CN106204154A - User based on analytic hierarchy process (AHP) and Information Entropy is worth marking system and method thereof - Google Patents
User based on analytic hierarchy process (AHP) and Information Entropy is worth marking system and method thereof Download PDFInfo
- Publication number
- CN106204154A CN106204154A CN201610573368.6A CN201610573368A CN106204154A CN 106204154 A CN106204154 A CN 106204154A CN 201610573368 A CN201610573368 A CN 201610573368A CN 106204154 A CN106204154 A CN 106204154A
- Authority
- CN
- China
- Prior art keywords
- user
- index
- worth
- weight
- level index
- 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
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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
Abstract
The invention discloses a kind of user based on analytic hierarchy process (AHP) and Information Entropy and be worth marking system and method thereof, relate to live platform technology field, including: evaluation module, it is provided with multiple first class index, the corresponding multiple two-level index of each first index in it.Data input module, it is for the relative importance grade received between any two first class index that user inputs and obtains multiple achievement data corresponding with two-level index.First class index weight computation module, it calculates, based on analytic hierarchy process (AHP), the weight that each first class index is worth relative to user.Two-level index weight computation module, it calculates the weight of each two-level index based on Information Entropy.Total weight computation module, it is for calculating the weight that each two-level index is worth relative to user, and calculates user and be worth score.The present invention can behavior based on user and judge that user is worth.
Description
Technical field
The present invention relates to live platform technology field, be specifically related to a kind of user's valency based on analytic hierarchy process (AHP) and Information Entropy
Value marking system and method thereof.
Background technology
Along with developing rapidly of live industry, enriching of platform service, userbase presents the trend of being skyrocketed through.User exists
Produce different various actions on platform, bring different value to website, such as gifts, charging information etc., how to evaluate
The comprehensive value score of user, which is high-value user, and which is low value user, facilitates marketing personnel according to user not
With being worth grade, plan different marketing program with maintaining and keep strategy, be a problem to be solved.
Summary of the invention
For defect present in prior art, it is an object of the invention to provide a kind of can behavior based on user and sentence
The user based on analytic hierarchy process (AHP) and Information Entropy that disconnected user is worth is worth marking system.
For reaching object above, the present invention adopts the technical scheme that: a kind of based on analytic hierarchy process (AHP) with the use of Information Entropy
Family is worth marking system, including:
Evaluation module, is provided with multiple first class index, and the corresponding multiple two-level index of each first index in it;
Data input module, its relative importance grade between any two first class index receiving user's input
And obtaining user behavior information, described user behavior information includes the achievement data that multiple and described two-level index is corresponding;
First class index weight computation module, it is according to the relative importance grade between described any two first class index,
And calculate, based on analytic hierarchy process (AHP), the weight that each described first class index is worth relative to user;
Two-level index weight computation module, it calculates the weight of each described two-level index based on Information Entropy;And,
Total weight computation module, it is according to the weight of each described two-level index, and corresponding to this two-level index
The weight that level index is worth relative to user, calculates the weight that each two-level index is worth, described total weight calculation mould relative to user
Weight and described achievement data that block is worth relative to user by each two-level index calculate user and are worth score.
On the basis of technique scheme, described first class index includes user's liveness, customer consumption ability and user
Supplementing ability with money, the two-level index that described user's liveness is corresponding includes that user watches duration, user watches natural law, user watches room
Between number and user send barrage number;Two-level index corresponding to described customer consumption ability includes that user gives virtual present amount and use
Virtual present room number is given at family;Described user supplements two-level index corresponding to ability with money and includes user's recharge amount, the most each
Recharge amount and supplement natural law with money.
On the basis of technique scheme, described first class index weight computation module refers to according to described any two one-level
Relative importance grade between mark constructs the judgment matrix of first class index, and is returned by after row summation by described judgment matrix again
One change processes and obtains characteristic vector, and tries to achieve the eigenvalue of maximum of described characteristic vector, described first class index weight computation module
Whether the discordance of the eigenvalue of maximum described judgment matrix of inspection according to described characteristic vector is in permissible range, if so,
Then using described characteristic vector as weight vectors, if it is not, evaluation first class index relative importance grade between any two the most again.
Meanwhile, the present invention also provide for a kind of behavior based on user and judge that user is worth based on analytic hierarchy process (AHP)
It is worth methods of marking with the user of Information Entropy.
For reaching object above, the present invention adopts the technical scheme that: one utilizes above-mentioned user to be worth marking system to enter
Row user is worth methods of marking, comprises the following steps:
S1. data input module receive user input any two first class index between relative importance grade and
Obtaining user behavior information, wherein user behavior information includes multiple achievement data corresponding with two-level index;
S2. first class index weight computation module uses analytic hierarchy process (AHP) to calculate the power that each first class index is worth relative to user
Weight;
S3. two-level index weight computation module uses Information Entropy to calculate the weight of each two-level index;
S4. the weight of each two-level index is multiplied by the first class index corresponding with this two-level index by total weight computation module
The weight that user is worth relatively obtains the weight that each two-level index is worth relative to user;
S5. the weight being worth relative to user according to each two-level index and achievement data use weighted sum to calculate each user
It is worth score.
