CN106127345A - The Forecasting Methodology of a kind of mobile subscriber complaint and prognoses system - Google Patents
The Forecasting Methodology of a kind of mobile subscriber complaint and prognoses system Download PDFInfo
- Publication number
- CN106127345A CN106127345A CN201610493634.4A CN201610493634A CN106127345A CN 106127345 A CN106127345 A CN 106127345A CN 201610493634 A CN201610493634 A CN 201610493634A CN 106127345 A CN106127345 A CN 106127345A
- Authority
- CN
- China
- Prior art keywords
- user
- training sample
- report
- complaint
- report user
- 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
- 238000000034 method Methods 0.000 title claims abstract description 28
- 238000012549 training Methods 0.000 claims abstract description 100
- 235000012054 meals Nutrition 0.000 claims abstract description 31
- 238000005192 partition Methods 0.000 claims description 4
- 230000000052 comparative effect Effects 0.000 claims description 3
- 241000208340 Araliaceae Species 0.000 claims 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims 1
- 235000003140 Panax quinquefolius Nutrition 0.000 claims 1
- 235000008434 ginseng Nutrition 0.000 claims 1
- 238000011156 evaluation Methods 0.000 abstract description 8
- 238000012545 processing Methods 0.000 abstract description 7
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/01—Customer relationship services
- G06Q30/015—Providing customer assistance, e.g. assisting a customer within a business location or via helpdesk
- G06Q30/016—After-sales
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/50—Network service management, e.g. ensuring proper service fulfilment according to agreements
- H04L41/5061—Network service management, e.g. ensuring proper service fulfilment according to agreements characterised by the interaction between service providers and their network customers, e.g. customer relationship management
- H04L41/5064—Customer relationship management
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- General Business, Economics & Management (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- Signal Processing (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Game Theory and Decision Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The present invention provides the Forecasting Methodology and prognoses system that a kind of mobile subscriber complains.This Forecasting Methodology includes: definition user characteristics is to be exceeded, beyond the accounting of set meal and call, the bivector that the accounting of set meal forms by customer flow;The feature of user and the situation of complaint in analyzing and training sample;According to the feature of user, complaint situation and the eigenvalue of user to be measured in training sample, it was predicted that whether user to be measured can complain.This Forecasting Methodology is by carrying out Scientific evaluation to the complaint of mobile subscriber tendency, it is possible to whether prediction user can complain exactly, thus improves mobile subscriber and complain the accuracy of prediction, helps speed up the processing speed of complaint, promotes the satisfaction of user.
Description
Technical field
The present invention relates to communication technical field, in particular it relates to the Forecasting Methodology that a kind of mobile subscriber complains with prediction is
System.
Background technology
It is a link particularly important in carrier service work that customer complaint processes, in existing complaint handling, general
All over passively listening attentively to and record client's demand for contact personnel, this method efficiency is low and Consumer's Experience is bad.
In current operator keen competition environment, do not require nothing more than and the customer complaint produced can be carried out efficiently
Process, require that the business to user uses perception to carry out real-time monitored simultaneously, the complaint of user is inclined to and carries out Scientific evaluation, right
The complaint behavior of user is predicted, and carries out pretreatment measure the most in advance.How look-ahead customer complaint reality
Paying a return visit the most in advance is a problem demanding prompt solution.
Summary of the invention
The present invention is directed to above-mentioned technical problem present in prior art, it is provided that the Forecasting Methodology that a kind of mobile subscriber complains
And prognoses system.This Forecasting Methodology is by carrying out Scientific evaluation to the complaint of mobile subscriber tendency, it is possible to predict user exactly
Whether can complain, thus improve mobile subscriber and complain the accuracy of prediction, help speed up the processing speed of complaint, promote and use
The satisfaction at family.
The present invention provides the Forecasting Methodology that a kind of mobile subscriber complains, including:
Definition user characteristics is to be exceeded, beyond the accounting of set meal and call, the two dimension that the accounting of set meal forms by customer flow
Vector;
The feature of user and the situation of complaint in analyzing and training sample;
According to the feature of user in described training sample, complain situation and the eigenvalue of user to be measured, it was predicted that described in treat
Survey whether user can complain.
