quality.
Off-network prediction technique in broadband provided in this embodiment is obtained by the user characteristics of existing net middle width strip user and in advance
The broadband off-network prediction model taken can accurately predict the off-network tendency and the possible reason of off-network of existing net middle width strip user,
And then can instruct operator that targetedly existing network users are maintained and serviced, realize precision marketing and dimension to user
System.Its middle width strip off-network prediction model is with the big data analysis Mining Thought for interconnecting networking, using a variety of machine learning algorithms
Study analysis is carried out to the behavioural characteristic of off-network user, cleaning is carried out to the data of magnanimity and is modeled with machine learning, influence is found out
Relationship between the behavior of user's off-network and user itself other characteristics of variables, the accurate off-network tendency for predicting broadband user can be into
The realization of one step is targetedly maintained and is serviced to existing network users.
Fig. 3 is the structure chart of off-network prediction meanss in broadband provided in an embodiment of the present invention.Broadband provided in this embodiment from
Net prediction meanss can execute the process flow of broadband off-network prediction technique embodiment offer, as shown in figure 3, the broadband off-network
Prediction meanss 30 include obtaining module 31 and prediction module 32.
Module 31 is obtained, for obtaining the user characteristics of any broadband user in existing net；
Prediction module 32 obtains institute for inputting the user characteristics in the broadband off-network prediction model obtained in advance
The off-network probability and off-network for stating existing net middle width strip user may reasons；It is true according to the off-network probability and the possible reason of the off-network
Surely show the off-network tendency of net middle width strip user.
Further, described device further include: model training module 33 is used for:
The a variety of user properties for obtaining off-network user, as training data and test data；
According to the training data at least two machine learning algorithms of training, alternative model is obtained respectively；
The predictablity rate of each alternative model is tested according to the test data；
Select the highest alternative model of predictablity rate as optimal alternative model, by neural network algorithm to it is described most
Excellent alternative model optimizes, and obtains the broadband off-network prediction model, and the broadband off-network prediction model is used for according to user
Feature exports off-network probability and influences the different degree ranking of the user characteristics of off-network probability.
Further, the model training module 33 is also used to:
After a variety of user properties for obtaining off-network user, there are the data of absent field for acquisition, and carry out to absent field
It fills up.
Further, the model training module 33 is specifically used for:
Judge whether the absent field belongs to Gaussian Profile；
If the absent field belongs to Gaussian Profile, the missing word is filled up using gauss of distribution function random value
Section；
If the absent field is not belonging to Gaussian Profile, gauss of distribution function is constructed using normalized device, is recycled
Gauss of distribution function random value is to fill up the absent field.
Further, the model training module 33 is also used to:
After a variety of user properties for obtaining off-network user, any two users in a variety of user properties of off-network user are obtained
Variance expansion factor between attribute；
The user property that variance expansion factor is less than preset threshold is chosen, is used as off-network in training data and test data
The user property at family.
Further, a variety of user properties of the off-network user comprise at least one of the following:
B numeric field data that the user of off-network user enters an item of expenditure in the accounts, the domain O network data, user behavior data, customer complaint data.
Further, the prediction module 32 is also used to:
It is defeated after determining the off-network tendency of existing net middle width strip user according to the off-network probability and the possible reason of the off-network
The possible reason of user list, off-network probability and off-network of off-network tendency is provided, to carry out user's dimension according to the user list
System and user service.
Off-network prediction meanss in broadband provided in an embodiment of the present invention can be specifically used for the above-mentioned Fig. 1 and Fig. 2 of execution and be provided
Embodiment of the method, details are not described herein again for concrete function.
Off-network prediction meanss in broadband provided in an embodiment of the present invention, it is special by the user for obtaining any broadband user in existing net
Sign；And in the broadband off-network prediction model for obtaining user characteristics input in advance, obtain the existing net middle width strip user's
Off-network probability and off-network may reasons；It then may the determining existing net middle width strip use of reason according to the off-network probability and the off-network
The off-network at family is inclined to.The broadband off-network prediction obtained in the present embodiment by the user characteristics of existing net middle width strip user and in advance
Model, can accurately predict the off-network tendency and the possible reason of off-network of existing net middle width strip user, and then can instruct to run
Quotient targetedly maintains and services to existing network users, realizes the precision marketing to user and maintains.
Fig. 4 is the structural schematic diagram of off-network pre- measurement equipment in broadband provided in an embodiment of the present invention.The embodiment of the present invention provides
The pre- measurement equipment of broadband off-network can execute broadband off-network prediction technique embodiment offer process flow, as shown in figure 4, broadband
The pre- measurement equipment 40 of off-network includes memory 41, processor 42, computer program and communication interface 43；Wherein, computer program is deposited
Storage is configured as executing broadband off-network prediction technique described in above embodiments as processor 42 in memory 41.