On the basis of technique scheme, calculate each first class index and include following step relative to the weight that user is worth
Rapid:
S21. compare two-by-two between first class index, evaluate first class index relative importance grade between any two,
And the judgment matrix A of first class index is constructed according to relative importance grade,
Wherein, aijFor judgment matrix A the i-th row jth row element, its represent first class index i and first class index j relative
Importance rate, i=1,2 ..., n, j=1,2 ..., n, and aij=1/aji;
S22. will determine that matrix A obtains matrix B=(b by row normalizationij)n×n, wherein bijFor matrix B the i-th row jth row
Element,
S23. matrix B is obtained Matrix C=(c by row summationi)n×1, wherein ciFor the element of Matrix C the i-th row,
S24. Matrix C normalization is obtained characteristic vector W=(w1,w2,...,wn)T, wherein wiIt is characterized vector W the i-th row
Element,
S25. the eigenvalue of maximum that calculating characteristic vector W is corresponding:
S26. the concordance of test and judge matrix A, weighs the deviation degree of consistency of judgment matrix A, and weighs judgement square
Whether battle array A discordance, beyond permissible range, if it is not, perform step S27, the most then returns step S21, again evaluates one-level
Index relative importance grade between any two, then performs step S22-S26, until judgment matrix A discordance is being allowed
In the range of;
S27. by characteristic vector W=(w1,w2,...wn)TAs weight vectors.
On the basis of technique scheme, the concordance of described test and judge matrix A uses coincident indicator CI weighing apparatus
The deviation degree of consistency of amount judgment matrix A:And it is inconsistent to use Consistency Ratio CR to weigh judgment matrix A
Property permissible range: CR=CI/RI, wherein RI is average homogeneity index, and it is selected with judgment matrix A with the data of valency.
On the basis of technique scheme, the weight calculating each two-level index comprises the following steps:
S31. sample the user behavior information of user, build and initialize matrix X,
Wherein, xijFor initializing the element of matrix X the i-th row jth row, the jth two-level index of its expression i-th user
Achievement data, i=1,2 ..., n, j=1,2 ..., m;
S32. the achievement data to the two-level index of each user is standardized processing,
If xjFor forward index, the achievement data of the two-level index after standardization:
If xjFor negative sense index, the achievement data of the two-level index after standardization:
Matrix after process is designated as X',
xi'jFor the element of matrix X' the i-th row jth row, it represents the jth two grades of the i-th user after standardization
The achievement data of index, i=1,2 ..., n, j=1,2 ..., m;
S33. the contribution degree of i-th user under calculating jth two-level index:Wherein i=1,2 ...,
N, j=1,2 ..., m;
S34. the entropy of calculating jth two-level index:Wherein k=1/ln (n), i=1,
2 ..., n, j=1,2 ... m;
S35. the coefficient of variation of jth two-level index: g is calculatedj=1-ej, j=1,2 ..., m;
S36. the weight of each two-level index is calculated:J=1,2 ..., m.
On the basis of technique scheme, described first class index includes user's liveness, customer consumption ability and user
Supplement ability with money.
On the basis of technique scheme, the two-level index that described user's liveness is corresponding include user watch duration,
User watches natural law, user watches room number and user sends barrage number;The two-level index bag that described customer consumption ability is corresponding
Include user and give virtual present amount and user gives virtual present room number;Described user supplements the two-level index bag that ability is corresponding with money
Include user's recharge amount, average recharge amount every time and supplement natural law with money.
On the basis of technique scheme, described aijValue have 9 kinds, it is 1/9,1/7,1/5,1/3,1/1,3/1,
5/1,7/1 or 9/1.
Compared with prior art, it is an advantage of the current invention that:
The title of the present invention includes first class index weight computation module, two-level index weight computation module and total weight calculation
Module, present invention behavior based on user, according to first class index weight computation module, two-level index weight computation module and always weigh
The value score of re-computation module energy quantitative Analysis user.Comprehensive value score according to user, divides different senior middle schools at a low price
Value grade.Such that it is able to preferably find the user of high value.Be worth grade according to different users, design different marketing and
Retention tactics, it is simple to orientation is marketed and maintains.
Accompanying drawing explanation
Fig. 1 is the structural representation that in the present invention, user based on analytic hierarchy process (AHP) and Information Entropy is worth marking system;
Fig. 2 is to carry out user in the present invention to be worth the flow chart of scoring;
Fig. 3 is the flow chart that in the present invention, first class index weight computation module calculates weight;
Fig. 4 is the flow chart that in the present invention, two-level index weight computation module calculates weight.