Preferably, in described analyzing and training sample, feature and the complaint situation of user include:
Extract described training sample, according to the complaint record of users all in described training sample, by described training sample
In user be divided into report user and non-report user;
Add up report user described in described training sample and the described respective number of users of non-report user, and according to statistics
Result calculates the first shared in the total number of users of described training sample ratio of described report user and described non-report user
The second ratio shared in the total number of users of described training sample;
Calculate the eigenvalue of user in described training sample, and be many by the feature value division of user in described training sample
Individual range of attributes;
Add up described report user and the described respective number of users of non-report user in each described range of attributes, and according to
Statistical result calculate each described range of attributes in described report user in described report user's sum of described training sample
Non-report user described in the 3rd shared ratio and each described range of attributes is not complaining use described in described training sample
The 4th ratio shared in the sum of family.
It is preferably, described according to the feature of user, complaint situation and the eigenvalue of user to be measured in described training sample,
Predict whether described prediction user can complain to include:
Calculate the eigenvalue of described user to be measured;
Assume that described user to be measured is described report user, according to described first ratio and the eigenvalue of described user to be measured
Described 3rd ratio in the described range of attributes at place, uses Bayes theorem to calculate and obtains the first value;
Assume that described user to be measured is described non-report user, according to described second ratio and the feature of described user to be measured
Described 4th ratio in the described range of attributes at value place, uses Bayes theorem to calculate and obtains the second value;
The most described first value and the size of described second value, if described first value is more than described second value, then described
User to be measured is meeting report user;If described first value is less than described second value, the most described user to be measured is for will not complain use
Family.
Preferably, also include: predict the complaint time of described meeting report user;
Being ranked up the complaint time of multiple described meeting report users, preferential the described of return visit sequence minimum can complain use
Family.
Preferably, the complaint time of described prediction described meeting report user includes:
According to the eigenvalue of user in described training sample and the time of complaint, calculate and determine the constant in Logistic model
Parameter;
The eigenvalue of described meeting report user is substituted in described Logist ic model, calculates described meeting report user's
The complaint time.
The present invention also provides for the prognoses system that a kind of mobile subscriber complains, including:
Definition module, is to be exceeded accounting for of set meal by customer flow beyond the accounting of set meal and call for defining user characteristics
Bivector than composition;
Analyze module, the feature of user and the situation of complaint in analyzing and training sample;
First prediction module, for according to the feature of user, complaint situation and user to be measured in described training sample
Eigenvalue, it was predicted that whether described user to be measured can complain.
Preferably, described analysis module includes:
Extract taxon, be used for extracting described training sample, remember according to the complaint of users all in described training sample
Record, is divided into report user and non-report user by the user in described training sample;
First statistical computation unit, is used for adding up report user described in described training sample and described non-report user is each
From number of users, and calculate described report user in the total number of users of described training sample shared first according to statistical result
The second ratio that ratio and described non-report user are shared in the total number of users of described training sample;
Computation partition unit, for calculating the eigenvalue of user in described training sample, and will use in described training sample
The feature value division at family is multiple range of attributes;
Second statistical computation unit, is used for adding up in each described range of attributes described report user and described does not complains use
The respective number of users in family, and in calculating each described range of attributes according to statistical result described report user at described training sample
Described report user's sum in non-report user described in the 3rd shared ratio and each described range of attributes in described instruction
Practice the 4th ratio shared in described non-report user's sum of sample.
Preferably, described first prediction module includes:
First computing unit, for calculating the eigenvalue of described user to be measured;
Second computing unit, for according to described first ratio and the described attribute at the eigenvalue place of described user to be measured
In the range of described 3rd ratio, use Bayes theorem calculate obtain first value;
3rd computing unit, for according to described second ratio and the described attribute at the eigenvalue place of described user to be measured
In the range of described 4th ratio, use Bayes theorem calculate obtain second value;
Relatively determine unit, for the most described first value and the size of described second value, and determine according to comparative result
Whether described user to be measured is can report user.