Specifically, the processor 42 runs the computer program stored in the memory 41 to realize: obtaining existing net
In any broadband user user characteristics；In the broadband off-network prediction model that user characteristics input is obtained in advance, obtain
The off-network probability and off-network of the existing net middle width strip user may reason；It may reason according to the off-network probability and the off-network
Determine the off-network tendency of existing net middle width strip user.
Further, the processor 42 is also configured to
The a variety of user properties for obtaining off-network user, as training data and test data；
According to the training data at least two machine learning algorithms of training, alternative model is obtained respectively；
The predictablity rate of each alternative model is tested according to the test data；
Select the highest alternative model of predictablity rate as optimal alternative model, by neural network algorithm to it is described most
Excellent alternative model optimizes, and obtains the broadband off-network prediction model, and the broadband off-network prediction model is used for according to user
Feature exports off-network probability and influences the different degree ranking of the user characteristics of off-network probability.
The pre- measurement equipment of broadband off-network of embodiment illustrated in fig. 4 can be used for executing the technical solution of above method embodiment,
The realization principle and technical effect are similar, and details are not described herein again.
In addition, the present embodiment also provides a kind of computer readable storage medium, it is stored thereon with computer program, the meter
Calculation machine program is executed by processor to realize off-network prediction technique in broadband described in above-described embodiment.
In several embodiments provided by the present invention, it should be understood that disclosed device and method can pass through it
Its mode is realized.For example, the apparatus embodiments described above are merely exemplary, for example, the division of the unit, only
Only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components can be tied
Another system is closed or is desirably integrated into, or some features can be ignored or not executed.Another point, it is shown or discussed
Mutual coupling, direct-coupling or communication connection can be through some interfaces, the INDIRECT COUPLING or logical of device or unit
Letter connection can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of hardware adds SFU software functional unit.
The above-mentioned integrated unit being realized in the form of SFU software functional unit can store and computer-readable deposit at one
In storage media.Above-mentioned SFU software functional unit is stored in a storage medium, including some instructions are used so that a computer
It is each that equipment (can be personal computer, server or the network equipment etc.) or processor (processor) execute the present invention
The part steps of embodiment the method.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (Read-
Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic or disk etc. it is various
It can store the medium of program code.
Those skilled in the art can be understood that, for convenience and simplicity of description, only with above-mentioned each functional module
Division progress for example, in practical application, can according to need and above-mentioned function distribution is complete by different functional modules
At the internal structure of device being divided into different functional modules, to complete all or part of the functions described above.On
The specific work process for stating the device of description, can refer to corresponding processes in the foregoing method embodiment, and details are not described herein.
Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention., rather than its limitations；To the greatest extent
Pipe present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: its according to
So be possible to modify the technical solutions described in the foregoing embodiments, or to some or all of the technical features into
Row equivalent replacement；And these are modified or replaceed, various embodiments of the present invention technology that it does not separate the essence of the corresponding technical solution
The range of scheme.
Claims (10)
1. a kind of broadband off-network prediction technique characterized by comprising
Obtain the user characteristics of any broadband user in existing net；
By in the broadband off-network prediction model that obtains in advance of user characteristics input, obtain the existing net middle width strip user from
Net probability and off-network may reasons；
The off-network tendency of existing net middle width strip user is determined according to the off-network probability and the possible reason of the off-network.
2. the method according to claim 1, wherein the method also includes:
The a variety of user properties for obtaining off-network user, as training data and test data；
According to the training data at least two machine learning algorithms of training, alternative model is obtained respectively；
The predictablity rate of each alternative model is tested according to the test data；
Select the highest alternative model of predictablity rate as optimal alternative model, by neural network algorithm to described optimal standby
Modeling type optimizes, and obtains the broadband off-network prediction model, and the broadband off-network prediction model is used for according to user characteristics
It exports off-network probability and influences the different degree ranking of the user characteristics of off-network probability.
3. according to the method described in claim 2, it is characterized in that, being gone back after a variety of user properties for obtaining off-network user
Include:
There are the data of absent field for acquisition, and fill up to absent field.
4. according to the method described in claim 3, it is characterized in that, described fill up absent field, comprising:
Judge whether the absent field belongs to Gaussian Profile；
If the absent field belongs to Gaussian Profile, the absent field is filled up using gauss of distribution function random value；
If the absent field is not belonging to Gaussian Profile, gauss of distribution function is constructed using method for normalizing, recycles Gauss
Distribution function random value is to fill up the absent field.
5. according to the method described in claim 2, it is characterized in that, being gone back after a variety of user properties for obtaining off-network user
Include:
Obtain the variance expansion factor in a variety of user properties of off-network user between any two users' attribute；
The user property that variance expansion factor is less than preset threshold is chosen, as off-network user in training data and test data
User property.