Detailed description of the invention
Below in conjunction with accompanying drawing, the present invention is described in further detail.
Shown in Figure 1, the present invention provides a kind of user based on analytic hierarchy process (AHP) and Information Entropy to be worth marking system, its
Including evaluation module, data input module, first class index weight computation module, two-level index weight computation module and total weight meter
Calculate module.
Evaluation module, is provided with multiple first class index and the multiple two-level index corresponding with each first class index in it.
Data input module, its relative importance grade between any two first class index receiving user's input
And obtaining user behavior information, user behavior information includes the achievement data that multiple and described two-level index is corresponding.
First class index in the present invention includes that user's liveness, customer consumption ability and user supplement ability with money.User enlivens
The two-level index that degree is corresponding includes that user watches duration, user watches natural law, user watches room number and user sends barrage number.
Two-level index corresponding to customer consumption ability includes that user gives virtual present amount and user gives virtual present room number.User
Two-level index corresponding to ability of supplementing with money includes user's recharge amount, average recharge amount every time and supplements natural law with money.
In order to determine that user is worth, the present invention constructs three-decker according to the behavioural information of user, and ground floor is user
It is worth;The second layer is first class index: user's liveness, customer consumption ability and user supplement ability with money;Third layer is two grades
Index: user watches duration, user watches natural law, user watches room number, user sends barrage number, user gives virtual present
Amount, user give virtual present room number, user's recharge amount, average recharge amount every time and supplement natural law with money.
First class index weight computation module, it is according to the relative importance grade between any two first class index, and base
The weight that each first class index is worth is calculated relative to user in analytic hierarchy process (AHP).
Two-level index weight computation module, it calculates the weight of each two-level index based on Information Entropy.Concrete, for two
For level index user watches duration, its achievement data is the data of the concrete viewing time of user.
Total weight computation module, it is according to the weight of each two-level index, and the one-level corresponding to this two-level index refers to
Marking the weight that relative user is worth, calculate the weight that each two-level index is worth relative to user, total weight computation module is by every
Weight and achievement data that individual two-level index is worth relative to user calculate user and are worth score.
In the present invention, the weight of each two-level index is multiplied by corresponding with this two-level index by total weight computation module
Level index obtains, relative to the weight that user is worth, the weight that each two-level index is worth relative to user, and according to each two-level index
The weight that user is worth relatively uses weighted sum to calculate each user and is worth score.
Specifically, the weight calculation being worth relative to user when first class index user's liveness out after, then calculate respectively
The two-level index user corresponding with user's liveness watches duration, user watches natural law, user watches room number and user sends out
Send the weight of barrage number, by the above-mentioned two kinds of multiplied by weight obtained, i.e. can get that user watches duration, user watches natural law,
User watches room number and sends the weight that barrage number is worth relative to user with user.Then it is worth relative to user according to two-level index
Weight and achievement data, use weighted sum calculate each user be worth score.It is worth score finally according to the user calculated,
Choose suitable score interval, divide different high, normal, basic value grades.
In sum, present invention behavior based on user, the value score of energy quantitative Analysis user.Comprehensive according to user
It is worth score, divides different high, normal, basic value grades.It is worth grade according to different users, designs different marketing and maintain
Strategy, it is simple to orientation is marketed and maintains.
The present invention also provides for one and utilizes above-mentioned user to be worth marking system to carry out user and be worth methods of marking, including following
Step:
S1. data input module receive user input any two first class index between relative importance grade and
Obtaining user behavior information, wherein user behavior information includes multiple achievement data corresponding with two-level index;
S2. first class index weight computation module uses analytic hierarchy process (AHP) to calculate the power that each first class index is worth relative to user
Weight;
The present invention calculates the weight that each first class index is worth relative to user comprise the following steps:
S21. compare two-by-two between first class index, evaluate first class index relative importance grade between any two,
And the judgment matrix A of first class index is constructed according to relative importance grade,
Wherein, aijFor judgment matrix A the i-th row jth row element, its represent first class index i and first class index j relative
Importance rate, i=1,2 ..., n, j=1,2 ..., n, and aij=1/aji;
A in the present inventionijThere are 9 kinds of values, respectively 1/9,1/7,1/5,1/3,1/1,3/1,5/1,7/1 and 9/1.It represents
I-th first class index is to the significance level of jth first class index from light to heavy.
Concrete, if aij=1/9, then it represents that i-th first class index is the most inessential to jth first class index, if aij=
9/1, represent that jth one-level is referred to extremely important by i-th first class index.Owing to the first class index in the present invention has 3, therefore judge
Matrix A is third-order matrix.
