CN104866626A - Method and device for recommending telecommunication service - Google Patents
Method and device for recommending telecommunication service Download PDFInfo
- Publication number
- CN104866626A CN104866626A CN201510329668.5A CN201510329668A CN104866626A CN 104866626 A CN104866626 A CN 104866626A CN 201510329668 A CN201510329668 A CN 201510329668A CN 104866626 A CN104866626 A CN 104866626A
- Authority
- CN
- China
- Prior art keywords
- telecommunication service
- testing accuracy
- individuality
- population
- recommended
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
Abstract
The embodiment of the invention provides a method and device for recommending a telecommunication service. The method comprises the following steps: performing telecommunication service matching prediction on target users according to the history information of the target users and a training result which is obtained in advance and is used for checking whether or not the users would like to accept a telecommunication service to be recommended; recognizing users who would like to accept the telecommunication service to be recommended according to the result of the telecommunication service matching prediction; and transmitting telecommunication service recommended information to the users who would like to accept the telecommunication service to be recommended. According to the embodiment of the invention, the telecommunication service can be recommended to the users accurately.
Description
Technical field
The present invention relates to communication technical field, the recommend method of particularly a kind of telecommunication service and device.
Background technology
Present stage, along with the significantly reduction of the universal of data traffic consumption with charging level, the traditional telecommunication service of three large telecom operators is faced with the destiny of homogeneity competition, telecom operators must be made the transition to intensive management and lean operation, from dependence traditional media in the past, as TV, broadcast, newspaper, the popular marketing model that outdoor advertising etc. " blanket type " are bombed, to the demand analysed in depth with hold user, accurately orient the type of service of user's adaptation, thus to the accurate marketing Mode change of " for client finds the business of applicable himself ".
At present, each basic telecom carrier all establishes respective operation analysis system, and the following several mode of main dependence is assisted and carried out accurate marketing excavation:
Mode one: business association rule and method that is traditional and that improve, forms multiple business association item assist and carries out marketing excavation;
Mode two: based on the basic document of this telecommunication user, brand classification (Global Link, walk in the Divine Land, M-ZONE), the history moon Back ground Information such as spending limit carry out simple analysis and data mining.
For mode one, owing to only considered the relevance between business in which, not from the angle of client, consider the true consumption needs of client, be in like fashion reflection be that service developer thinks recommendable business, obviously so business recommended result draws from service developer angle, is not very accurately, really cannot meet the business demand of client.
For mode two, because the real demand of user cannot obtain from the basic document of self, client's set meal classification, history moon spending limit, such as, the business of the client of monthly high consumption not all to expensive is interested in history, and low consumption client is not interested in cheap business entirely yet equally in history.So obviously so business recommended result is also inaccurate, also cannot really meet client's real demand.
Summary of the invention
The object of the embodiment of the present invention is the recommend method and the device that provide a kind of telecommunication service, accurately can recommend telecommunication service to user.
In order to achieve the above object, The embodiment provides a kind of recommend method of telecommunication service, the method comprises:
According to the historical information of targeted customer and the training result that whether can accept telecommunication service to be recommended for inspection user that obtains in advance, the prediction of telecommunication service coupling is carried out to targeted customer, obtains the result of telecommunication service coupling prediction;
According to the result of telecommunication service coupling prediction, identify the user that can accept telecommunication service to be recommended;
Telecommunication service recommendation information is sent to the user that can accept telecommunication service to be recommended.
Wherein, according to the result of telecommunication service coupling prediction, identify the user that can accept telecommunication service to be recommended, specifically comprise:
Judge whether the result of telecommunication service coupling prediction is greater than preset value;
If the result of telecommunication service coupling prediction is greater than preset value, then determine that this targeted customer is the user that can accept telecommunication service to be recommended.
Wherein, according to the historical information of targeted customer and the training result that whether can accept telecommunication service to be recommended for inspection user that obtains in advance, carry out the prediction of telecommunication service coupling to targeted customer, before obtaining the result of telecommunication service coupling prediction, method also comprises:
Obtain and be used for the training result whether inspection user can accept telecommunication service to be recommended.
Wherein, obtain and be used for the training result whether inspection user can accept telecommunication service to be recommended, specifically comprise:
Obtain the basic data required for training;
Pre-service is carried out to basic data, obtains intermediate data;
By genetic algorithm and neural network algorithm, intermediate data is trained, obtain training result.
Wherein, pre-service is carried out to basic data, obtains intermediate data, specifically comprise:
Require to screen basic data according to the screening preset, obtain the data after screening;
According to the data layout preset, the data after screening are sorted;
Data after sequence are normalized, obtain the intermediate data with the data structure that genetic algorithm can identify.
Wherein, by genetic algorithm and neural network algorithm, intermediate data is trained, obtains training result, specifically comprise:
By genetic algorithm, interative computation is carried out to intermediate data, obtain initial population;
Initial population is decoded, obtains each individuality in initial population;
Be optimized by the testing accuracy of neural network algorithm to each individuality in initial population, obtain transitional population;
Judge whether to exist in transitional population the individuality that testing accuracy is greater than default testing accuracy;
If there is the individuality that testing accuracy is greater than default testing accuracy in transitional population, then using this individuality as training result.
Wherein, method also comprises:
If there is not the individuality that testing accuracy is greater than default testing accuracy in transitional population, then according to preset testing accuracy, by genetic algorithm, transitional population is selected, crossover and mutation operation, obtain new colony;
New colony is decoded, and is optimized by the testing accuracy of neural network algorithm to decoded each individuality, obtain new transitional population;
Judge whether to exist in new transitional population the individuality that testing accuracy is greater than default testing accuracy;
If there is the individuality that testing accuracy is greater than default testing accuracy in new transitional population, then using this individuality as training result;
If there is not the individuality that testing accuracy is greater than default testing accuracy in new transitional population, then continue through that genetic algorithm is selected new transitional population, crossover and mutation operation, obtain new colony, and continue through the testing accuracy of neural network algorithm to each individuality in new colony and be optimized, obtain new transitional population, until there is the individuality that testing accuracy is greater than default testing accuracy in new transitional population, and using this individuality as training result.