Preferably, the second prediction module, order module and return visit module are also included;
Described second prediction module is for predicting the complaint time of described meeting report user;
Described order module is for being ranked up the complaint time of multiple described meeting report users;
Described return visit module pays a return visit described meeting report user for the sequence according to described order module.
Preferably, described second prediction module includes:
Calculating determines unit, and for according to the eigenvalue of user in described training sample and the time of complaint, calculating determines
Constant parameter in Logistic model;
4th computing unit, for substituting in described Logistic model by the eigenvalue of described meeting report user, calculates
The complaint time of described meeting report user.
Beneficial effects of the present invention: the Forecasting Methodology that mobile subscriber provided by the present invention complains, by mobile subscriber
Complaint tendency carry out Scientific evaluation, it is possible to whether prediction user can complain exactly, thus improves mobile subscriber and complains pre-
The accuracy surveyed, helps speed up the processing speed of complaint, promotes the satisfaction of user.
The prognoses system that mobile subscriber provided by the present invention complains, it is possible to obtain the technology identical with above-mentioned Forecasting Methodology
Effect.
Accompanying drawing explanation
Fig. 1 is the flow chart of the Forecasting Methodology that mobile subscriber complains in the embodiment of the present invention 1;
Fig. 2 is the flow chart of the Forecasting Methodology that mobile subscriber complains in the embodiment of the present invention 2;
Fig. 3 is the flow chart of step 2 in the embodiment of the present invention 2;
Fig. 4 is the flow chart of step 3 in the embodiment of the present invention 2;
Fig. 5 is the theory diagram of the prognoses system that mobile subscriber complains in the embodiment of the present invention 3;
Fig. 6 is the theory diagram of the prognoses system that mobile subscriber complains in the embodiment of the present invention 4.
Description of reference numerals therein:
1. definition module;2. analyze module;21. extract taxon;22. first statistical computation unit;23. computation partition
Unit;24. second statistical computation unit;3. the first prediction module;31. first computing units;32. second computing units;33.
Three computing units;34. compare and determine unit;4. the second prediction module;41. calculating determine unit;42. the 4th computing units;5.
Order module;6. pay a return visit module.
Detailed description of the invention
For making those skilled in the art be more fully understood that technical scheme, below in conjunction with the accompanying drawings and be embodied as
Forecasting Methodology and prognoses system that a kind of mobile subscriber provided by the present invention is complained by mode are described in further detail.
Embodiment 1:
The present embodiment provides the Forecasting Methodology that a kind of mobile subscriber complains, as it is shown in figure 1, include:
Step S1: definition user characteristics is that the accounting being exceeded set meal beyond the accounting of set meal and call by customer flow forms
Bivector.
Step S2: the feature of user and the situation of complaint in analyzing and training sample.
Step S3: according to the feature of user, complaint situation and the eigenvalue of user to be measured in training sample, it was predicted that to be measured
Whether user can complain.
The complaint of mobile subscriber can be inclined to and carry out Scientific evaluation by this Forecasting Methodology, thus improves mobile subscriber and complain
The accuracy of prediction, helps speed up the processing speed of complaint, promotes the satisfaction of user.
Embodiment 2:
The present embodiment provides the Forecasting Methodology that a kind of mobile subscriber complains, as in figure 2 it is shown, include:
Step S1: definition user characteristics is that the accounting being exceeded set meal beyond the accounting of set meal and call by customer flow forms
Bivector.
In this step, as user characteristics is defined as a bivector (f, t), wherein,
F represents that flow exceeds part and accounts for the ratio of flow package;fdRepresent that to inquire about the user of acquisition from client management system of that month
Flow package;frRepresent the of that month actually used flow of user inquiring about acquisition from user's internet records system;T represents that call accounts for the ratio of call set meal beyond part;tdRepresent and inquire about from client management system
The call set meal that the user obtained is of that month;trRepresent from BSS system, inquire about the actual air time that the user of acquisition is of that month.
Step S2: the feature of user and the situation of complaint in analyzing and training sample.
This step specifically includes: as it is shown on figure 3,
Step S21: extract training sample, according to the complaint record of users all in training sample, by training sample
User is divided into report user and non-report user.