S22. will determine that matrix A obtains matrix B=(b by row normalizationij)n×n, wherein bijFor matrix B the i-th row jth row
Element,
S23. matrix B is obtained Matrix C=(c by row summationi)n×1, wherein ciFor the element of Matrix C the i-th row,
S24. Matrix C normalization is obtained characteristic vector W=(w1,w2,...,wn)T, wherein wiIt is characterized vector W the i-th row
Element,
S25. the eigenvalue of maximum that calculating characteristic vector W is corresponding:
S26. the concordance of test and judge matrix A, weighs the deviation degree of consistency of judgment matrix A, and weighs judgement square
Whether battle array A discordance, beyond permissible range, if it is not, perform step S27, the most then returns step S21, again evaluates one-level
Index relative importance grade between any two, then performs step S22-S26, until judgment matrix A discordance is being allowed
In the range of.
The present invention use coincident indicator CI weigh the deviation one of judgment matrix A in the concordance of test and judge matrix A
Cause property degree:If CI value is the biggest, show that the degree of judgment matrix A deviation crash consistency is the biggest;If CI value
The least, show that the degree of judgment matrix A deviation crash consistency is the least, i.e. judgment matrix A concordance is the best.
The present invention uses Consistency Ratio CR to weigh the permissible range of judgment matrix A discordance: CR=CI/RI, wherein
RI is average homogeneity index, and it is selected with judgment matrix A with the data of valency.The present invention is third moment due to judgment matrix A
Battle array, therefore RI also chooses the value on three rank, RI is shown in Table 1 with the relation ginseng of exponent number.Permissible range in the present invention is 0.1, i.e.
During the CR < 0.1 of judgment matrix A, its inconsistent type degree is in permissible range, otherwise, then needs again to evaluate aijValue, directly
To the discordance degree of judgment matrix A in permissible range.
The value of table 1 average homogeneity index RI and the relation of exponent number
n | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 |
RI | 0 | 0 | 0.58 | 0.9 | 1.12 | 1.24 | 1.32 | 1.41 | 1.45 | 1.49 | 1.51 |
S27. by characteristic vector W=(w1,w2,...wn)TAs weight vectors;
S3. two-level index weight computation module uses Information Entropy to calculate the weight of each two-level index;
The weight calculating each two-level index in the present invention comprises the following steps:
S31. sample the user behavior information of user, build and initialize matrix X,
Wherein, xijFor initializing the element of matrix X the i-th row jth row, the jth two-level index of its expression i-th user
Achievement data, i=1,2 ..., n, j=1,2 ..., m;
Two-level index in the present invention includes that user watches duration, user watches natural law, user watches room number, Yong Hufa
Send barrage number, user to give virtual present amount, user gives virtual present room number, user's recharge amount, averagely supplement with money every time
The amount of money and supplement natural law with money.One has 9 two-level index, so place m=9.
S32. the achievement data to the two-level index of each user is standardized processing,
If xjFor forward index, the achievement data of the two-level index after standardization:
If xjFor negative sense index, the achievement data of the two-level index after standardization:
Matrix after process is designated as X',
x′ijFor the element of matrix X' the i-th row jth row, it represents the jth two grades of the i-th user after standardization
The achievement data of index, i=1,2 ..., n, j=1,2 ..., m;
Owing to the dimension of each two-level index is inconsistent, therefore before calculating user and being worth score, two grades chosen are referred to
Mark is standardized processing.Additionally, due to forward index is different with the implication that negative sense index value represents: forward index value is more
Height is the best, and negative sense index value is the lowest more good, therefore, during standardized index, needs to refer to different algorithms and is standardized.
S33. the contribution degree of i-th user under calculating jth two-level index:
S34. the entropy of calculating jth two-level index:Wherein k=1/ln (n), i=1,
2 ..., n, j=1,2 ..., m;
S35. the coefficient of variation of jth two-level index: g is calculatedj=1-ej;
For jth two-level index, when its entropy is the least, then coefficient of variation gjThe biggest, and gjThe biggest, represent these two grades
Index is the most important.
S36. the weight of each two-level index is calculated:
S4. the weight of each two-level index is multiplied by the first class index corresponding with this two-level index by total weight computation module
The weight that user is worth relatively obtains the weight that each two-level index is worth relative to user.
S5. the weight being worth relative to user according to each two-level index and achievement data use weighted sum to calculate each user
It is worth score.
In the present invention, first class index includes that user's liveness, customer consumption ability and user supplement ability with money.User's liveness
Corresponding two-level index includes that user watches duration, user watches natural law, user watches room number and user sends barrage number.With
The two-level index that the family consuming capacity are corresponding includes that user gives virtual present amount and user gives virtual present room number.User fills
Two-level index corresponding to value ability includes user's recharge amount, average recharge amount every time and supplements natural law with money.