Wherein, be optimized continuing through the testing accuracy of neural network algorithm to each individuality in new colony, after obtaining new transitional population, method also comprises:
Judge whether the interative computation number of times of genetic algorithm reaches preset times;
If the interative computation number of times of genetic algorithm reaches preset times, then deconditioning, and the highest individuality of testing accuracy in the new transitional population last computing obtained is as training result.
Embodiments of the invention additionally provide a kind of recommendation apparatus of telecommunication service, and this device comprises:
Prediction module, for according to the historical information of targeted customer and the training result that whether can accept telecommunication service to be recommended for inspection user that obtains in advance, carries out the prediction of telecommunication service coupling to targeted customer, obtains the result of telecommunication service coupling prediction;
Identification module, for the result according to the prediction of telecommunication service coupling, identifies the user that can accept telecommunication service to be recommended;
Sending module, for sending telecommunication service recommendation information to the user that can accept telecommunication service to be recommended.
Wherein, identification module comprises:
Judging unit, for judging whether the result of telecommunication service coupling prediction is greater than preset value, and when the result of telecommunication service coupling prediction is greater than preset value, triggers determining unit;
Determining unit, for the triggering according to judging unit, determines that this targeted customer is the user that can accept telecommunication service to be recommended.
Wherein, device also comprises:
Acquisition module, for obtaining the training result that whether can accept telecommunication service to be recommended for inspection user.
Wherein, acquisition module comprises:
Acquiring unit, for obtaining the basic data of training;
Processing unit, for carrying out pre-service to basic data, obtains intermediate data;
Training unit, for being trained intermediate data by genetic algorithm and neural network algorithm, obtains training result.
Wherein, processing unit comprises:
Screening subelement, for requiring to screen basic data according to the screening preset, obtains the data after screening;
Sequence subelement, for according to the data layout preset, sorts to the data after screening;
Normalizing unit, for being normalized the data after sequence, obtains the intermediate data with the data structure that genetic algorithm can identify.
Wherein, training unit comprises:
Operator unit, for carrying out interative computation by genetic algorithm to intermediate data, obtains initial population;
Decoding subelement, for decoding to initial population, obtains each individuality in initial population;
Optimizing subelement, for being optimized by the testing accuracy of neural network algorithm to each individuality in initial population, obtaining transitional population;
Judgment sub-unit, for judging whether to exist in transitional population the individuality that testing accuracy is greater than default testing accuracy, and when there is testing accuracy in transitional population and being greater than default testing accuracy individual, triggers first and determines subelement;
First determines subelement, for the triggering according to judgment sub-unit, using this individuality as training result.
Wherein, device also comprises:
First operational module, for when there is not testing accuracy in transitional population and being greater than default testing accuracy individual, according to default testing accuracy, is selected transitional population by genetic algorithm, crossover and mutation operation, obtains new colony;
Decoder module, for decoding to new colony, and being optimized by the testing accuracy of neural network algorithm to decoded each individuality, obtaining new transitional population;
First judge module, for judging whether to exist in new transitional population the individuality that testing accuracy is greater than default testing accuracy, and in new transitional population, there is testing accuracy when being greater than default testing accuracy individual, trigger determination module, in new transitional population, there is not testing accuracy when being greater than default testing accuracy individual, trigger the second operational module;
Determination module, for the triggering according to the first judge module, using this individuality as training result;
Second operational module, for the triggering according to the first judge module, continue through that genetic algorithm is selected new transitional population, crossover and mutation operation, obtain new colony, and continue through the testing accuracy of neural network algorithm to each individuality in new colony and be optimized, obtain new transitional population, until there is the individuality that testing accuracy is greater than default testing accuracy in new transitional population, and using this individuality as training result.
Wherein, device also comprises:
Second judge module, for judging whether the interative computation number of times of genetic algorithm reaches preset times, and when the interative computation number of times of genetic algorithm reaches preset times, triggers stopping modular;
Stopping modular, for the triggering according to the second judge module, deconditioning, and the highest individuality of testing accuracy in the new transitional population last computing obtained is as training result.
Such scheme of the present invention at least comprises following beneficial effect:
In an embodiment of the present invention, by the historical information of targeted customer and the training result that whether can accept telecommunication service to be recommended for inspection user that obtains in advance, the user that can accept telecommunication service to be recommended is identified from targeted customer, and send telecommunication service recommendation information to the user that can accept telecommunication service to be recommended, solve the problem can not recommending telecommunication service exactly to user, reach the effect of accurately recommending telecommunication service to user.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of the recommend method of telecommunication service in first embodiment of the invention;
Fig. 2 is the process flow diagram of the recommend method of telecommunication service in second embodiment of the invention;
Fig. 3 is the process flow diagram of the recommend method of telecommunication service in third embodiment of the invention;
Fig. 4 is the process flow diagram obtaining training result in third embodiment of the invention;
Fig. 5 is the structural representation of the recommendation apparatus of telecommunication service in fourth embodiment of the invention.
Embodiment
Below with reference to accompanying drawings exemplary embodiment of the present disclosure is described in more detail.Although show exemplary embodiment of the present disclosure in accompanying drawing, however should be appreciated that can realize the disclosure in a variety of manners and not should limit by the embodiment set forth here.On the contrary, provide these embodiments to be in order to more thoroughly the disclosure can be understood, and complete for the scope of the present disclosure can be conveyed to those skilled in the art.