6. according to the described in any item methods of claim 2-5, which is characterized in that
A variety of user properties of the off-network user comprise at least one of the following:
B numeric field data that the user of off-network user enters an item of expenditure in the accounts, the domain O network data, user behavior data, customer complaint data.
7. the method according to claim 1, wherein described can be active according to the off-network probability and the off-network
After off-network tendency because determining existing net middle width strip user, further includes:
The possible reason of the user list, off-network probability and off-network with off-network tendency is exported, to carry out according to the user list
User maintains and user service.
8. a kind of pre- measurement equipment of broadband off-network characterized by comprising
Memory, for storing computer program；
Processor, for running the computer program stored in the memory to realize: obtaining any broadband user in existing net
User characteristics；In the broadband off-network prediction model that user characteristics input is obtained in advance, the existing net middle width strip is obtained
The off-network probability and off-network of user may reason；It may the determining existing net middle width strip of reason according to the off-network probability and the off-network
The off-network of user is inclined to.
9. equipment according to claim 8, which is characterized in that the processor is also configured to
The a variety of user properties for obtaining off-network user, as training data and test data；
According to the training data at least two machine learning algorithms of training, alternative model is obtained respectively；
The predictablity rate of each alternative model is tested according to the test data；
Select the highest alternative model of predictablity rate as optimal alternative model, by neural network algorithm to described optimal standby
Modeling type optimizes, and obtains the broadband off-network prediction model, and the broadband off-network prediction model is used for according to user characteristics
It exports off-network probability and influences the different degree ranking of the user characteristics of off-network probability.
10. a kind of computer readable storage medium, which is characterized in that be stored thereon with computer program；
Such as method of any of claims 1-6 is realized when the computer program is executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811615555.1A CN109741098A (en) | 2018-12-27 | 2018-12-27 | Broadband off-network prediction technique, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811615555.1A CN109741098A (en) | 2018-12-27 | 2018-12-27 | Broadband off-network prediction technique, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109741098A true CN109741098A (en) | 2019-05-10 |
Family
ID=66361461
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811615555.1A Pending CN109741098A (en) | 2018-12-27 | 2018-12-27 | Broadband off-network prediction technique, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109741098A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110971460A (en) * | 2019-12-03 | 2020-04-07 | 北京红山信息科技研究院有限公司 | Off-network prediction method, device, server and storage medium |
| CN111740866A (en) * | 2020-06-24 | 2020-10-02 | 中国联合网络通信集团有限公司 | Off-grid prediction method and device |
| CN112671573A (en) * | 2020-12-17 | 2021-04-16 | 北京神州泰岳软件股份有限公司 | Method and device for identifying potential off-network users in broadband service |
| CN113543117A (en) * | 2020-04-22 | 2021-10-22 | 中国移动通信集团重庆有限公司 | Prediction method and device for number portability user and computing equipment |
| CN113543178A (en) * | 2021-07-28 | 2021-10-22 | 北京红山信息科技研究院有限公司 | Service optimization method, device, equipment and storage medium based on user perception |
| CN113727332A (en) * | 2021-08-12 | 2021-11-30 | 中国联合网络通信集团有限公司 | Method and device for recovering internet surfing |
| CN114389962A (en) * | 2021-12-27 | 2022-04-22 | 中国电信股份有限公司 | Broadband loss user determination method and device, electronic equipment and storage medium |
| CN114786173A (en) * | 2022-03-29 | 2022-07-22 | 中国联合网络通信集团有限公司 | Number portability identification method, device, equipment and storage medium based on broadband |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20120139637A (en) * | 2009-07-09 | 2012-12-27 | 액센츄어 글로벌 서비시즈 리미티드 | Marketing model determination system |
| CN106022505A (en) * | 2016-04-28 | 2016-10-12 | 华为技术有限公司 | Method and device of predicting user off-grid |
| CN108629506A (en) * | 2018-05-03 | 2018-10-09 | 深圳市牛鼎丰科技有限公司 | Modeling method, device, computer equipment and the storage medium of air control model |
| CN109086931A (en) * | 2018-08-01 | 2018-12-25 | 中国联合网络通信集团有限公司 | Predict user's off-network method and system |
-
2018
- 2018-12-27 CN CN201811615555.1A patent/CN109741098A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20120139637A (en) * | 2009-07-09 | 2012-12-27 | 액센츄어 글로벌 서비시즈 리미티드 | Marketing model determination system |