The weight calculation being worth relative to user when first class index user's liveness out after, then calculate respectively and enliven with user
Spend that corresponding two-level index user watches duration, user watches natural law, user watches room number and user sends barrage number
Weight, by the above-mentioned two kinds of multiplied by weight obtained, i.e. can get that user watches duration, user watches natural law, user watches room
Between count and send the weight that barrage number is worth relative to user with user.The weight that two-level index is worth relative to user is designated as qj, use x'j
Represent the two-level index after standardization, but the achievement data of two-level index simply handled it by two standardizations,
x'jSubstantially and xjIdentical.
Then the weight being worth relative to user according to two-level index and the achievement data corresponding with two-level index, employing adds
Weigh and calculate each user and be worth score.Its concrete calculation sees below formula:
Si=q1*x′i1+q2*x′i2+…+qj*x′ij, j=1,2 ..., m;
It is worth score finally according to the user calculated, chooses suitable score interval, divide different high, normal, basic value etc.
Level.
The present invention is not limited to above-mentioned embodiment, for those skilled in the art, without departing from
On the premise of the principle of the invention, it is also possible to make some improvements and modifications, these improvements and modifications are also considered as the protection of the present invention
Within the scope of.The content not being described in detail in this specification belongs to prior art known to professional and technical personnel in the field.
Claims (10)
1. a user based on analytic hierarchy process (AHP) and Information Entropy is worth marking system, it is characterised in that including:
Evaluation module, is provided with multiple first class index, and the corresponding multiple two-level index of each first index in it;
Data input module, its for receive user input any two first class index between relative importance grade and
Obtaining user behavior information, described user behavior information includes the achievement data that multiple and described two-level index is corresponding;
First class index weight computation module, it is according to the relative importance grade between described any two first class index, and base
The weight that each described first class index is worth is calculated relative to user in analytic hierarchy process (AHP);
Two-level index weight computation module, it calculates the weight of each described two-level index based on Information Entropy;And,
Total weight computation module, it is according to the weight of each described two-level index, and the one-level corresponding to this two-level index refers to
Marking the weight that relative user is worth, calculate the weight that each two-level index is worth relative to user, described total weight computation module is led to
Cross weight that each two-level index is worth relative to user and described achievement data calculates user and is worth score.
2. user based on analytic hierarchy process (AHP) and Information Entropy as claimed in claim 1 is worth marking system, it is characterised in that: institute
State first class index and include that user's liveness, customer consumption ability and user supplement ability, two grades that described user's liveness is corresponding with money
Index includes that user watches duration, user watches natural law, user watches room number and user sends barrage number;Described customer consumption
Two-level index corresponding to ability includes that user gives virtual present amount and user gives virtual present room number;Described user supplements with money
Two-level index corresponding to ability includes user's recharge amount, average recharge amount every time and supplements natural law with money.
3. user based on analytic hierarchy process (AHP) and Information Entropy as claimed in claim 1 is worth marking system, it is characterised in that: institute
State first class index weight computation module and construct one-level according to the relative importance grade between described any two first class index
The judgment matrix of index, and described judgment matrix is obtained characteristic vector by renormalization process after row summation, and try to achieve described
The eigenvalue of maximum of characteristic vector, described first class index weight computation module is checked according to the eigenvalue of maximum of described characteristic vector
Whether the discordance of described judgment matrix in permissible range, the most then using described characteristic vector as weight vectors, if
No, evaluation first class index relative importance grade between any two the most again.
4. one kind utilizes the user described in claim 1 to be worth marking system to carry out user and be worth methods of marking, it is characterised in that
Comprise the following steps:
S1. data input module receives the relative importance grade between any two first class index of user's input and acquisition
User behavior information, wherein user behavior information includes multiple achievement data corresponding with two-level index;
S2. first class index weight computation module uses analytic hierarchy process (AHP) to calculate the weight that each first class index is worth relative to user;
S3. two-level index weight computation module uses Information Entropy to calculate the weight of each two-level index;
S4. that the weight of each two-level index is multiplied by the first class index corresponding with this two-level index is relative for total weight computation module
The weight that user is worth obtains the weight that each two-level index is worth relative to user;
S5. the weight being worth relative to user according to each two-level index and achievement data use weighted sum to calculate each user and are worth
Score.