First embodiment
As shown in Figure 1, the first embodiment of the present invention provides a kind of recommend method of telecommunication service, and the method comprises:
Step S11, according to the historical information of targeted customer and the training result that whether can accept telecommunication service to be recommended for inspection user that obtains in advance, carries out the prediction of telecommunication service coupling to targeted customer, obtains the result of telecommunication service coupling prediction;
In the first embodiment of the present invention, the historical information of above-mentioned targeted customer mainly comprises customer flow information data and user account Back ground Information, wherein customer flow information data is mainly through gateway packet radio service technology (GPRS, General Packet Radio Service) support node collects, and user account Back ground Information collects mainly through business operation support system (BOSS).
Step S12, according to the result of telecommunication service coupling prediction, identifies the user that can accept telecommunication service to be recommended;
Step S13, sends telecommunication service recommendation information to the user that can accept telecommunication service to be recommended.
In the first embodiment of the present invention, telecommunication service recommendation information can be sent by the mode of note to the user that can accept telecommunication service to be recommended.Be understandable that, do not limit the concrete mode sending telecommunication service recommendation information to the user that can accept telecommunication service to be recommended in the first embodiment of the present invention.
In the first embodiment of the present invention, the training result being used for inspection user and whether can accepting telecommunication service to be recommended can be interpreted as a kind of function, therefore, in time the historical information of targeted customer being imported wherein, the result of telecommunication service coupling prediction can be obtained.And then judge whether this targeted customer is the user that can accept telecommunication service to be recommended according to the result of the telecommunication service coupling prediction obtained, when it is the user that can accept telecommunication service to be recommended, send telecommunication service recommendation information to it.So just, achieve the object of accurately recommending telecommunication service to user.
Second embodiment
As shown in Figure 2, the second embodiment of the present invention provides a kind of recommend method of telecommunication service, and the method comprises:
Step S21, according to the historical information of targeted customer and the training result that whether can accept telecommunication service to be recommended for inspection user that obtains in advance, carries out the prediction of telecommunication service coupling to targeted customer, obtains the result of telecommunication service coupling prediction;
In the second embodiment of the present invention, the historical information of above-mentioned targeted customer mainly comprises customer flow information data and user account Back ground Information, wherein customer flow information data is mainly through gateway packet radio service technology (GPRS, General Packet Radio Service) support node collects, and user account Back ground Information collects mainly through business operation support system (BOSS).
Step S22, judges whether the result of telecommunication service coupling prediction is greater than preset value;
Step S23, if the result of telecommunication service coupling prediction is greater than preset value, then determines that this targeted customer is the user that can accept telecommunication service to be recommended;
Step S24, sends telecommunication service recommendation information to the user that can accept telecommunication service to be recommended.
In the second embodiment of the present invention, telecommunication service recommendation information can be sent by the mode of note to the user that can accept telecommunication service to be recommended.Be understandable that, in the second embodiment of the present invention, do not limit the concrete mode sending telecommunication service recommendation information to the user that can accept telecommunication service to be recommended.
In the second embodiment of the present invention, the training result being used for inspection user and whether can accepting telecommunication service to be recommended can be interpreted as a kind of function, therefore, in time the historical information of targeted customer being imported wherein, the result of telecommunication service coupling prediction can be obtained.And then judge whether the result of this telecommunication service coupling prediction is greater than preset value (such as 0.8), when it is greater than preset value, thinks that this targeted customer is the user that can accept telecommunication service to be recommended, and sends telecommunication service recommendation information to it.If the result of telecommunication service coupling prediction is certainly less than preset value, then think that this targeted customer is the user that can not accept telecommunication service to be recommended, do not need to send telecommunication service recommendation information to it, thus realize accurate business marketing pattern, reach the object of accurately recommending telecommunication service to user.
3rd embodiment
As shown in Figure 3, the third embodiment of the present invention provides a kind of recommend method of telecommunication service, and the method comprises:
Step S31, obtains and is used for the training result whether inspection user can accept telecommunication service to be recommended;
Step S32, according to the historical information of targeted customer and the training result that whether can accept telecommunication service to be recommended for inspection user that obtains in advance, carries out the prediction of telecommunication service coupling to targeted customer, obtains the result of telecommunication service coupling prediction;
In the third embodiment of the present invention, the historical information of above-mentioned targeted customer mainly comprises customer flow information data and user account Back ground Information, wherein customer flow information data is mainly through gateway packet radio service technology (GPRS, General Packet Radio Service) support node collects, and user account Back ground Information collects mainly through business operation support system (BOSS).
Step S33, according to the result of telecommunication service coupling prediction, identifies the user that can accept telecommunication service to be recommended;
Step S34, sends telecommunication service recommendation information to the user that can accept telecommunication service to be recommended.
In the third embodiment of the present invention, telecommunication service recommendation information can be sent by the mode of note to the user that can accept telecommunication service to be recommended.Be understandable that, in the third embodiment of the present invention, do not limit the concrete mode sending telecommunication service recommendation information to the user that can accept telecommunication service to be recommended.
In the third embodiment of the present invention, the training result being used for inspection user and whether can accepting telecommunication service to be recommended can be interpreted as a kind of function, therefore, in time the historical information of targeted customer being imported wherein, the result of telecommunication service coupling prediction can be obtained.And then judge whether this targeted customer is the user that can accept telecommunication service to be recommended according to the result of the telecommunication service coupling prediction obtained, when it is the user that can accept telecommunication service to be recommended, send telecommunication service recommendation information to it.So just, accurate business marketing pattern can be realized, reach the object of accurately recommending telecommunication service to user.
Wherein, in the third embodiment of the present invention, above-mentioned steps S31 specifically comprises: first obtain the basic data required for training; Then pre-service is carried out to basic data, obtain intermediate data; Finally by genetic algorithm and neural network algorithm, intermediate data is trained, obtain training result.