In this step, as extracted flow from internet records data system, BSS system and client management system beyond set meal
With call beyond multiple users of set meal, using these users as training sample;Simultaneously according to the complaint record of these users, will
These users are divided into two classes: i.e. report user C1 and non-report user C2.
Step S22: report user and the respective number of users of non-report user in statistics training sample, and according to statistical result
Calculate the first shared in the total number of users of training sample ratio of report user and the non-report user user at training sample
The second ratio shared in sum.
In this step, being M1 as statistics obtains report user's sum in training sample, non-report user's sum is M2, then throw
The first ratio telling user shared in the total number of users of training sample is:Non-report user is in training
The second ratio shared in the total number of users of sample is:
Step S23: calculate the eigenvalue of user in training sample, and be many by the feature value division of user in training sample
Individual range of attributes.
In this step, it is divided into following four as the flow in user characteristics value in training sample exceeded accounting f of set meal
Individual range of attributes: 1. .f≤0.01;2. .0.01 < f≤0.05;3. .0.05 < f≤0.5;4. .f > 0.5.By in training sample
Call in user characteristics value is divided into three below range of attributes beyond accounting t of set meal: 1. .t≤0.1;2. .0.1 < t≤
0.5;3. .t > 0.5.
Step S24: add up report user and the respective number of users of non-report user in each range of attributes, and according to statistics
Result calculates the 3rd ratio and each that report user is shared in report user's sum of training sample in each range of attributes
The 4th ratio that in range of attributes, non-report user is shared in non-report user's sum of training sample.
In this step, such as the number of users A respectively of the report user in four range of attributes of f in statistics training sample11、
A12、A13、A14;The 3rd ratio that report user in four range of attributes of f is shared in report user's sum of training sample
Example is respectively as follows: The number of users of the non-report user in four range of attributes of f A respectively in statistics training sample21、A22、
A23、A24;The non-report user in four range of attributes of f the 4th ratio shared by non-report user's sum of training sample is divided
It is not:
The number of users of the report user in three range of attributes of t B respectively in statistics training sample11、B12、B13;At t three
The report user in individual range of attributes the 3rd ratio shared by report user's sum of training sample is respectively as follows: T's in statistics training sample
The number of users of the non-report user in three range of attributes B respectively21、B22、B23;Use is not complained in three range of attributes of t
The family the 4th ratio shared by non-report user's sum of training sample is respectively as follows:
Step S3: according to the feature of user, complaint situation and the eigenvalue of user to be measured in training sample, it was predicted that to be measured
Whether user can complain.
This step specifically includes: as shown in Figure 4,
Step S31: calculate the eigenvalue of user to be measured.
In this step, calculate the flow accounting f ' and the accounting t ' beyond set meal that converses beyond set meal of user to be measured.
Step S32: assume that user to be measured is report user, according to the genus at the eigenvalue place of the first ratio and user to be measured
The 3rd ratio in the range of property, uses Bayes theorem to calculate and obtains the first value.
In this step, as assume the flow of user to be measured beyond the accounting f ' of set meal in the range of attributes of f≤0.01, treat
Survey the call of user beyond the accounting t ' of set meal in the range of attributes of t > 0.5, if the first value is X1, then X1=argmax (P
(C1)P(f≤0.01|C1) P (t > 0.5 | C1))。
Step S33: assume that user to be measured is non-report user, according to the eigenvalue place of the second ratio and user to be measured
The 4th ratio in range of attributes, uses Bayes theorem to calculate and obtains the second value.
In this step, with step S32 in it is similarly assumed that the flow of user to be measured beyond the accounting f ' of set meal in f≤0.01
Range of attributes in, the call of user to be measured beyond the accounting t ' of set meal in the range of attributes of t > 0.5.If the second value is X2,
Then X2=argmax (P (C2)P(f≤0.01|C2) P (t > 0.5 | C2))。
Step S34: compare the first value and the size of the second value, if the first value is more than the second value, then execution step S35:
User to be measured is meeting report user.If the first value is less than the second value, then perform step S36: user to be measured is for will not complain use
Family.