| CN106022505A (en) * | 2016-04-28 | 2016-10-12 | 华为技术有限公司 | Method and device of predicting user off-grid |
| CN108629506A (en) * | 2018-05-03 | 2018-10-09 | 深圳市牛鼎丰科技有限公司 | Modeling method, device, computer equipment and the storage medium of air control model |
| CN109086931A (en) * | 2018-08-01 | 2018-12-25 | 中国联合网络通信集团有限公司 | Predict user's off-network method and system |
Non-Patent Citations (1)
| Title |
|---|
| 夏立玲等: "水环境数据缺失处理方法之优化研究", 《水电能源科学》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110971460A (en) * | 2019-12-03 | 2020-04-07 | 北京红山信息科技研究院有限公司 | Off-network prediction method, device, server and storage medium |
| CN113543117B (en) * | 2020-04-22 | 2022-10-04 | 中国移动通信集团重庆有限公司 | Prediction method and device for number portability user and computing equipment |
| CN113543117A (en) * | 2020-04-22 | 2021-10-22 | 中国移动通信集团重庆有限公司 | Prediction method and device for number portability user and computing equipment |
| CN111740866A (en) * | 2020-06-24 | 2020-10-02 | 中国联合网络通信集团有限公司 | Off-grid prediction method and device |
| CN112671573A (en) * | 2020-12-17 | 2021-04-16 | 北京神州泰岳软件股份有限公司 | Method and device for identifying potential off-network users in broadband service |
| CN112671573B (en) * | 2020-12-17 | 2023-05-16 | 北京神州泰岳软件股份有限公司 | Method and device for identifying potential off-network users in broadband service |
| CN113543178A (en) * | 2021-07-28 | 2021-10-22 | 北京红山信息科技研究院有限公司 | Service optimization method, device, equipment and storage medium based on user perception |
| CN113543178B (en) * | 2021-07-28 | 2024-04-09 | 北京红山信息科技研究院有限公司 | Service optimization method, device, equipment and storage medium based on user perception |
| CN113727332A (en) * | 2021-08-12 | 2021-11-30 | 中国联合网络通信集团有限公司 | Method and device for recovering internet surfing |
| CN113727332B (en) * | 2021-08-12 | 2022-09-02 | 中国联合网络通信集团有限公司 | Method and device for recovering internet surfing |
| CN114389962A (en) * | 2021-12-27 | 2022-04-22 | 中国电信股份有限公司 | Broadband loss user determination method and device, electronic equipment and storage medium |
| CN114786173A (en) * | 2022-03-29 | 2022-07-22 | 中国联合网络通信集团有限公司 | Number portability identification method, device, equipment and storage medium based on broadband |
| CN114786173B (en) * | 2022-03-29 | 2023-06-09 | 中国联合网络通信集团有限公司 | Broadband-based number-carrying network-switching identification method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109741098A (en) | Broadband off-network prediction technique, equipment and storage medium | |
| CN110033314B (en) | Advertisement data processing method and device | |
| CN108337316A (en) | Information-pushing method, device, computer equipment and storage medium | |
| US20190272553A1 (en) | Predictive Modeling with Entity Representations Computed from Neural Network Models Simultaneously Trained on Multiple Tasks | |
| CN104992347B (en) | A kind of method and device of video matching advertisement | |
| CN111784508A (en) | Enterprise risk assessment method and device and electronic equipment | |
| US11514368B2 (en) | Methods, apparatuses, and computing devices for trainings of learning models | |
| CN107689008A (en) | A kind of user insures the method and device of behavior prediction | |
| CN107330445A (en) | The Forecasting Methodology and device of user property | |
| CN104679743A (en) | Method and device for determining preference model of user | |
| CN105096152A (en) | Commodity popularity-based operation execution method and device | |
| CN103778474A (en) | Resource load capacity prediction method, analysis prediction system and service operation monitoring system | |
| CN110147803A (en) | Customer churn early-warning processing method and device | |
| CN107220867A (en) | object control method and device | |
| CN104574124A (en) | Method and device for determining exhibition effect of advertisement data | |
| CN110443265A (en) | A kind of behavioral value method and apparatus based on corporations | |
| CN107808220A (en) | A kind of revenue prediction method and device | |
| CN103186604A (en) | Method, device and equipment for determining satisfaction degree of user on search result | |
| CN109428760B (en) | User credit evaluation method based on operator data | |
| CN108228950A (en) | A kind of information processing method and device | |
| CN112765452A (en) | Search recommendation method and device and electronic equipment | |
| CN106407316A (en) | Topic model-based software question and answer recommendation method and device | |
| CN112581185A (en) | Method, system, computer and storage medium for estimating gender and age of advertisement user | |
| CN114757637A (en) | Credit approval method, credit approval device, credit approval equipment and credit approval storage medium based on decision engine | |
| CN109359850A (en) | A kind of method and device generating risk assessment scale |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190510 |
|
| RJ01 | Rejection of invention patent application after publication |