In the third embodiment of the present invention, basic data required for training mainly comprises: a part orders the user of telecommunication service to be recommended and a part of customer flow information data and user account Back ground Information not ordering the user of telecommunication service to be recommended, wherein customer flow information data is mainly through gateway packet radio service technology (GPRS, General Packet Radio Service) support node collects, and user account Back ground Information collects mainly through business operation support system (BOSS).Correspondingly, after getting the basic data required for training, pre-service can be carried out to it, obtain intermediate data, finally by genetic algorithm and neural network algorithm, intermediate data is trained, obtain training result, can elaborate later as concrete training process.
In the third embodiment of the present invention, for 10 yuan of flows upshift set meal (hereinafter referred to as A set meal), elaborate the recommend method of above-mentioned telecommunication service.Find that having eight attribute column and user to order A set meal in user's ticket exists the larger degree of association by association analysis in early stage.Thus user's call bill data in Heilongjiang Province's 1-6 month in 2014 can be used, therefrom extract the basic data of 5,000,000 Subscriber Number tickets (comprise and order A set meal user and do not order A set meal user) as training, and pre-service is carried out to basic data, and then trained through pretreated data by genetic algorithm and neural network algorithm, obtain training result.And when needing to recommend A set meal to the user not ordering A set meal, in order to make the degree of accuracy of the recommend method of the telecommunication service in third embodiment of the invention more distinct, extracting at this 600,000 users 1-6 month call bill data in 2014 not ordering A set meal and identifying the user wherein accepting A set meal.Be divided into 3 groups, often organize 200,000 users, use the algorithm in the third embodiment of the present invention respectively, neural network algorithm, artificial experience method identifies the user that the meeting predicted separately accepts A set meal, can accept each taking-up 5000 numbers the user of A set meal from three groups, sends 10086 port A set meals marketing notes for these users that can accept A set meal, user can handle by replying letter, as shown in table 1 from BOSS system statistics user order situation.In addition, similar, use above same method, the user 100,000 not being ordered to emotional affection net telephone expenses set meal identifies, obtain the user that 29744 meetings accept emotional affection net telephone expenses set meal, emotional affection net is carried out to these users that can accept emotional affection net telephone expenses set meal and orders sending short messages in groups, have 10321 family users and order this business by uplink short message mode, discrimination reaches 34.7%, and normal short message mass-sending marketing user order rate is 1 ~ 5%.It can thus be appreciated that the recommend method of the telecommunication service that third embodiment of the invention provides is much more accurate than the method for conventional recommendation telecommunication service.
| In 3rd embodiment | Neural network is calculated | Artificial experience method |
| Algorithm | Method | ||
| Subscribed users number | 454 | 339 | 246 |
| Recognition accuracy (%) | 9.08％ | 6.78％ | 4.92％ |
Table 1
Wherein, in the third embodiment of the present invention, above-mentioned pre-service is carried out to basic data, obtain intermediate data, specifically comprise: first require to screen basic data according to the screening preset, remove unwanted data in follow-up genetic algorithm and neural network algorithm, obtain the data after screening; Then according to the data layout preset, the data after screening are sorted, that is, the data after screening are spliced into data line by line, and the unified data layout according to presetting stores; Be normalized the data after sequence, the absolute value by each attribute value becomes certain relative value relation, obtains the intermediate data with the data structure that genetic algorithm can identify.
By genetic algorithm and neural network algorithm, intermediate data is trained in elaboration, before obtaining training result, first simply set forth the method that genetic algorithm and neural network algorithm are combined.Concrete steps are as follows:
(1) by genetic algorithm, initial weight distribution is optimized, in solution space, orients some good search volumes:
1. parameter coding, connects weights (w by random initializtion
i) and Node B threshold (￠ i) be shown as the genotype string structure data in hereditary space by coding schedule, as shown in table 2:
| W1 | .. | Wk | ￠1 | .. | ￠m |
Table 2
2. make t=0, determine initial population, the node serial number of neural network, connection weight and threshold value are formed one group of data, represent body one by one, several such data groups just form initial population.
3. calculate fitness function F, based on the network energy function E of the output node error of neural network, selected F=C/E, in formula, C is a constant.
4. selecting (copying) operation, for preventing the optimal result searched from losing, directly entering into colony of future generation maximum for fitness in previous generation colony 10%.The ratio that copies of 90% individuality is in addition determined by fitness:
5. carry out swap operation, suitably choose the probability P of exchange
c, generally with 0.85 as well.
6. carry out mutation operation, effect is useful may the separating that prevent loss, and makes algorithm have global convergence.
7. make t=t+1, continue to calculate fitness f
i, i=1,2 ... n.
8. judge whether end condition meets, namely t=N (N is the iterations of specifying in advance) then makes s=0, enters (2), otherwise returns 4..
(2) in the colony obtained (1) step, each new individuality adopts the some steps of neural network algorithm iteration, and then operates by genetic algorithm:
1. make s=s+1, each individuality in colony is decoded, obtain the connection weights and threshold of corresponding network; According to the maximum frequency of training in local, revised by the connection weights and threshold of neural network algorithm to this network, then being carried out by the connection weights and threshold of network encoding generates corresponding new individuality, the colony that all these new individuality compositions one are new.
2. for new colony, carry out the calculating of ideal adaptation degree, if the error corresponding to the individuality that in colony, fitness is maximum is less than lower limit or s=M (M is the iterations of specifying in advance), then proceed to 3.; Otherwise select (copying) to operate to new colony, swap operation, 1. mutation operation, turn to.
3. individuality maximum for fitness is decoded, obtain the connection weights and threshold of required network.