5. user based on analytic hierarchy process (AHP) and Information Entropy as claimed in claim 4 is worth methods of marking, it is characterised in that meter
Calculate the weight that each first class index is worth relative to user to comprise the following steps:
S21. compare two-by-two between first class index, evaluate first class index relative importance grade between any two, and root
The judgment matrix A of first class index is constructed according to relative importance grade,
Wherein, aijFor judgment matrix A the i-th row jth row element, its represent first class index i and first class index j relatively important
Property grade, i=1,2 ..., n, j=1,2 ..., n, and aij=1/aji;
S22. will determine that matrix A obtains matrix B=(b by row normalizationij)n×n, wherein bijUnit for matrix B the i-th row jth row
Element,I=1,2 ..., n, j=1,2 ..., n;
S23. matrix B is obtained Matrix C=(c by row summationi)n×1, wherein ciFor the element of Matrix C the i-th row,I=
1,2,...,n;
S24. Matrix C normalization is obtained characteristic vector W=(w1,w2,...,wn)T, wherein wiIt is characterized the unit of vector W the i-th row
Element,I=1,2 ..., n;
S25. the eigenvalue of maximum that calculating characteristic vector W is corresponding:
S26. the concordance of test and judge matrix A, weighs the deviation degree of consistency of judgment matrix A, and weighs judgment matrix A not
Concordance, whether beyond permissible range, if it is not, perform step S27, the most then returns step S21, again evaluation first class index two
Relative importance grade between two, then performs step S22-S26, until judgment matrix A discordance is in permissible range;
S27. by characteristic vector W=(w1,w2,...wn)TAs weight vectors.
6. user based on analytic hierarchy process (AHP) and Information Entropy as claimed in claim 5 is worth methods of marking, it is characterised in that: institute
State in the concordance of test and judge matrix A use coincident indicator CI weigh judgment matrix A the deviation degree of consistency:And use Consistency Ratio CR to weigh the permissible range of judgment matrix A discordance: CR=CI/RI, wherein
RI is average homogeneity index, and it is selected with judgment matrix A with the data of valency.
7. user based on analytic hierarchy process (AHP) and Information Entropy as claimed in claim 4 is worth methods of marking, it is characterised in that meter
The weight calculating each two-level index comprises the following steps:
S31. sample the user behavior information of user, build and initialize matrix X,
Wherein, xijFor initializing the element of matrix X the i-th row jth row, the index of the jth two-level index of its expression i-th user
Data, i=1,2 ..., n, j=1,2 ..., m;
S32. the achievement data to the two-level index of each user is standardized processing,
If xjFor forward index, the achievement data of the two-level index after standardization:
If xjFor negative sense index, the achievement data of the two-level index after standardization:
Matrix after process is designated as X',
x′ijFor the element of matrix X' the i-th row jth row, it represents the jth two-level index of the i-th user after standardization
Achievement data, i=1,2 ..., n, j=1,2 ..., m;
S33. the contribution degree of i-th user under calculating jth two-level index:Wherein i=1,2 ..., n, j=
1,2,...,m;
S34. the entropy of calculating jth two-level index:Wherein k=1/ln (n), i=1,2 ...,
N, j=1,2 ... m;
S35. the coefficient of variation of jth two-level index: g is calculatedj=1-ej, j=1,2 ..., m;
S36. the weight of each two-level index is calculated:J=1,2 ..., m.
8. user based on analytic hierarchy process (AHP) and Information Entropy as claimed in claim 5 is worth methods of marking, it is characterised in that: institute
State first class index and include that user's liveness, customer consumption ability and user supplement ability with money.
9. user based on analytic hierarchy process (AHP) and Information Entropy as claimed in claim 8 is worth methods of marking, it is characterised in that: institute
State two-level index corresponding to user's liveness and include that user watches duration, user watches natural law, user watches room number and user
Send barrage number;Two-level index corresponding to described customer consumption ability includes that user gives virtual present amount and user gives virtual
Present room number;Described user supplement with money two-level index corresponding to ability include user's recharge amount, average recharge amount every time and
Supplement natural law with money.