As in figure 2 it is shown, the Forecasting Methodology in the present embodiment also includes: step S4: the complaint time of prediction meeting report user.
This step specifically includes:
Step S41: according to the eigenvalue of user in training sample and the time of complaint, calculate and determine in Logistic model
Constant parameter.
In this step, the formula of Logistic model is:Wherein, y is dependent variable, at this
Embodiment representing, user is occurring flow to exceed call set complaint time interval after the meal (such as sky beyond flow package or call
Number), K is parameter of saturation, represents the upper limit of complaint time, can rule of thumb set.F and t is independent variable, and f represents that flow exceeds
Part accounts for the accounting of flow package, and t represents that call accounts for the accounting of call set meal beyond part;α and β is weight parameter, can basis
Empirical value determines, a and b be in Logistic model it needs to be determined that constant parameter.
By in training sample each user eigenvalue (f, t) and complain (i.e. complaining time interval) y generation respectively time
Enter in above-mentioned Logistic model formation, estimate constant parameter a's and b in Logistic model according to method of least square
Value.
Step S42: will the eigenvalue of report user substitute in Logistic model, calculate can the complaint of report user time
Between.
In this step, the flow package of inquiry acquisition meeting report user and call set meal from client management system, and from
When in user's internet records system and BSS system, inquiry obtains the actually used flow of meeting report user and actual call respectively
Between, according to formula (1) and formula (2) calculate can report user eigenvalue (f, t), it will the eigenvalue of report user substitutes into step
Rapid S41 has determined that in the Logistic model of constant parameter, thus the complaint time finally calculating meeting report user (i.e. throws
Tell time interval).
Step S5: be ranked up the complaint time of multiple meeting report users, the preferential meeting complaint paying a return visit sequence minimum is used
Family.
In this step, owing to can report user frequently include multiple, so arranging of this step can be by the complaint time
Sequence, in advance the meeting report user that the complaint time is nearer is preferentially paid a return visit, thus improves Consumer's Experience, improve use
The satisfaction at family.
The beneficial effect of embodiment 1-2: the Forecasting Methodology that the mobile subscriber provided in embodiment 1-2 complains, by right
The complaint tendency of mobile subscriber carries out Scientific evaluation, it is possible to whether prediction user can complain exactly, thus improves mobile use
The accuracy of prediction is complained at family, helps speed up the processing speed of complaint, promotes the satisfaction of user.
Embodiment 3:
The present embodiment provides the prognoses system that a kind of mobile subscriber complains, as it is shown in figure 5, include: definition module 1, is used for
Definition user characteristics is to be exceeded, beyond the accounting of set meal and call, the bivector that the accounting of set meal forms by customer flow.Analyze
Module 2, the feature of user and the situation of complaint in analyzing and training sample.First prediction module 3, for according in training sample
The feature of user, complaint situation and the eigenvalue of user to be measured, it was predicted that whether user to be measured can complain.
In the present embodiment, analyze module 2 and include: extract taxon 21, be used for extracting training sample, according to training sample
In the complaint record of all users, the user in training sample is divided into report user and non-report user.First statistical computation
Unit 22, is used for adding up report user and the respective number of users of non-report user in training sample, and calculates according to statistical result
The first ratio that report user is shared in the total number of users of training sample and non-report user are in the total number of users of training sample
The second ratio shared by.Computation partition unit 23, for calculating the eigenvalue of user in training sample, and by training sample
The feature value division of user is multiple range of attributes.Second statistical computation unit 24, complains in being used for adding up each range of attributes
User and the respective number of users of non-report user, and calculate in each range of attributes report user at training sample according to statistical result
Not non-report user's not complaining at training sample in the 3rd ratio shared in this report user's sum and each range of attributes
The 4th ratio shared in total number of users.
In the present embodiment, the first prediction module 3 includes: the first computing unit 31, for calculating the eigenvalue of user to be measured.
Second computing unit 32, is used for according to the 3rd ratio in the range of attributes at the eigenvalue place of the first ratio and user to be measured,
Use Bayes theorem to calculate and obtain the first value.3rd computing unit 33, for according to the second ratio and the feature of user to be measured
The 4th ratio in the range of attributes at value place, uses Bayes theorem to calculate and obtains the second value.Relatively determine unit 34, be used for
Relatively first value and the size of the second value, and determine whether user to be measured is can report user according to comparative result.