In whole searching process, the effect of often kind of method is all limited in " part " scope, thus is ensured the global convergence of study on the one hand by genetic algorithm, overcomes Gauss-Newton method to the dependence of initial value and local convergence problem; On the other hand, with the combination of " accurately " Gauss-Newton learning algorithm also overcome simple genetic algorithm with randomness and probability problem, and contribute to the search efficiency improving it.The result done like this, will be expected to significantly improve convergence, reduction it to the degree of dependence of starting condition while ensure its convergence direction, even if when limited known to problem, still can obtain satisfied training result.
As shown in Figure 4, in the third embodiment of the present invention, above by genetic algorithm and neural network algorithm, intermediate data is trained, obtains training result, specifically comprise:
Step S401, carries out interative computation by genetic algorithm to intermediate data, obtains initial population;
Step S402, decodes to initial population, obtains each individuality in initial population;
Step S403, is optimized by the testing accuracy of neural network algorithm to each individuality in initial population, obtains transitional population;
Step S404, judges whether to exist in transitional population the individuality that testing accuracy is greater than default testing accuracy; If there is the individuality that testing accuracy is greater than default testing accuracy in transitional population, perform step S405, otherwise perform step S406;
Step S405, using this individuality as training result;
Step S406, according to default testing accuracy, by genetic algorithm, transitional population is selected, crossover and mutation operation, obtain new colony;
Step S407, decodes to new colony, and is optimized by the testing accuracy of neural network algorithm to decoded each individuality, obtains new transitional population;
Step S408, judges whether to exist in new transitional population the individuality that testing accuracy is greater than default testing accuracy, and the individuality that testing accuracy is greater than default testing accuracy if exist in new transitional population, then perform step S409, otherwise perform step S410;
Step S409, using this individuality as training result;
Step S410, continue through that genetic algorithm is selected new transitional population, crossover and mutation operation, obtain new colony, and continue through the testing accuracy of neural network algorithm to each individuality in new colony and be optimized, obtain new transitional population, until there is the individuality that testing accuracy is greater than default testing accuracy in new transitional population, and using this individuality as training result.
In the third embodiment of the present invention, obtaining in the process of training result by genetic algorithm and neural network algorithm, can judge whether the interative computation number of times of genetic algorithm reaches preset times, when the interative computation number of times of genetic algorithm reaches preset times, also can deconditioning even if there is not individuality that testing accuracy is greater than default testing accuracy, and the highest individuality of testing accuracy in the new transitional population last computing obtained is as training result.
Wherein, neural network algorithm realizes to the switching of genetic algorithm by the maximum step number of limited default testing accuracy or local, genetic algorithm realizes to the switching of neural network algorithm by once complete genetic operator operation (such as selection opertor, report to the leadship after accomplishing a task operator or mutation operator), mutual like this using the training result of the other side as oneself initial weight or initial population, alternately training repeatedly, until reach the preset times of the interative computation number of times of limited default testing accuracy or genetic algorithm.
4th embodiment
The embodiment of the recommendation apparatus of a kind of telecommunication service provided for fourth embodiment of the invention below.Embodiment and the above-mentioned embodiment of the method for the recommendation apparatus of described telecommunication service belong to same design, and the detail content of not detailed description in the embodiment of the recommendation apparatus of telecommunication service can with reference to said method embodiment.
As shown in Figure 5, the fourth embodiment of the present invention provides a kind of recommendation apparatus of telecommunication service, and this device comprises:
Prediction module 51, for according to the historical information of targeted customer and the training result that whether can accept telecommunication service to be recommended for inspection user that obtains in advance, carries out the prediction of telecommunication service coupling to targeted customer, obtains the result of telecommunication service coupling prediction;
Identification module 52, for the result according to the prediction of telecommunication service coupling, identifies the user that can accept telecommunication service to be recommended;
Sending module 53, for sending telecommunication service recommendation information to the user that can accept telecommunication service to be recommended.
Wherein, identification module 52 comprises:
Judging unit, for judging whether the result of telecommunication service coupling prediction is greater than preset value, and when the result of telecommunication service coupling prediction is greater than preset value, triggers determining unit;
Determining unit, for the triggering according to judging unit, determines that this targeted customer is the user that can accept telecommunication service to be recommended.
Wherein, device also comprises:
Acquisition module, for obtaining the training result that whether can accept telecommunication service to be recommended for inspection user.
Wherein, acquisition module comprises:
Acquiring unit, for obtaining the basic data of training;
Processing unit, for carrying out pre-service to basic data, obtains intermediate data;
Training unit, for being trained intermediate data by genetic algorithm and neural network algorithm, obtains training result.
Wherein, processing unit comprises:
Screening subelement, for requiring to screen basic data according to the screening preset, obtains the data after screening;
Sequence subelement, for according to the data layout preset, sorts to the data after screening;
Normalizing unit, for being normalized the data after sequence, obtains the intermediate data with the data structure that genetic algorithm can identify.
Wherein, training unit comprises:
Operator unit, for carrying out interative computation by genetic algorithm to intermediate data, obtains initial population;
Decoding subelement, for decoding to initial population, obtains each individuality in initial population;
Optimizing subelement, for being optimized by the testing accuracy of neural network algorithm to each individuality in initial population, obtaining transitional population;
Judgment sub-unit, for judging whether to exist in transitional population the individuality that testing accuracy is greater than default testing accuracy, and when there is testing accuracy in transitional population and being greater than default testing accuracy individual, triggers first and determines subelement;
First determines subelement, for the triggering according to judgment sub-unit, using this individuality as training result.