10. user based on analytic hierarchy process (AHP) and Information Entropy as claimed in claim 4 is worth methods of marking, it is characterised in that:
Described aijValue have 9 kinds, it is 1/9,1/7,1/5,1/3,1/1,3/1,5/1,7/1 or 9/1.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610573368.6A CN106204154A (en) | 2016-07-20 | 2016-07-20 | User based on analytic hierarchy process (AHP) and Information Entropy is worth marking system and method thereof |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610573368.6A CN106204154A (en) | 2016-07-20 | 2016-07-20 | User based on analytic hierarchy process (AHP) and Information Entropy is worth marking system and method thereof |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106204154A true CN106204154A (en) | 2016-12-07 |
Family
ID=57494533
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610573368.6A Pending CN106204154A (en) | 2016-07-20 | 2016-07-20 | User based on analytic hierarchy process (AHP) and Information Entropy is worth marking system and method thereof |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106204154A (en) |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106850613A (en) * | 2017-01-24 | 2017-06-13 | 中国科学院信息工程研究所 | A kind of user behavior method for evaluating trust and system based on advanced AHP |
CN107153677A (en) * | 2017-04-18 | 2017-09-12 | 北京思特奇信息技术股份有限公司 | A kind of data processing method and system for searching value user |
CN107784519A (en) * | 2017-10-13 | 2018-03-09 | 厦门集微科技有限公司 | A kind of advertising results appraisal procedure and server |
CN107871211A (en) * | 2017-11-06 | 2018-04-03 | 网易乐得科技有限公司 | Generation method, medium, device and the computing device of user's growth curve |
CN107948742A (en) * | 2017-11-10 | 2018-04-20 | 北京潘达互娱科技有限公司 | A kind of any active ues lookup method and device |
CN108833143A (en) * | 2018-05-28 | 2018-11-16 | 郑州村村联网络技术有限公司 | Calculate the method, apparatus and system of user activity |
CN108960562A (en) * | 2018-05-18 | 2018-12-07 | 深圳市和讯华谷信息技术有限公司 | A kind of regional influence appraisal procedure and device |
CN109255724A (en) * | 2018-09-13 | 2019-01-22 | 北京工商大学 | A kind of pesticide residual contamination evaluation method based on multiple-factor and AHP-E model |
CN109345118A (en) * | 2018-09-30 | 2019-02-15 | 浙江大学华南工业技术研究院 | Measurement method, device, the computer equipment of user experience quantitative evaluation value |
CN109460926A (en) * | 2018-11-14 | 2019-03-12 | 江苏电力信息技术有限公司 | Platform area group of assets comprehensive performance evaluation method based on analytic hierarchy process (AHP) and Information Entropy |
CN109492890A (en) * | 2018-10-26 | 2019-03-19 | 浙江大学华南工业技术研究院 | Measurement method, device, the computer equipment of user experience quantitative evaluation value |
CN109598421A (en) * | 2018-11-20 | 2019-04-09 | 国家电网有限公司 | Charging station operation data processing method and processing device |
CN109740967A (en) * | 2019-01-23 | 2019-05-10 | 广州虎牙信息科技有限公司 | A kind of method for processing business, device, computer equipment and storage medium |
CN110390587A (en) * | 2018-04-18 | 2019-10-29 | 苏宁易购集团股份有限公司 | A kind of customer evaluation method and system |
CN111078945A (en) * | 2019-12-19 | 2020-04-28 | 北京奇艺世纪科技有限公司 | Video pushing method and device |
CN111768873A (en) * | 2020-06-03 | 2020-10-13 | 中国地质大学(武汉) | COVID-19 real-time risk prediction method |
CN112464166A (en) * | 2020-11-26 | 2021-03-09 | 上海移卓网络科技有限公司 | Channel popularization evaluation method and device, computer equipment and storage medium |
CN112734233A (en) * | 2021-01-08 | 2021-04-30 | 上海移卓网络科技有限公司 | Method and device for confirming quality of newly added customer of APP (application) promotion channel |
CN113312559A (en) * | 2021-06-07 | 2021-08-27 | 北京百度网讯科技有限公司 | Contribution information determining method and device and electronic equipment |
CN113688120A (en) * | 2021-08-30 | 2021-11-23 | 维沃移动通信有限公司 | Quality detection method and device for data warehouse and electronic equipment |
-
2016
- 2016-07-20 CN CN201610573368.6A patent/CN106204154A/en active Pending
Cited By (23)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106850613A (en) * | 2017-01-24 | 2017-06-13 | 中国科学院信息工程研究所 | A kind of user behavior method for evaluating trust and system based on advanced AHP |
CN106850613B (en) * | 2017-01-24 | 2019-10-25 | 中国科学院信息工程研究所 | A kind of user behavior method for evaluating trust and system based on advanced AHP |
CN107153677A (en) * | 2017-04-18 | 2017-09-12 | 北京思特奇信息技术股份有限公司 | A kind of data processing method and system for searching value user |
CN107153677B (en) * | 2017-04-18 | 2020-11-24 | 北京思特奇信息技术股份有限公司 | Data processing method and system for searching value user |
CN107784519A (en) * | 2017-10-13 | 2018-03-09 | 厦门集微科技有限公司 | A kind of advertising results appraisal procedure and server |
CN107871211A (en) * | 2017-11-06 | 2018-04-03 | 网易乐得科技有限公司 | Generation method, medium, device and the computing device of user's growth curve |
CN107948742A (en) * | 2017-11-10 | 2018-04-20 | 北京潘达互娱科技有限公司 | A kind of any active ues lookup method and device |
CN110390587A (en) * | 2018-04-18 | 2019-10-29 | 苏宁易购集团股份有限公司 | A kind of customer evaluation method and system |