Prognoses system in the present embodiment, it is possible to the complaint tendency of mobile subscriber is carried out Scientific evaluation, thus improves
Mobile subscriber complains the accuracy of prediction, helps speed up the processing speed of complaint, promotes the satisfaction of user.
Embodiment 4:
The present embodiment provides a kind of mobile subscriber the prognoses system of complaint, as different from Example 3, as shown in Figure 6,
In embodiment 3 on the basis of prognoses system, the prognoses system in the present embodiment also includes: the second prediction module 4, order module 5
With return visit module 6.Second prediction module 4 is for predicting the complaint time of meeting report user.Order module 5 is for throwing multiple
The complaint time telling user is ranked up.Pay a return visit module 6 for the meeting of the return visit report user of the sequence according to order module 5.
By arranging the second prediction module 4, order module 5 and paying a return visit module 6, it is possible in advance to the meeting that the time of complaint is nearer
Report user preferentially pays a return visit, thus improves Consumer's Experience, improves the satisfaction of user.
In the present embodiment, the second prediction module 4 includes: calculates and determines unit 41, for according to user in training sample
Eigenvalue and the time of complaint, calculate and determine the constant parameter in Logistic model.4th computing unit 42, for complaining
The eigenvalue of user substitutes in Logistic model, calculates the complaint time of meeting report user.
In the present embodiment, other principle modules and the function thereof of prognoses system are in the same manner as in Example 3, and here is omitted.
The beneficial effect of embodiment 3-4: the prognoses system that the mobile subscriber provided in embodiment 3-4 complains, by right
The complaint tendency of mobile subscriber carries out Scientific evaluation, it is possible to whether prediction user can complain exactly, thus improves mobile use
The accuracy of prediction is complained at family, helps speed up the processing speed of complaint, promotes the satisfaction of user.
It is understood that the principle that is intended to be merely illustrative of the present of embodiment of above and the exemplary enforcement that uses
Mode, but the invention is not limited in this.For those skilled in the art, in the essence without departing from the present invention
In the case of god and essence, can make various modification and improvement, these modification and improvement are also considered as protection scope of the present invention.
Claims (10)
1. the Forecasting Methodology that a mobile subscriber complains, it is characterised in that including:
Definition user characteristics is to be exceeded, beyond the accounting of set meal and call, the bivector that the accounting of set meal forms by customer flow;
The feature of user and the situation of complaint in analyzing and training sample;
According to the feature of user, complaint situation and the eigenvalue of user to be measured in described training sample, it was predicted that described use to be measured
Whether family can complain.
The Forecasting Methodology that mobile subscriber the most according to claim 1 complains, it is characterised in that in described analyzing and training sample
Feature and the complaint situation of user include:
Extract described training sample, according to the complaint record of users all in described training sample, by described training sample
User is divided into report user and non-report user;
Add up report user described in described training sample and the described respective number of users of non-report user, and according to statistical result
Calculate the first shared in the total number of users of described training sample ratio of described report user and described non-report user in institute
State the second ratio shared in the total number of users of training sample;
Calculate the eigenvalue of user in described training sample, and be multiple genus by the feature value division of user in described training sample
Property scope;
Add up described report user and the described respective number of users of non-report user in each described range of attributes, and according to statistics
In result calculates each described range of attributes, described report user is shared in described report user's sum of described training sample
The 3rd ratio and each described range of attributes in described in non-report user total report user non-described in described training sample
The 4th ratio shared in number.
The Forecasting Methodology that mobile subscriber the most according to claim 2 complains, it is characterised in that described according to described training sample
The feature of user, complaint situation and the eigenvalue of user to be measured in Ben, it was predicted that whether described prediction user can complain includes:
Calculate the eigenvalue of described user to be measured;
Assume that described user to be measured is described report user, according to described first ratio and the eigenvalue place of described user to be measured
Described range of attributes in described 3rd ratio, use Bayes theorem calculate obtain first value;
Assume that described user to be measured is described non-report user, according to the eigenvalue institute of described second ratio and described user to be measured
Described range of attributes in described 4th ratio, use Bayes theorem calculate obtain second value;
The most described first value and the size of described second value, if described first value is more than described second value, the most described to be measured
User is meeting report user;If described first value is less than described second value, the most described user to be measured is will not report user.
The Forecasting Methodology that mobile subscriber the most according to claim 3 complains, it is characterised in that also include: predict described meeting
The complaint time of report user;
The complaint time of multiple described meeting report users is ranked up, the preferential described meeting report user paying a return visit sequence minimum.
The Forecasting Methodology that mobile subscriber the most according to claim 4 complains, it is characterised in that described prediction is described can complain
The complaint time of user includes:
According to the eigenvalue of user in described training sample and the time of complaint, calculate the constant ginseng determining in Logistic model
Number;
The eigenvalue of described meeting report user is substituted in described Logistic model, when calculating the complaint of described meeting report user
Between.
6. the prognoses system that a mobile subscriber complains, it is characterised in that including:
Definition module, is the accounting group being exceeded set meal by customer flow beyond the accounting of set meal and call for defining user characteristics
The bivector become;
Analyze module, the feature of user and the situation of complaint in analyzing and training sample;
First prediction module, for according to the feature of user, complaint situation and the feature of user to be measured in described training sample
Value, it was predicted that whether described user to be measured can complain.
The prognoses system that mobile subscriber the most according to claim 6 complains, it is characterised in that described analysis module includes:
Extract taxon, be used for extracting described training sample, according to the complaint record of users all in described training sample, will
User in described training sample is divided into report user and non-report user;
First statistical computation unit, is used for adding up report user described in described training sample and described non-report user is respective
Number of users, and calculate, according to statistical result, the first ratio that described report user is shared in the total number of users of described training sample
Second ratio shared in the total number of users of described training sample with described non-report user;
Computation partition unit, for calculating the eigenvalue of user in described training sample, and by user in described training sample
Feature value division is multiple range of attributes;
Second statistical computation unit, is used for adding up described report user and described non-report user in each described range of attributes each
From number of users, and calculate in each described range of attributes described report user in the institute of described training sample according to statistical result
Described in stating in report user's sum in the 3rd shared ratio and each described range of attributes, non-report user is at described training sample
The 4th ratio shared in this described non-report user's sum.
The prognoses system that mobile subscriber the most according to claim 7 complains, it is characterised in that described first prediction module bag
Include:
First computing unit, for calculating the eigenvalue of described user to be measured;
Second computing unit, for according to described first ratio and the described range of attributes at the eigenvalue place of described user to be measured
Interior described 3rd ratio, uses Bayes theorem to calculate and obtains the first value;
3rd computing unit, for according to described second ratio and the described range of attributes at the eigenvalue place of described user to be measured
Interior described 4th ratio, uses Bayes theorem to calculate and obtains the second value;
Relatively determine unit, for the most described first value and the size of described second value, and determine according to comparative result described
Whether user to be measured is can report user.
The prognoses system that mobile subscriber the most according to claim 8 complains, it is characterised in that also include the second prediction mould
Block, order module and return visit module;
Described second prediction module is for predicting the complaint time of described meeting report user;
Described order module is for being ranked up the complaint time of multiple described meeting report users;
Described return visit module pays a return visit described meeting report user for the sequence according to described order module.
The prognoses system that mobile subscriber the most according to claim 9 complains, it is characterised in that described second prediction module
Including:
Calculating determines unit, and for according to the eigenvalue of user in described training sample and the time of complaint, calculating determines
Constant parameter in Logistic model;
4th computing unit, for substituting in described Logistic model by the eigenvalue of described meeting report user, calculates described
The complaint time of meeting report user.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610493634.4A CN106127345A (en) | 2016-06-28 | 2016-06-28 | The Forecasting Methodology of a kind of mobile subscriber complaint and prognoses system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610493634.4A CN106127345A (en) | 2016-06-28 | 2016-06-28 | The Forecasting Methodology of a kind of mobile subscriber complaint and prognoses system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106127345A true CN106127345A (en) | 2016-11-16 |
Family
ID=57284535
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610493634.4A Pending CN106127345A (en) | 2016-06-28 | 2016-06-28 | The Forecasting Methodology of a kind of mobile subscriber complaint and prognoses system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106127345A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106910075A (en) * | 2017-01-26 | 2017-06-30 | 合肥工业大学 | Intelligent processing system and method that client mobile communication is complained |
CN109982367A (en) * | 2017-12-28 | 2019-07-05 | 中国移动通信集团四川有限公司 | Mobile terminal Internet access customer complaint prediction technique, device, equipment and storage medium |
CN117745328A (en) * | 2023-12-29 | 2024-03-22 | 深圳市南方网通网络技术开发有限公司 | Multi-platform-based network marketing data processing method and system |
-
2016
- 2016-06-28 CN CN201610493634.4A patent/CN106127345A/en active Pending
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106910075A (en) * | 2017-01-26 | 2017-06-30 | 合肥工业大学 | Intelligent processing system and method that client mobile communication is complained |
CN106910075B (en) * | 2017-01-26 | 2018-07-24 | 合肥工业大学 | The intelligent processing system and method that client mobile communication is complained |
CN109982367A (en) * | 2017-12-28 | 2019-07-05 | 中国移动通信集团四川有限公司 | Mobile terminal Internet access customer complaint prediction technique, device, equipment and storage medium |
CN109982367B (en) * | 2017-12-28 | 2022-04-29 | 中国移动通信集团四川有限公司 | Complaint prediction method, device, equipment and storage medium for internet users of mobile terminals |
CN117745328A (en) * | 2023-12-29 | 2024-03-22 | 深圳市南方网通网络技术开发有限公司 | Multi-platform-based network marketing data processing method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102158879B (en) | Essential factor lost score data processing method and equipment | |
Monroy et al. | Activity based costing, time-driven activity based costing and lean accounting: Differences among three accounting systems’ approach to manufacturing | |
Durmanov et al. | Simulation modeling, analysis and performance assessment | |
US20100280861A1 (en) | Service Level Agreement Negotiation and Associated Methods | |
CN101958917B (en) | Cloud manufacturing system-oriented method for measuring and enhancing flexibility of resource service composition | |
CN107018004A (en) | User satisfaction management system and method | |
CN110264039A (en) | A kind of generation method and equipment of working report | |
CN111626767B (en) | Resource data issuing method, device and equipment | |
CN106096810A (en) | Method and system for planning based on power distribution network operation data Yu geographical topology information | |
CN108133390A (en) | For predicting the method and apparatus of user behavior and computing device | |
CN106127345A (en) | The Forecasting Methodology of a kind of mobile subscriber complaint and prognoses system | |
WO2017071369A1 (en) | Method and device for predicting user unsubscription | |
CN104794644A (en) | Task crowdsourcing method oriented to intelligent service engine | |
CN111105129B (en) | Detection task intelligent distribution method based on detection mechanism evaluation | |
CN105787010B (en) | Acquisition process and method for pushing and system based on personal data | |
CN110717620A (en) | Order meal delivery time prediction method and device, electronic equipment and storage medium | |
CN107274001A (en) | A kind of Electricity customers based on data mining are lost in power predicating method | |
CN108805458A (en) | A kind of enterprise technology Competitiveness Assessment method and device | |
Cerit et al. | Quality function deployment and its application on a smartphone design | |
CN109754221A (en) | A kind of method and device improving power grid enterprises' audit coverage rate | |
Makhijani et al. | QUEST: queue simulation for content moderation at scale | |
Steinmann et al. | Using simulation to evaluate call forecasting algorithms for inbound call center | |
KR101927273B1 (en) | Risk value evaluating system for unclaimed construction and risk value evaluating apparatus for unclaimed construction | |
Li et al. | Learning to bundle proactively for on-demand meal delivery | |
CN114693265A (en) | Supply chain multi-user docking method and system of cloud switching platform |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20161116 |
|
RJ01 | Rejection of invention patent application after publication |