Wherein, device also comprises:
First operational module, for when there is not testing accuracy in transitional population and being greater than default testing accuracy individual, according to default testing accuracy, is selected transitional population by genetic algorithm, crossover and mutation operation, obtains new colony;
Decoder module, for decoding to new colony, and being optimized by the testing accuracy of neural network algorithm to decoded each individuality, obtaining new transitional population;
First judge module, for judging whether to exist in new transitional population the individuality that testing accuracy is greater than default testing accuracy, and in new transitional population, there is testing accuracy when being greater than default testing accuracy individual, trigger determination module, in new transitional population, there is not testing accuracy when being greater than default testing accuracy individual, trigger the second operational module;
Determination module, for the triggering according to the first judge module, using this individuality as training result;
Second operational module, for the triggering according to the first judge module, continue through that genetic algorithm is selected new transitional population, crossover and mutation operation, obtain new colony, and continue through the testing accuracy of neural network algorithm to each individuality in new colony and be optimized, obtain new transitional population, until there is the individuality that testing accuracy is greater than default testing accuracy in new transitional population, and using this individuality as training result.
Wherein, device also comprises:
Second judge module, for judging whether the interative computation number of times of genetic algorithm reaches preset times, and when the interative computation number of times of genetic algorithm reaches preset times, triggers stopping modular;
Stopping modular, for the triggering according to the second judge module, deconditioning, and the highest individuality of testing accuracy in the new transitional population last computing obtained is as training result.
In the fourth embodiment of the present invention, the recommendation apparatus of telecommunication service according to the historical information of targeted customer and the training result that whether can accept telecommunication service to be recommended for inspection user obtained in advance, can obtain the result of telecommunication service coupling prediction.And then judge whether this targeted customer is the user that can accept telecommunication service to be recommended according to the result of the telecommunication service coupling prediction obtained, when it is the user that can accept telecommunication service to be recommended, send telecommunication service recommendation information to it.So just, accurate business marketing pattern can be realized, reach the object of accurately recommending telecommunication service to user.
It should be noted that, the recommendation apparatus of the telecommunication service that the embodiment of the present invention provides is the device of application said method, and namely all embodiments of said method are all applicable to this device, and all can reach same or analogous beneficial effect.
The above is the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the prerequisite not departing from principle of the present invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.
Claims (16)
1. a recommend method for telecommunication service, is characterized in that, described method comprises:
According to the historical information of targeted customer and the training result that whether can accept telecommunication service to be recommended for inspection user that obtains in advance, the prediction of telecommunication service coupling is carried out to described targeted customer, obtains the result of telecommunication service coupling prediction;
According to the result of described telecommunication service coupling prediction, identify the user that can accept telecommunication service to be recommended;
Telecommunication service recommendation information is sent to the user that can accept telecommunication service to be recommended.
2. recommend method as claimed in claim 1, is characterized in that, the described result according to the prediction of described telecommunication service coupling, identifies the user that can accept telecommunication service to be recommended, specifically comprises:
Judge whether the result of described telecommunication service coupling prediction is greater than preset value;
If the result of described telecommunication service coupling prediction is greater than preset value, then determine that this targeted customer is the user that can accept telecommunication service to be recommended.
3. recommend method as claimed in claim 1, it is characterized in that, at the described historical information according to targeted customer and the training result that whether can accept telecommunication service to be recommended for inspection user that obtains in advance, the prediction of telecommunication service coupling is carried out to described targeted customer, before obtaining the result of telecommunication service coupling prediction, described method also comprises:
Obtain and be used for the training result whether inspection user can accept telecommunication service to be recommended.
4. recommend method as claimed in claim 3, is characterized in that, described acquisition is used for the training result whether inspection user can accept telecommunication service to be recommended, specifically comprises:
Obtain the basic data required for training;
Pre-service is carried out to described basic data, obtains intermediate data;
By genetic algorithm and neural network algorithm, described intermediate data is trained, obtain described training result.
5. recommend method as claimed in claim 4, is characterized in that, describedly carries out pre-service to described basic data, obtains intermediate data, specifically comprises:
Require to screen described basic data according to the screening preset, obtain the data after screening;
According to the data layout preset, the data after described screening are sorted;
Data after sequence are normalized, obtain the intermediate data with the data structure that described genetic algorithm can identify.
6. recommend method as claimed in claim 4, is characterized in that, is describedly trained described intermediate data by genetic algorithm and neural network algorithm, obtains described training result, specifically comprises:
By genetic algorithm, interative computation is carried out to described intermediate data, obtain initial population;
Described initial population is decoded, obtains each individuality in described initial population;
Be optimized by the testing accuracy of neural network algorithm to each individuality in described initial population, obtain transitional population;
Judge whether to exist in described transitional population the individuality that testing accuracy is greater than default testing accuracy;
If there is the individuality that testing accuracy is greater than default testing accuracy in described transitional population, then using this individuality as training result.
7. recommend method as claimed in claim 6, it is characterized in that, described method also comprises:
If there is not the individuality that testing accuracy is greater than default testing accuracy in described transitional population, then according to preset testing accuracy, by genetic algorithm, described transitional population is selected, crossover and mutation operation, obtain new colony;
Described new colony is decoded, and is optimized by the testing accuracy of neural network algorithm to decoded each individuality, obtain new transitional population;
Judge whether to exist in new transitional population the individuality that testing accuracy is greater than default testing accuracy;
If there is the individuality that testing accuracy is greater than default testing accuracy in described new transitional population, then using this individuality as training result;
If there is not the individuality that testing accuracy is greater than default testing accuracy in described new transitional population, then continue through that genetic algorithm is selected new transitional population, crossover and mutation operation, obtain new colony, and continue through the testing accuracy of neural network algorithm to each individuality in new colony and be optimized, obtain new transitional population, until there is the individuality that testing accuracy is greater than default testing accuracy in new transitional population, and using this individuality as training result.
8. recommend method as claimed in claim 7, it is characterized in that, be optimized at the described testing accuracy of neural network algorithm to each individuality in new colony that continue through, after obtaining new transitional population, described method also comprises:
Judge whether the interative computation number of times of genetic algorithm reaches preset times;
If the interative computation number of times of genetic algorithm reaches preset times, then deconditioning, and the highest individuality of testing accuracy in the new transitional population last computing obtained is as training result.