CN108960562A (en) * | 2018-05-18 | 2018-12-07 | 深圳市和讯华谷信息技术有限公司 | A kind of regional influence appraisal procedure and device |
CN108833143A (en) * | 2018-05-28 | 2018-11-16 | 郑州村村联网络技术有限公司 | Calculate the method, apparatus and system of user activity |
CN109255724B (en) * | 2018-09-13 | 2022-02-11 | 北京工商大学 | Pesticide residue pollution evaluation method based on multifactor and AHP-E model |
CN109255724A (en) * | 2018-09-13 | 2019-01-22 | 北京工商大学 | A kind of pesticide residual contamination evaluation method based on multiple-factor and AHP-E model |
CN109345118A (en) * | 2018-09-30 | 2019-02-15 | 浙江大学华南工业技术研究院 | Measurement method, device, the computer equipment of user experience quantitative evaluation value |
CN109492890A (en) * | 2018-10-26 | 2019-03-19 | 浙江大学华南工业技术研究院 | Measurement method, device, the computer equipment of user experience quantitative evaluation value |
CN109460926A (en) * | 2018-11-14 | 2019-03-12 | 江苏电力信息技术有限公司 | Platform area group of assets comprehensive performance evaluation method based on analytic hierarchy process (AHP) and Information Entropy |
CN109598421A (en) * | 2018-11-20 | 2019-04-09 | 国家电网有限公司 | Charging station operation data processing method and processing device |
CN109740967A (en) * | 2019-01-23 | 2019-05-10 | 广州虎牙信息科技有限公司 | A kind of method for processing business, device, computer equipment and storage medium |
CN111078945A (en) * | 2019-12-19 | 2020-04-28 | 北京奇艺世纪科技有限公司 | Video pushing method and device |
CN111768873A (en) * | 2020-06-03 | 2020-10-13 | 中国地质大学(武汉) | COVID-19 real-time risk prediction method |
CN112464166A (en) * | 2020-11-26 | 2021-03-09 | 上海移卓网络科技有限公司 | Channel popularization evaluation method and device, computer equipment and storage medium |
CN112734233A (en) * | 2021-01-08 | 2021-04-30 | 上海移卓网络科技有限公司 | Method and device for confirming quality of newly added customer of APP (application) promotion channel |
CN113312559A (en) * | 2021-06-07 | 2021-08-27 | 北京百度网讯科技有限公司 | Contribution information determining method and device and electronic equipment |
CN113688120A (en) * | 2021-08-30 | 2021-11-23 | 维沃移动通信有限公司 | Quality detection method and device for data warehouse and electronic equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106204154A (en) | User based on analytic hierarchy process (AHP) and Information Entropy is worth marking system and method thereof | |
CN106228403A (en) | A kind of user based on step analysis algorithm is worth methods of marking and system | |
CN103853786B (en) | The optimization method and system of database parameter | |
CN107576920A (en) | A kind of satellite power supply health degree computational methods based on fuzzy multi-attribute evaluation | |
CN106228286A (en) | A kind of data analysing method for the assessment of artificial customer service work quality | |
Hosseinian et al. | A new linear programming method for weights generation and group decision making in the analytic hierarchy process | |
CN102902882A (en) | Method for evaluating operation quality of information systems | |
CN105160149B (en) | A kind of demand response scheduling evaluation system construction method for simulating regulating units | |
CN109191189A (en) | Power sales decontrol lower power customer value assessment method | |
CN104021300B (en) | Comprehensive assessment method based on effect of distribution type electrical connection on power distribution network | |
CN106780127A (en) | Evaluation method containing distributed photovoltaic power distribution network | |
CN106952052A (en) | Based on hybrid weight core principle component analysis enterprise supplier evaluation method | |
CN106548272A (en) | A kind of electric automobile fills the evaluation methodology of facility combination property soon | |
CN105956768A (en) | Power generation enterprise competitiveness evaluation method based on combined weight determining and improved TOPSIS | |
CN105007170A (en) | WLAN load comprehensive evaluation method based on FAHP-SVM theory | |
CN106548413A (en) | A kind of power system energy storage fitness-for-service assessment method and system | |
CN108205783A (en) | A kind of automation credit scoring system in agricultural credit field | |
CN104331613B (en) | The evaluation method of the communication equipment antijamming capability of multiple types | |
CN109993463A (en) | A kind of engineering quality control evaluation method of pipe gallery | |
Kurniawati et al. | Implementation of AHP and SAW methods for optimization of decision recommendations | |
CN111882198A (en) | Project performance evaluation method and system | |
CN106960288A (en) | A kind of electric energy substitutes evaluation method and system | |
CN109460926A (en) | Platform area group of assets comprehensive performance evaluation method based on analytic hierarchy process (AHP) and Information Entropy | |
CN106056274A (en) | Power construction main body benefit analysis method based on PCA-DEA two-dimensional comprehensive evaluation model | |
CN104199744B (en) | A kind of supercomputer application performance stability judging method and device |
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: 20161207 |
|
RJ01 | Rejection of invention patent application after publication |