9. a recommendation apparatus for telecommunication service, is characterized in that, described device comprises:
Prediction module, for according to the historical information of targeted customer and the training result that whether can accept telecommunication service to be recommended for inspection user that obtains in advance, carries out the prediction of telecommunication service coupling to described targeted customer, obtains the result of telecommunication service coupling prediction;
Identification module, for the result according to the prediction of described telecommunication service coupling, identifies the user that can accept telecommunication service to be recommended;
Sending module, for sending telecommunication service recommendation information to the user that can accept telecommunication service to be recommended.
10. recommendation apparatus as claimed in claim 9, it is characterized in that, described identification module comprises:
Judging unit, for judging whether the result of described telecommunication service coupling prediction is greater than preset value, and when the result of described telecommunication service coupling prediction is greater than preset value, triggers determining unit;
Determining unit, for the triggering according to described judging unit, determines that this targeted customer is the user that can accept telecommunication service to be recommended.
11. recommendation apparatus as claimed in claim 9, it is characterized in that, described device also comprises:
Acquisition module, for obtaining the training result that whether can accept telecommunication service to be recommended for inspection user.
12. recommendation apparatus as claimed in claim 11, it is characterized in that, described acquisition module comprises:
Acquiring unit, for obtaining the basic data of training;
Processing unit, for carrying out pre-service to described basic data, obtains intermediate data;
Training unit, for being trained described intermediate data by genetic algorithm and neural network algorithm, obtains described training result.
13. recommendation apparatus as claimed in claim 12, it is characterized in that, described processing unit comprises:
Screening subelement, for requiring to screen described basic data according to the screening preset, obtains the data after screening;
Sequence subelement, for according to the data layout preset, sorts to the data after described screening;
Normalizing unit, for being normalized the data after sequence, obtains the intermediate data with the data structure that described genetic algorithm can identify.
14. recommendation apparatus as claimed in claim 12, it is characterized in that, described training unit comprises:
Operator unit, for carrying out interative computation by genetic algorithm to described intermediate data, obtains initial population;
Decoding subelement, for decoding to described initial population, obtains each individuality in described initial population;
Optimizing subelement, for being optimized by the testing accuracy of neural network algorithm to each individuality in described initial population, obtaining transitional population;
Judgment sub-unit, for judging whether to exist in described transitional population the individuality that testing accuracy is greater than default testing accuracy, and when there is testing accuracy in described transitional population and being greater than default testing accuracy individual, triggers first and determines subelement;
First determines subelement, for the triggering according to described judgment sub-unit, using this individuality as training result.
15. recommendation apparatus as claimed in claim 14, it is characterized in that, described device also comprises:
First operational module, for when there is not testing accuracy in described transitional population and being greater than default testing accuracy individual, according to default testing accuracy, is selected described transitional population by genetic algorithm, crossover and mutation operation, obtains new colony;
Decoder module, for decoding to described new colony, and being optimized by the testing accuracy of neural network algorithm to decoded each individuality, obtaining new transitional population;
First judge module, for judging whether to exist in new transitional population the individuality that testing accuracy is greater than default testing accuracy, and when there is testing accuracy in described new transitional population and being greater than default testing accuracy individual, trigger determination module, when there is not testing accuracy in described new transitional population and being greater than default testing accuracy individual, trigger the second operational module;
Determination module, for the triggering according to described first judge module, using this individuality as training result;
Second operational module, for the triggering according to described first judge module, continue through that genetic algorithm is selected new transitional population, crossover and mutation operation, obtain new colony, and continue through the testing accuracy of neural network algorithm to each individuality in new colony and be optimized, obtain new transitional population, until there is the individuality that testing accuracy is greater than default testing accuracy in new transitional population, and using this individuality as training result.
16. recommendation apparatus as claimed in claim 15, it is characterized in that, described device also comprises:
Second judge module, for judging whether the interative computation number of times of genetic algorithm reaches preset times, and when the interative computation number of times of genetic algorithm reaches preset times, triggers stopping modular;
Stopping modular, for the triggering according to described second judge module, deconditioning, and the highest individuality of testing accuracy in the new transitional population last computing obtained is as training result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510329668.5A CN104866626B (en) | 2015-06-15 | 2015-06-15 | A kind of recommendation method and device of telecommunication service |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510329668.5A CN104866626B (en) | 2015-06-15 | 2015-06-15 | A kind of recommendation method and device of telecommunication service |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104866626A true CN104866626A (en) | 2015-08-26 |
| CN104866626B CN104866626B (en) | 2018-06-26 |
Family
ID=53912452
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510329668.5A Active CN104866626B (en) | 2015-06-15 | 2015-06-15 | A kind of recommendation method and device of telecommunication service |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104866626B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105894028A (en) * | 2016-03-31 | 2016-08-24 | 百度在线网络技术（北京）有限公司 | User identification method and device |
| CN106600369A (en) * | 2016-12-09 | 2017-04-26 | 广东奡风科技股份有限公司 | Real-time recommendation system and method of financial products of banks based on Naive Bayesian classification |
| CN106815221A (en) * | 2015-11-27 | 2017-06-09 | 华为软件技术有限公司 | A kind of telecommunication service relation decision method and device |
| CN107766556A (en) * | 2017-11-03 | 2018-03-06 | 福建工程学院 | A kind of interactive Ontology Matching method and computer equipment based on evolution algorithm |
| CN109600757A (en) * | 2018-11-29 | 2019-04-09 | 南京亚信软件有限公司 | Prediction technique, device, computer equipment and the storage medium of base station dilatation |
| CN109919675A (en) * | 2019-03-04 | 2019-06-21 | 深圳微品致远信息科技有限公司 | Communication user upshift prediction probability recognition methods neural network based and system |
| CN110378739A (en) * | 2019-07-23 | 2019-10-25 | 中国联合网络通信集团有限公司 | A kind of data traffic matching process and device |
| TWI726398B (en) * | 2019-08-27 | 2021-05-01 | 中華電信股份有限公司 | Self-adapted telecommunication service recommend system and method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7013238B1 (en) * | 2003-02-24 | 2006-03-14 | Microsoft Corporation | System for delivering recommendations |
| US8380548B2 (en) * | 2007-02-15 | 2013-02-19 | So-ling Carmen Ng | Method for managing intellectual property |
| CN103107936A (en) * | 2011-11-11 | 2013-05-15 | 中国移动通信集团上海有限公司 | Method and device for sending information |
-
2015
- 2015-06-15 CN CN201510329668.5A patent/CN104866626B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7013238B1 (en) * | 2003-02-24 | 2006-03-14 | Microsoft Corporation | System for delivering recommendations |
| US8380548B2 (en) * | 2007-02-15 | 2013-02-19 | So-ling Carmen Ng | Method for managing intellectual property |
| CN103107936A (en) * | 2011-11-11 | 2013-05-15 | 中国移动通信集团上海有限公司 | Method and device for sending information |
Non-Patent Citations (4)
| Title |
|---|
| LADYS RODRIGUEZ.ETC: "Color recommendation system combining design concepts with interactive customers preference modeling from context changes", 《EVOLUTIONARY COMPUTATION (CEC)》 * |
| 兰慧: "基于消费行为的电信套餐推荐模型研究", 《中国优秀硕士学位论文全文数据库》 * |
| 吴仕勇: "基于数值计算方法的BP神经网络及遗传算法的优化研究", 《中国优秀硕士学位论文全文数据库》 * |
| 吴建生等: "遗传算法BP神经网络的预报研究和应用", 《数学的实践与认识》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106815221A (en) * | 2015-11-27 | 2017-06-09 | 华为软件技术有限公司 | A kind of telecommunication service relation decision method and device |
| CN105894028A (en) * | 2016-03-31 | 2016-08-24 | 百度在线网络技术（北京）有限公司 | User identification method and device |
| CN106600369A (en) * | 2016-12-09 | 2017-04-26 | 广东奡风科技股份有限公司 | Real-time recommendation system and method of financial products of banks based on Naive Bayesian classification |
| CN107766556A (en) * | 2017-11-03 | 2018-03-06 | 福建工程学院 | A kind of interactive Ontology Matching method and computer equipment based on evolution algorithm |
| CN107766556B (en) * | 2017-11-03 | 2021-07-30 | 福建工程学院 | Interactive ontology matching method based on evolutionary algorithm and computer equipment |
| CN109600757A (en) * | 2018-11-29 | 2019-04-09 | 南京亚信软件有限公司 | Prediction technique, device, computer equipment and the storage medium of base station dilatation |
| CN109600757B (en) * | 2018-11-29 | 2022-01-18 | 南京亚信软件有限公司 | Prediction method and device for base station capacity expansion, computer equipment and storage medium |
| CN109919675A (en) * | 2019-03-04 | 2019-06-21 | 深圳微品致远信息科技有限公司 | Communication user upshift prediction probability recognition methods neural network based and system |
| CN110378739A (en) * | 2019-07-23 | 2019-10-25 | 中国联合网络通信集团有限公司 | A kind of data traffic matching process and device |
| CN110378739B (en) * | 2019-07-23 | 2022-03-29 | 中国联合网络通信集团有限公司 | Data traffic matching method and device |
| TWI726398B (en) * | 2019-08-27 | 2021-05-01 | 中華電信股份有限公司 | Self-adapted telecommunication service recommend system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104866626B (en) | 2018-06-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104866626A (en) | Method and device for recommending telecommunication service | |
| CN103761254B (en) | Method for matching and recommending service themes in various fields | |
| CN105975479B (en) | A kind of telecommunication user interest-degree analysis method and system based on tag library | |
| CN101951617B (en) | Mobile network communication quality evaluation method based on analytic hierarchy process | |
| CN104750760B (en) | A kind of implementation method and device for recommending application software | |
| CN101923689A (en) | Method for determining advertising information launched audience and related device thereof | |
| CN106651232B (en) | Freight note number data analysis method and device | |
| CN103678531A (en) | Friend recommendation method and friend recommendation device | |
| CN107169063A (en) | A kind of user property Forecasting Methodology and system based on social information | |
| CN109522475A (en) | A kind of merchant recommendation method based on user's history consumption data | |
| CN105824818A (en) | Informationized management method, platform and system | |
| CN107527240A (en) | A kind of operator's industry product Praise effect identification system and method | |
| CN102664744B (en) | Group-sending recommendation method in network message communication | |
| CN103250376A (en) | Method and system for carrying out predictive analysis relating to nodes of a communication network | |
| CN108429776A (en) | Method for pushing, device, client, interactive device and the system of network object | |
| CN110781256B (en) | Method and device for determining POI matched with Wi-Fi based on sending position data | |
| CN105300398A (en) | Method, device and system for acquiring site information | |
| CN104102694B (en) | Tree node sort method and tree node collator | |
| CN109308564A (en) | The recognition methods of crowd's performance ratings, device, storage medium and computer equipment | |
| CN105740434A (en) | Network information scoring method and device | |
| Puri et al. | Pragmatics and semantics to connect specific local laws with public reactions | |
| CN116188115A (en) | Supplier recommendation system and method based on supply chain network demand and preference mining | |
| CN110163525A (en) | Terminal recommended method and terminal recommender system | |
| CN106682030A (en) | Method and device for information processing | |
| WO2024001102A1 (en) | Method and apparatus for intelligently identifying family circle in communication industry, and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |