CN106372775A - Assessment method and system of comprehensive value of power grid client - Google Patents
Assessment method and system of comprehensive value of power grid client Download PDFInfo
- Publication number
- CN106372775A CN106372775A CN201610696848.1A CN201610696848A CN106372775A CN 106372775 A CN106372775 A CN 106372775A CN 201610696848 A CN201610696848 A CN 201610696848A CN 106372775 A CN106372775 A CN 106372775A
- Authority
- CN
- China
- Prior art keywords
- index
- value
- electricity
- client
- level
- 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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- 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
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The invention provides an assessment method and system of the comprehensive value of a power grid client. The method comprises the following steps of: dividing power grid client indexes to obtain first-class indexes, second-level indexes and third-level indexes; screening related true and effective fields from a database, and determining n basic indexes in the three-level indexes; adopting a hierarchical clustering method and a Pareto law to carry out level judgment on the n basic indexes; on the basis of a Delphi expert grading method, obtaining a collection index consultation score graph and a client total score; and according to a multiple linear regression method, analyzing and determining a basic index weight. The power grid client indexes are divided, the hierarchical clustering method and the Pareto law are adopted to carry out level judgment on the n basic indexes, the Delphi expert grading method is combined to obtain the collection index consultation score graph and the client total score, the basic index weight is analyzed and determined according to the multiple linear regression method, the comprehensive value of the power grid client is assessed from multiple aspects in an omnibearing way, and a power enterprise is effectively assisted in appointing a personalized power utilization strategy to clients.
Description
Technical field
The present invention relates to electrical network customer analysis technical field, particularly relate to a kind of electrical network customer general value appraisal procedure and
System.
Background technology
In recent years, State Council in 2002 formally issues " power system reform scheme ", and the power industry marketization of China changes
Leather formally raises the curtain.The monopolistic general layout of conventional vertical integration is broken, and sends out and is completely separated for main body.Study and set up
Science, objective operation of power networks state index system, to propulsion electrical network scientific development, improve administration of power networks level and have important meaning
Justice.
South electric network propose integrated operation of power networks intelligence system (operation smart system, oss) and its
Operation of Electric Systems driver's cabin module (power system operation cockpit, poc) is that operation of power networks state is referred to
The important application of mark system, its objective is to analyse in depth operation of power networks feature, true reflection operation of power networks state, and then improves fortune
The manipulation ability to electrical network for the administrative staff.Operation of power networks state index is large number of, how to set up rational index system, distribution phase
Close index weights, be the major issue perplexing research worker always.
Content of the invention
The technical problem to be solved in the present invention is to provide a kind of electrical network customer general value appraisal procedure and system, Neng Gouquan
Orientation multi-angle assesses the comprehensive value of electrical network client.
For solving above-mentioned technical problem, embodiments of the invention provide a kind of electrical network customer general value appraisal procedure, institute
State electrical network customer general value appraisal procedure to include:
Electrical network client's index is divided, obtains first class index, two-level index and three-level index, described include working as present value
Value and potential value, described current value includes revenue contributions, stability, social benefit and operation benefits, described potential value
Including loyalty, exhibition prospect, credit rating and policy guidance, described revenue contributions, stability, social benefit and operation benefits are two
Level index；
Screen related authentic and valid field to be mated from data base, determine the n item base values in three-level index；
Using hierarchy clustering method and Pareto Empiric Rule to n item base values evaluation rank；
Obtain collection index in conjunction with Delphi method expert point rating method and seek the opinion of score graph and client's total score；
Analyzed according to multiple linear regression analysis method and determine base values weight.
Preferably, described revenue contributions include purchase of electricity, electricity charge area contribution rate, enterprise's purchase of electricity contribution degree ratio contribution
Degree, business electrical average price, execution electricity price standard, account for the indexs such as the ratio of field capacity weigh；Stability include valley power consumption rate,
The indexs such as Peak power use rate, electric capacity change electricity consumption situation, electric capacity change cycle, purchase of electricity rate of increase, purchase of electricity increment are weighed；
Social benefit include whether for highly energy-consuming, be whether the forceful electric power requirement such as military project medical treatment enterprise, load significance level, electric pressure,
Scale of consumer, plant-grid connection situation, whether thousand grades of high pressure clients, whether participate in protecting electric task and weigh；Operation benefits include telegram in reply,
The quantifier number of stoppages, client's load type, double multi-source client, height jeopardize Very Important Person grade, change electricity consumption situation, appearance
Quantitative change more frequency, proof cycle are weighed；Whether loyalty includes data completeness, is tripartite client, accumulative tariff recovery rate, basis
Phase tariff recovery rate, tariff recovery are spent on time, tariff recovery speed, repay electricity charge ability, payment duration, price sensitivity, visitor
Family interbehavior, electricity consumption duration are weighed；Development prospect includes electricity charge troughput, electricity charge rate of increase, business impact degree, Industrial Cycle
Degree, place industrial advantages, enterprise management level, enterprise's place life cycle are weighed；Credit rating includes history arrearage volume, adds up to owe
Guarantee level, arrearage power-off condition, the situation that collects charges for electricity in advance, metering fault letter are repaid in expense time percentage, average arrearage number of times, the electricity charge
Breath, preventive trial information, potential safety hazard information, transgression for using electricity situation, violation number of times of checking meter, the quantifier number of stoppages are weighed；
Policy guidance includes whether that listing policy in limits, whether lists policy encouragement measurement in；
Described purchase of electricity, electricity charge area contribution rate, enterprise's purchase of electricity contribution degree ratio contribution degree, business electrical average price, hold
Row electricity price standard, account for the measurement of the indexs such as the ratio of field capacity, valley power consumption rate, Peak power use rate, electric capacity change electricity consumption situation,
Electric capacity change cycle, purchase of electricity rate of increase, purchase of electricity increment, whether be highly energy-consuming, whether be the forceful electric power requirement such as military project medical treatment
Enterprise, load significance level, electric pressure, scale of consumer, plant-grid connection situation, whether thousand grades of high pressure clients, whether participate in protecting
Electric task measurement, telegram in reply, the quantifier number of stoppages, client's load type, double multi-source client, height jeopardize Very Important Person etc.
Whether level, change electricity consumption situation, capacity change frequency, proof cycle measurement, data completeness, are tripartite client, the accumulative electricity charge
The response rate, current period tariff recovery rate, tariff recovery are spent on time, tariff recovery speed, repayment electricity charge ability, payment duration, price
Sensitivity, customer interaction behavior, the measurement of electricity consumption duration, electricity charge troughput, electricity charge rate of increase, business impact degree, Industrial Cycle degree,
Place industrial advantages, enterprise management level, enterprise's place life cycle measurement, history arrearage volume, accumulative arrearage time percentage, flat
All guarantee level, arrearage power-off condition, the situation that collects charges for electricity in advance, metering fault information, preventive trial are repaid in arrearage number of times, the electricity charge
Information, potential safety hazard information, transgression for using electricity situation, violation number of times of checking meter, the quantifier number of stoppages are weighed, whether are listed policy in
Limiting, whether listing policy in and encourage to weigh is three-level index.
Preferably, described screen related authentic and valid field from data base and mated, determine the n item in three-level index
Base values, comprising:
Will be corresponding with Database field for the three-level index determining；
The index of field vacancy, error in data is deleted from index storehouse, runs into some index time cycle inapplicable feelings
Condition, is supplied with average；
Three-level index after Database field is corresponding is updated to n item base values.
Preferably, the method for described employing hierarchical clustering and Pareto Empiric Rule are to n item base values evaluation rank, comprising:
To electric power data normalization pretreatment；
Determine k value scope, and calculate the silhouette coefficient of k value；
Extract corresponding k value and obtain cluster analysis result, determine threshold value and the score value of divided rank according to the numerical value of different clusters.
Preferably, described combination Delphi method expert point rating method obtain collect index seek the opinion of score graph and client's total score, bag
Include:
With n item base values for affecting the factor design value analyses object consultation table of bond；
Carry out fraction integration with the index score that addition evaluation type Delphi method expert point rating method is chosen to expert, will count
Result feeds back to expert；
The fraction of each index according to expert amendment and object total score are seeked the opinion of with reference to anonymity and suggestion feedback, obtain index
Final score value and total score, obtain index and seek the opinion of score graph.
Preferably, described analysis according to multiple linear regression analysis method determines base values weight, comprising:
With the corresponding score value of n item base values evaluation rank as independent variable, seek the opinion of the client in score graph to collect index
Total score is dependent variable, carries out multiple linear regression analysis method using matlab or r language tool, determines that each base values is weighed
Weight.
The present invention also provides a kind of electrical network customer general value assessment system, described electrical network customer general value assessment system
Including:
Client's index division module, for dividing to electrical network client's index, obtains first class index, two-level index and three
Level index, described inclusion current value and potential value, described current value include revenue contributions, stability, social benefit and
Operation benefits, described potential value includes loyalty, exhibition prospect, credit rating and policy guidance, described revenue contributions, stability,
Social benefit and operation benefits are two-level index, and described revenue contributions, stability, social benefit and operation benefits refer to for two grades
Mark；
Database matching module, is mated for screening related authentic and valid field from data base, is determined that three-level refers to
N item base values in mark；
Level evaluation module, for using hierarchy clustering method and Pareto Empiric Rule to n item base values evaluation rank；
Grading module, seeks the opinion of score graph and client's total score for obtaining collection index with reference to Delphi method expert point rating method；
Weight determination module, determines base values weight for analyzing according to multiple linear regression analysis method.
Preferably, described level evaluation module includes:
Normalization unit, for electric power data normalization pretreatment；
Coefficient calculation unit, for determining k value scope, and calculates the silhouette coefficient of k value；
Grade threshold unit, obtains cluster analysis result for extracting corresponding k value, determines divided rank according to the numerical value of different clusters
Threshold value and scoring.
Preferably, institute's scoring module includes:
Opinionaire design cell, for being seeked the opinion of with n item base values for the factor design value analyses object affecting bond
Opinionaire；
Fraction integral unit, for being entered to the index score that expert chooses with addition evaluation type Delphi method expert point rating method
Row fraction is integrated, and statistical result is fed back to expert；
Fraction determining unit, the fraction for each index according to expert amendment and object total score seek the opinion of with reference to anonymity and
Suggestion feedback, obtains the final score value of index and total score, obtains index and seek the opinion of score graph.
Preferably, described weight determination module includes:
Weight determining unit, for the corresponding score value of n item base values evaluation rank as independent variable, being levied with collecting index
The client's total score ask in score graph is dependent variable, carries out multiple linear regression using matlab or r language tool, determines each
Base values weight.
The having the beneficial effect that of the technique scheme of the present invention:
In such scheme, by dividing to electrical network client's index, using hierarchy clustering method and Pareto Empiric Rule to n
Item base values evaluation rank, obtains collection index in conjunction with Delphi method expert point rating method and seeks the opinion of score graph and client's total score, and
Analyzed according to PCA and determine base values weight, all-dimensional multi-angle have evaluated the comprehensive value of electrical network client, has
Effect auxiliary power business to customer specifies personalization electricity consumption strategy.
Brief description
Fig. 1 is the electrical network customer general value appraisal procedure flow chart of steps of the embodiment of the present invention；
Fig. 2 is that the electrical network customer general value assessment system structure of the embodiment of the present invention connects block diagram.
Specific embodiment
For making the technical problem to be solved in the present invention, technical scheme and advantage clearer, below in conjunction with accompanying drawing and tool
Body embodiment is described in detail.
As shown in figure 1, a kind of electrical network customer general value appraisal procedure of the embodiment of the present invention, described electrical network client comprehensive
Valuation Method includes:
Step 101: electrical network client's index is divided, obtains first class index, two-level index and three-level index, described bag
Include current value and potential value, described current value includes revenue contributions, stability, social benefit and operation benefits, described
Potential value includes loyalty, exhibition prospect, credit rating and policy guidance, described revenue contributions, stability, social benefit and operation
Benefit is two-level index；
Wherein, after the index item of three-level index designs, can be according to the featured configuration of the different indexs different time
Dimension.The index of the quantizations such as such as business electrical amount, violation number of times of checking meter, purchase of electricity increment, this method focuses on to consider index
Total amount, the with due regard to point metering of different periods, we have parameter over year/season/moon respectively；And purchase of electricity rate of increase,
The index that current period tariff recovery rate, Peak power use rate etc. are related to multiplying power is more focused on periodically, time dimension with greater need for subdivision, I
Mainly have week/moon/season parameter.Such design had both considered the time dimension difference that different indexs are focused on, and rich
The rich polymorphism of index, makes index system more objective rationally, the later stage does when grade classification and weight analysis also more added with evidence
Can follow.
Step 102: screen related authentic and valid field from data base and mated, determine the n item base in three-level index
Plinth index；
Step 103: using hierarchy clustering method and Pareto Empiric Rule to n item base values evaluation rank；
Step 104: obtain collection index with reference to Delphi method expert point rating method and seek the opinion of score graph and client's total score；
Step 105: analyzed according to multiple linear regression analysis method and determine base values weight.
Preferably, described revenue contributions include purchase of electricity, electricity charge area contribution rate, enterprise's purchase of electricity contribution degree ratio contribution
Degree, business electrical average price, execution electricity price standard, account for the indexs such as the ratio of field capacity weigh；Stability include valley power consumption rate,
The indexs such as Peak power use rate, electric capacity change electricity consumption situation, electric capacity change cycle, purchase of electricity rate of increase, purchase of electricity increment are weighed；
Social benefit include whether for highly energy-consuming, be whether the forceful electric power requirement such as military project medical treatment enterprise, load significance level, electric pressure,
Scale of consumer, plant-grid connection situation, whether thousand grades of high pressure clients, whether participate in protecting electric task and weigh；Operation benefits include telegram in reply,
The quantifier number of stoppages, client's load type, double multi-source client, height jeopardize Very Important Person grade, change electricity consumption situation, appearance
Quantitative change more frequency, proof cycle are weighed；Whether loyalty includes data completeness, is tripartite client, accumulative tariff recovery rate, basis
Phase tariff recovery rate, tariff recovery are spent on time, tariff recovery speed, repay electricity charge ability, payment duration, price sensitivity, visitor
Family interbehavior, electricity consumption duration are weighed；Development prospect includes electricity charge troughput, electricity charge rate of increase, business impact degree, Industrial Cycle
Degree, place industrial advantages, enterprise management level, enterprise's place life cycle are weighed；Credit rating includes history arrearage volume, adds up to owe
Guarantee level, arrearage power-off condition, the situation that collects charges for electricity in advance, metering fault letter are repaid in expense time percentage, average arrearage number of times, the electricity charge
Breath, preventive trial information, potential safety hazard information, transgression for using electricity situation, violation number of times of checking meter, the quantifier number of stoppages are weighed；
Policy guidance includes whether that listing policy in limits, whether lists policy encouragement measurement in；
Described purchase of electricity, electricity charge area contribution rate, enterprise's purchase of electricity contribution degree ratio contribution degree, business electrical average price, hold
Row electricity price standard, account for the measurement of the indexs such as the ratio of field capacity, valley power consumption rate, Peak power use rate, electric capacity change electricity consumption situation,
Whether whether the indexs such as electric capacity change cycle, purchase of electricity rate of increase, purchase of electricity increment weighed, be highly energy-consuming, be military project medical treatment
Deng forceful electric power require enterprise, load significance level, electric pressure, scale of consumer, plant-grid connection situation, whether thousand grades of high pressure clients,
Electric task measurements, telegram in reply, the quantifier number of stoppages, client's load type are protected in participation, double multi-source client, height jeopardize again
Want customer grade, change electricity consumption situation, capacity change frequency, proof cycle measurement, data completeness, be whether tripartite client,
Accumulative tariff recovery rate, current period tariff recovery rate, tariff recovery are spent on time, tariff recovery speed, repay electricity charge ability, payment when
Length, price sensitivity, customer interaction behavior, the measurement of electricity consumption duration, electricity charge troughput, electricity charge rate of increase, business impact degree, industry
The degree of prosperity, place industrial advantages, enterprise management level, enterprise's place life cycle measurement, history arrearage volume, accumulative arrearage number of times
Ratio, average arrearage number of times, the electricity charge repay guarantee level, arrearage power-off condition, the situation that collects charges for electricity in advance, metering fault information, pre-
Whether anti-property Test Information, potential safety hazard information, transgression for using electricity situation, violation number of times of checking meter, the quantifier number of stoppages are weighed, are arranged
Enter policy to limit, whether list policy in and encourage to weigh is three-level index.
Preferably, described screen related authentic and valid field from data base and mated, determine the n item in three-level index
Base values, comprising:
Will be corresponding with Database field for the three-level index determining；
The index of field vacancy, error in data is deleted from index storehouse, runs into some index time cycle inapplicable feelings
Condition, is supplied with average；
Three-level index after Database field is corresponding is updated to n item base values.
Preferably, the method for described employing hierarchical clustering and Pareto Empiric Rule are to n item base values evaluation rank, comprising:
To electric power data normalization pretreatment；
Determine k value scope, and calculate the silhouette coefficient of k value；
Extract corresponding k value and obtain cluster analysis result, determine threshold value and the score value of divided rank according to the numerical value of different clusters, as follows
Table.
| The numerical value of k | Every corresponding score value of one-level |
| 3 | 33、66、100 |
| 4 | 25、50、75、100 |
| 5 | 20、40、60、80、100 |
| 6 | 16、32、48、60、72、84、100 |
| 7 | 14、28、43、57、71、86、100 |
| 8 | 12、25、47、50、62、75、87、100 |
| 9 | 11、22、33、44、55、66、77、88、100 |
| 10 | 10、20、30、40、50、60、70、80、90、100 |
Wherein, k value is typically chosen as 3-10, carries out k=3,4 ..., 10 cluster calculation respectively afterwards, and calculates the wheel of k value
Wide coefficient.The value of k in the range of be the bigger the better, k value is bigger, and the cluster of cluster is just more cohesion.Can quantify for electricity, capacity etc.
Index considers Pareto Law, professional knowledge and silhouette coefficient simultaneously, chooses k value more than 0.5 for the silhouette coefficient.
Specifically, Euclidean distance:
K-means algorithm is the clustering algorithm applied widely, is the typical object function cluster side based on prototype
The representative of method, obtains the regulation rule of interative computation using the method that function seeks extreme value.K-means algorithm using Euclidean distance as
Similarity measure.The basic procedure of k-means algorithm is: randomly choose k object in data set, as can individual class initial
Center；To remaining object, according to its distance with each class center, assign them to most like apoplexy due to endogenous wind；To each class,
Using distributing to such object, calculate new average, as such new center；Redistributing all objects, repeatedly
Iteration, until distribution is stable.
1) arbitrarily select k object from d as initial class center；
2)repeat
(1) according to the distance with each class center, each object is distributed to the class of " most like "；
(2) meansigma methodss of each cluster are recalculated, as Xin Lei center；
3) until no longer changes
K-means algorithm time complexity is o (nkl), and n is the number of data here, and k is the quantity of cluster, and l is algorithm
Reach the required iterationses of convergence.Normally, k < < n, and l < < n.Therefore for processing large data sets, this algorithm is relatively can
Flexible and effective.
K value is chosen and is determined by silhouette coefficient and actual index business demand
Assume that we have passed through certain algorithm, data to be sorted is clustered.Conventional such as using k-
Means, data to be sorted has been divided into k cluster.For each vector in cluster.Calculate their silhouette coefficient respectively.
For one of point i:
Calculate a (i)=average (distances of i vector other points in all clusters that it belongs to)
Calculate b (i)=min (average distance of the point to all place clusters non-itself for the i vector)
So i vector silhouette coefficient is just:
The value of visual profile coefficient is between [- 1,1], more level off to 1 represent cohesion degree and separating degree all relatively excellent.
By silhouette coefficient a little be averaging it is simply that the total silhouette coefficient of this cluster result.
The bigger k value of silhouette coefficient is chosen in practical operation.
In the present embodiment, the threshold value of timing more New Set can be elapsed according to the time, ensure that updatability and many simultaneously
Sample.
Preferably, described combination Delphi method expert point rating method obtain collect index seek the opinion of score graph and client's total score, bag
Include:
With n item base values for affecting the factor design value analyses object consultation table of bond；
Carry out fraction integration with the index score that addition evaluation type Delphi method expert point rating method is chosen to expert, will count
Result feeds back to expert；
The fraction of each index according to expert amendment and object total score are seeked the opinion of with reference to anonymity and suggestion feedback, obtain index
Final score value and total score, obtain index and seek the opinion of score graph.
Wherein, Delphi method is a kind of expert point rating method, and it selectes several according to the specific requirement of evaluation object first
Assessment item, works up evaluation criterion further according to assessment item.Seek the opinion of the suggestion of relevant expert by anonymous way, expert is anticipated
See and carry out counting, process, analyze and concluding, objectively comprehensive majority expertises and subjective judgment, to being difficult in a large number adopt skill
The factor that art method carries out quantitative analyses makes reasonable estimation, after seeking the opinion of, feed back and adjust through excessive wheel suggestion, credits is worth
Can achieve the method that degree is analyzed with value.Operating procedure is:
1. selection expert:
2. determine the factor that impact credits are worth, design value analysis object consultation table:
3. provide credits background information to expert, expert opinion seeked the opinion of with anonymous way:
4. expert opinion is analyzed collecting, statistical result is fed back to expert:
5. expert is according to the feedback result correction suggestion of oneself:
6. seek the opinion of and suggestion feedback through the anonymity of excessive wheel, form final analytical conclusions.
The computational methods of expert's fraction have:
1. addition evaluation type
The score value addition summation of each index subjet gained will be evaluated, to represent evaluation result by total score.This method is used for index
Between the simple person of relation.Formula is:
Wherein: w is evaluation object total score；wiFor i-th index score value；N is index item number.This method has two kinds of sides
Formula: even add point system and a point meter addition assessment method.
2. connect long-pending evaluation type
The score value of each project is even taken advantage of, and to show performance effect by its product size.This method sensitivity is very high,
The relation being evaluated between each index of object is especially close, and the fraction of one of which is related to have influence on the total of other items
As a result, that is,
Have that a certain index is unqualified, just play, to overall, the feature that negative acts on.
Formula is:
Wherein, w is evaluation object total score, wiFor i project score value, n is index item mesh number.
3. sum multiplication evaluation type
The evaluation index of evaluation object is divided into some groups, first calculates each group score value sum, then again each group is commented
Score value is even taken advantage of, and gained is total scoring.Person allows for degree difference in close relations and the side of influencing each other between each factor
Formula difference is determining.
Formula is:
Wherein: wijFor i-th group of j-th desired value in evaluation object, m is the group number of evaluation object, and n is to contain in i group
Index item number
4. weighting evaluation type
By the indices project in evaluation object according to the significance level of evaluation index, give different weights, that is, right
The significance level of each factor is treated with a certain discrimination.
Wherein, w is evaluation object PTS, wiFor the i index item score of evaluation object, aiWeights for i index item.And:
1)
Preferably, described analysis according to multiple linear regression analysis method determines base values weight, comprising:
With the corresponding score value of n item base values evaluation rank as independent variable, seek the opinion of the client in score graph to collect index
Total score is dependent variable, carries out multiple linear regression analysis method using matlab or r language tool, determines that each base values is weighed
Weight.
Wherein, for one-variable linear regression, data straight line models.Linear regression is simplest regression forms.Double changes
Amount returns the linear letter that a stochastic variable y (referred to as response variable) is considered as another stochastic variable x (referred to as predictor variable)
Number.That is:
Y=alpha+beta x
Wherein, the variance of y is constant；α and β is regression coefficient, represents that straight line is oblique with straight line in blocking of y-axis respectively
Rate.These coefficients can be solved with least square method, and this makes error between real data and the estimation of this straight line minimum.Given
S sample or shape such as (x1, y1), (x2, y2) .., the data point of (xs, ys), regression coefficient α and β can be calculated with following formula:
Wherein,It is x1, the meansigma methodss of x2 .., xs, and y is y1, the meansigma methodss of y2 .., ys.With other complicated recurrence
Method is compared, and linear regression is usually given approximate well.
For multiple linear regression, multiple linear regression is the popularization of simple linear regression, refers to multiple dependent variable pair
The recurrence of multiple independent variables.Being to be only limited to a dependent variable but having the situation of multiple independent variables of most common of which, is also named multiple
Return.The general type of multiple regression is as follows:
Y=a+b1x1+b2x2+b3x3+...+bkxk
A represents intercept, b1,b2,b3,...,bkFor regression coefficient.
In the method, multiple regression agrees with x1 extremely with the model of index system, and x2 .., xs are considered as respectively here
The index score value of item base values variable；The comprehensive total score that y is given for expert；b1,b2,b3,...,bkIt is that we need to obtain
Indices weight, b1,b2,b3,...,bkMay be negative value, this represents that this index plays to the point value of evaluation of client
Retroaction；A is set as the initial factor of index system.
In the method, multiple linear regression agrees with extremely with the model of index system.X1, x2 .., xs can see here
Make the index score value of every base values variable；The comprehensive total score that y is given for expert；b1,b2,b3,...,bkIt is our needs
The weight of the indices obtaining, b1,b2,b3,...,bkMay be negative value, this represents the point value of evaluation to client for this index
Play retroaction；A is set as the initial factor of index system.
In the present embodiment, make to calculate complexity using the different brackets data independent variable the most of third level n item basic index
Degree reduces, reduction hardware requirement, and reasonable employment cluster result is more objective efficient.
As shown in Fig. 2 a kind of electrical network customer general value assessment system of the embodiment of the present invention, described electrical network client comprehensive
Valve estimating system includes:
Client's index division module 201, for dividing to electrical network client's index, obtains first class index, two-level index
With three-level index, described inclusion current value and potential value, described current value include revenue contributions, stability, society effect
Benefit and operation benefits, described potential value includes loyalty, exhibition prospect, credit rating and policy guidance, described revenue contributions, stable
Property, social benefit and operation benefits be two-level index, described revenue contributions, stability, social benefit and operation benefits are two grades
Index；
Database matching module 202, is mated for screening related authentic and valid field from data base, is determined three-level
N item base values in index；
Level evaluation module 203, for being passed judgment on to n item base values using hierarchy clustering method and Pareto Empiric Rule etc.
Level；
Grading module 204, for obtaining with reference to Delphi method expert point rating method, collection index seeks the opinion of score graph and client is total
Point；
Weight determination module 205, determines base values weight for analyzing according to multiple linear regression analysis method.
Preferably, described level evaluation module includes:
Normalization unit, for electric power data normalization pretreatment；
Coefficient calculation unit, for determining k value scope, and calculates the silhouette coefficient of k value；
Grade threshold unit, obtains cluster analysis result for extracting corresponding k value, determines divided rank according to the numerical value of different clusters
Threshold value and scoring.
Preferably, institute's scoring module includes:
Opinionaire design cell, for being seeked the opinion of with n item base values for the factor design value analyses object affecting bond
Opinionaire；
Fraction integral unit, for being entered to the index score that expert chooses with addition evaluation type Delphi method expert point rating method
Row fraction is integrated, and statistical result is fed back to expert；
Fraction determining unit, the fraction for each index according to expert amendment and object total score seek the opinion of with reference to anonymity and
Suggestion feedback, obtains the final score value of index and total score, obtains index and seek the opinion of score graph.
Preferably, described weight determination module includes:
Weight determining unit, for the corresponding score value of n item base values evaluation rank as independent variable, being levied with collecting index
The client's total score ask in score graph is dependent variable, carries out multiple linear regression using matlab or r language tool, determines each
Base values weight.
The electrical network customer general value assessment system of the embodiment of the present invention, the method for employing is commented for electrical network customer general value
Estimate method, the therefore feature of electrical network customer general value assessment system is identical with electrical network customer general value appraisal procedure, here
Repeat no more.
The above is the preferred embodiment of the present invention it is noted that for those skilled in the art
For, on the premise of without departing from principle of the present invention, some improvements and modifications can also be made, these improvements and modifications
Should be regarded as protection scope of the present invention.
Claims (10)
1. a kind of electrical network customer general value appraisal procedure is it is characterised in that described electrical network customer general value appraisal procedure bag
Include:
Electrical network client's index is divided, obtains first class index, two-level index and three-level index, described inclusion current value and
Potential value, described current value includes revenue contributions, stability, social benefit and operation benefits, and described potential value includes
Loyalty, exhibition prospect, credit rating and policy guidance, described revenue contributions, stability, social benefit and operation benefits refer to for two grades
Mark；
Screen related authentic and valid field to be mated from data base, determine the n item base values in three-level index；
Using hierarchy clustering method and Pareto Empiric Rule to n item base values evaluation rank；
Obtain collection index in conjunction with Delphi method expert point rating method and seek the opinion of score graph and client's total score；
Analyzed according to multiple linear regression analysis method and determine base values weight.
2. electrical network customer general value appraisal procedure according to claim 1 is it is characterised in that described revenue contributions include
Purchase of electricity, electricity charge area contribution rate, enterprise's purchase of electricity contribution degree ratio contribution degree, business electrical average price, execution electricity price standard, account for
The indexs such as the ratio of field capacity are weighed；Stability includes valley power consumption rate, Peak power use rate, electric capacity change electricity consumption situation, electricity
Hold the indexs such as change cycle, purchase of electricity rate of increase, purchase of electricity increment to weigh；Whether social benefit includes whether as highly energy-consuming,
For the forceful electric power such as military project medical treatment require enterprise, load significance level, electric pressure, scale of consumer, plant-grid connection situation, whether thousand grades
High pressure client, whether participate in protecting electric task and weigh；Operation benefits include telegram in reply, the quantifier number of stoppages, client's load type, are
No pair of multi-source client, height jeopardize Very Important Person grade, change electricity consumption situation, capacity change frequency, proof cycle measurement；Loyalty
Including data completeness, be whether tripartite client, accumulative tariff recovery rate, current period tariff recovery rate, tariff recovery spend on time, electricity
Take recovery rate, repay electricity charge ability, payment duration, price sensitivity, customer interaction behavior, the measurement of electricity consumption duration；Before development
Scape include electricity charge troughput, electricity charge rate of increase, business impact degree, Industrial Cycle degree, place industrial advantages, enterprise management level,
Enterprise's place life cycle is weighed；Credit rating includes history arrearage volume, accumulative arrearage time percentage, average arrearage number of times, the electricity charge
Repay guarantee level, arrearage power-off condition, the situation that collects charges for electricity in advance, metering fault information, preventive trial information, potential safety hazard letter
Breath, transgression for using electricity situation, violation number of times of checking meter, the quantifier number of stoppages are weighed；Policy guidance includes whether to list policy limit in
Make, whether list policy encouragement measurement in；
Described purchase of electricity, electricity charge area contribution rate, enterprise's purchase of electricity contribution degree ratio contribution degree, business electrical average price, execute electricity
Price card is accurate, account for the measurement of the indexs such as the ratio of field capacity, valley power consumption rate, Peak power use rate, electric capacity change electricity consumption situation, electric capacity
Whether whether the indexs such as change cycle, purchase of electricity rate of increase, purchase of electricity increment weighed, be highly energy-consuming, be that military project medical treatment etc. is strong
Electricity requires enterprise, load significance level, electric pressure, scale of consumer, plant-grid connection situation, whether thousand grades of high pressure clients, whether
Participate in protecting electric task measurement, telegram in reply, the quantifier number of stoppages, client's load type, whether double multi-source client, height jeopardize important visitor
Family grade, change electricity consumption situation, capacity change frequency, proof cycle measurements, data completeness, be whether tripartite client, add up
Tariff recovery rate, current period tariff recovery rate, tariff recovery are spent on time, tariff recovery speed, repay electricity charge ability, payment duration,
Price sensitivity, customer interaction behavior, the measurement of electricity consumption duration, electricity charge troughput, electricity charge rate of increase, business impact degree, industry scape
Manner, place industrial advantages, enterprise management level, enterprise's place life cycle measurement, history arrearage volume, accumulative arrearage number of times ratio
Guarantee level, arrearage power-off condition, the situation that collects charges for electricity in advance, metering fault information, prevention are repaid in rate, average arrearage number of times, the electricity charge
Property Test Information, potential safety hazard information, transgression for using electricity situation, violation number of times of checking meter, the quantifier number of stoppages weigh, whether list in
Policy limits, whether lists policy in and encourage to weigh is three-level index.
3. electrical network customer general value appraisal procedure according to claim 1 and 2 it is characterised in that described from data base
The related authentic and valid field of middle screening is mated, and determines the n item base values in three-level index, comprising:
Will be corresponding with Database field for the three-level index determining；
The index of field vacancy, error in data is deleted from index storehouse, runs into some index time cycle inapplicable situations, with
Average is supplied；
Three-level index after Database field is corresponding is updated to n item base values.
4. electrical network customer general value appraisal procedure according to claim 1 and 2 is it is characterised in that described employing level
The method of cluster and Pareto Empiric Rule are to n item base values evaluation rank, comprising:
To electric power data normalization pretreatment；
Determine k value scope, and calculate the silhouette coefficient of k value；
Extract corresponding k value and obtain cluster analysis result, determine threshold value and the score value of divided rank according to the numerical value of different clusters.
5. electrical network customer general value appraisal procedure according to claim 4 is it is characterised in that described combination Delphi method
Expert point rating method obtains collection index and seeks the opinion of score graph and client's total score, comprising:
With n item base values for affecting the factor design value analyses object consultation table of bond；
Carry out fraction integration with the index score that addition evaluation type Delphi method expert point rating method is chosen to expert, by statistical result
Feed back to expert；
The fraction of each index according to expert amendment and object total score are seeked the opinion of with reference to anonymity and suggestion feedback, obtain index final
Score value and total score, obtain index and seek the opinion of score graph.
6. electrical network customer general value appraisal procedure according to claim 5 it is characterised in that described according to multiple linear
Homing method analysis determines base values weight, comprising:
With the corresponding score value of n item base values evaluation rank as independent variable, seek the opinion of the client's total score in score graph to collect index
For dependent variable, carry out multiple linear regression analysis method using mat l ab or r language tool, determine each base values weight.
7. a kind of electrical network customer general value assessment system is it is characterised in that described electrical network customer general value assessment system bag
Include:
Client's index division module, for dividing to electrical network client's index, obtains first class index, two-level index and three-level and refers to
Mark, described inclusion current value and potential value, described current value includes revenue contributions, stability, social benefit and operation
Benefit, described potential value includes loyalty, exhibition prospect, credit rating and policy guidance, described revenue contributions, stability, society
Benefit and operation benefits are two-level index, and described revenue contributions, stability, social benefit and operation benefits are two-level index；
Database matching module, is mated for screening related authentic and valid field from data base, is determined in three-level index
N item base values；
Level evaluation module, for using hierarchy clustering method and Pareto Empiric Rule to n item base values evaluation rank；
Grading module, seeks the opinion of score graph and client's total score for obtaining collection index with reference to Delphi method expert point rating method；
Weight determination module, determines base values weight for analyzing according to multiple linear regression analysis method.
8. electrical network customer general value assessment system according to claim 7 is it is characterised in that described level evaluation module
Including:
Normalization unit, for electric power data normalization pretreatment；
Coefficient calculation unit, for determining k value scope, and calculates the silhouette coefficient of k value；
Grade threshold unit, obtains cluster analysis result for extracting corresponding k value, determines the threshold of divided rank according to the numerical value of different clusters
Value and scoring.
9. electrical network customer general value assessment system according to claim 8 is it is characterised in that institute's scoring module bag
Include:
Opinionaire design cell, for n item base values for affecting the factor design value analyses object consultation of bond
Table；
Fraction integral unit, for being carried out with the index score that addition evaluation type Delphi method expert point rating method is chosen to expert point
Number is integrated, and statistical result is fed back to expert；
Fraction determining unit, the fraction for each index according to expert amendment and object total score are seeked the opinion of with reference to anonymity and suggestion
Feedback, obtains the final score value of index and object total score, obtains index and seek the opinion of score graph.
10. electrical network customer general value assessment system according to claim 9 is it is characterised in that described weight determines mould
Block includes:
Weight determining unit, for the corresponding score value of n item base values evaluation rank as independent variable, is seeked the opinion of point with collecting index
Client's total score in number table is dependent variable, carries out multiple linear regression using mat l ab or r language tool, determines each base
Plinth index weights.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610696848.1A CN106372775A (en) | 2016-08-19 | 2016-08-19 | Assessment method and system of comprehensive value of power grid client |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610696848.1A CN106372775A (en) | 2016-08-19 | 2016-08-19 | Assessment method and system of comprehensive value of power grid client |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106372775A true CN106372775A (en) | 2017-02-01 |
Family
ID=57878583
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610696848.1A Pending CN106372775A (en) | 2016-08-19 | 2016-08-19 | Assessment method and system of comprehensive value of power grid client |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106372775A (en) |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107133734A (en) * | 2017-04-28 | 2017-09-05 | 浙江极赢信息技术有限公司 | A kind of Channel Quality evaluation method and system |
| CN107180392A (en) * | 2017-05-18 | 2017-09-19 | 北京科技大学 | A kind of electric power enterprise tariff recovery digital simulation method |
| CN107292518A (en) * | 2017-06-21 | 2017-10-24 | 中国农业科学院农田灌溉研究所 | Topsoil index acquisition methods and device |
| CN107729651A (en) * | 2017-10-17 | 2018-02-23 | 黄河水利委员会黄河水利科学研究院 | Domatic rill developmental morphology characteristic synthetic quantization method based on various dimensions |
| CN108021739A (en) * | 2017-11-22 | 2018-05-11 | 中国北方发动机研究所(天津) | A kind of high-power military diesel machine Real-Time Model parameter Impact analysis method |
| CN108074108A (en) * | 2017-11-02 | 2018-05-25 | 平安科技（深圳）有限公司 | A kind of display methods and its terminal of net recommendation |
| CN108446834A (en) * | 2018-03-02 | 2018-08-24 | 国网湖北省电力公司 | A kind of residential electricity consumption boot policy Potentials method based on fuzzy evaluation |
| CN108931755A (en) * | 2018-06-11 | 2018-12-04 | 宁波三星智能电气有限公司 | A kind of electric energy meter power grid quality determining method |
| CN109508855A (en) * | 2018-09-27 | 2019-03-22 | 国网福建省电力有限公司信息通信分公司 | A kind of sales service compliance discriminatory analysis method based on big data processing |
| CN110163706A (en) * | 2018-02-13 | 2019-08-23 | 北京京东尚科信息技术有限公司 | Method and apparatus for generating information |
| CN110197313A (en) * | 2018-02-27 | 2019-09-03 | 顺丰科技有限公司 | Employee's evaluation method and device, equipment and storage medium |
| CN110288395A (en) * | 2019-06-20 | 2019-09-27 | 卓尔智联(武汉)研究院有限公司 | Outdoor advertising position Valuation Method, electronic equipment and storage medium |
| CN110858343A (en) * | 2018-08-23 | 2020-03-03 | 国信优易数据有限公司 | Data asset value evaluation system and method |
| CN111709327A (en) * | 2020-05-29 | 2020-09-25 | 中国人民财产保险股份有限公司 | Fuzzy matching method and device based on OCR recognition |
| CN112102003A (en) * | 2020-09-18 | 2020-12-18 | 国网辽宁省电力有限公司电力科学研究院 | Big data platform-based electricity customer core resource management system and method |
| CN112100246A (en) * | 2020-09-22 | 2020-12-18 | 国网辽宁省电力有限公司电力科学研究院 | Customer electricity value mining method based on multi-dimensional graph code label |
| CN113407827A (en) * | 2021-06-11 | 2021-09-17 | 广州三七极创网络科技有限公司 | Information recommendation method, device, equipment and medium based on user value classification |
| CN113591018A (en) * | 2021-07-30 | 2021-11-02 | 中国联合网络通信集团有限公司 | Communication client classification management method, system, electronic device and storage medium |
| CN113780861A (en) * | 2021-09-18 | 2021-12-10 | 深圳供电局有限公司 | Component index evaluation method and system based on user daily electric quantity adjustment value |
-
2016
- 2016-08-19 CN CN201610696848.1A patent/CN106372775A/en active Pending
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107133734A (en) * | 2017-04-28 | 2017-09-05 | 浙江极赢信息技术有限公司 | A kind of Channel Quality evaluation method and system |
| CN107180392A (en) * | 2017-05-18 | 2017-09-19 | 北京科技大学 | A kind of electric power enterprise tariff recovery digital simulation method |
| CN107292518A (en) * | 2017-06-21 | 2017-10-24 | 中国农业科学院农田灌溉研究所 | Topsoil index acquisition methods and device |
| CN107729651A (en) * | 2017-10-17 | 2018-02-23 | 黄河水利委员会黄河水利科学研究院 | Domatic rill developmental morphology characteristic synthetic quantization method based on various dimensions |
| CN108074108A (en) * | 2017-11-02 | 2018-05-25 | 平安科技（深圳）有限公司 | A kind of display methods and its terminal of net recommendation |
| CN108074108B (en) * | 2017-11-02 | 2021-02-09 | 平安科技（深圳）有限公司 | Method and terminal for displaying net recommendation value |
| CN108021739A (en) * | 2017-11-22 | 2018-05-11 | 中国北方发动机研究所(天津) | A kind of high-power military diesel machine Real-Time Model parameter Impact analysis method |
| CN110163706A (en) * | 2018-02-13 | 2019-08-23 | 北京京东尚科信息技术有限公司 | Method and apparatus for generating information |
| CN110163706B (en) * | 2018-02-13 | 2024-04-19 | 北京京东尚科信息技术有限公司 | Method and device for generating information |
| CN110197313A (en) * | 2018-02-27 | 2019-09-03 | 顺丰科技有限公司 | Employee's evaluation method and device, equipment and storage medium |
| CN108446834A (en) * | 2018-03-02 | 2018-08-24 | 国网湖北省电力公司 | A kind of residential electricity consumption boot policy Potentials method based on fuzzy evaluation |
| CN108931755A (en) * | 2018-06-11 | 2018-12-04 | 宁波三星智能电气有限公司 | A kind of electric energy meter power grid quality determining method |
| CN110858343A (en) * | 2018-08-23 | 2020-03-03 | 国信优易数据有限公司 | Data asset value evaluation system and method |
| CN109508855A (en) * | 2018-09-27 | 2019-03-22 | 国网福建省电力有限公司信息通信分公司 | A kind of sales service compliance discriminatory analysis method based on big data processing |
| CN110288395A (en) * | 2019-06-20 | 2019-09-27 | 卓尔智联(武汉)研究院有限公司 | Outdoor advertising position Valuation Method, electronic equipment and storage medium |
| CN111709327A (en) * | 2020-05-29 | 2020-09-25 | 中国人民财产保险股份有限公司 | Fuzzy matching method and device based on OCR recognition |
| CN111709327B (en) * | 2020-05-29 | 2023-06-27 | 中国人民财产保险股份有限公司 | Fuzzy matching method and device based on OCR (optical character recognition) |
| CN112102003A (en) * | 2020-09-18 | 2020-12-18 | 国网辽宁省电力有限公司电力科学研究院 | Big data platform-based electricity customer core resource management system and method |
| CN112100246A (en) * | 2020-09-22 | 2020-12-18 | 国网辽宁省电力有限公司电力科学研究院 | Customer electricity value mining method based on multi-dimensional graph code label |
| CN113407827A (en) * | 2021-06-11 | 2021-09-17 | 广州三七极创网络科技有限公司 | Information recommendation method, device, equipment and medium based on user value classification |
| CN113591018A (en) * | 2021-07-30 | 2021-11-02 | 中国联合网络通信集团有限公司 | Communication client classification management method, system, electronic device and storage medium |
| CN113780861A (en) * | 2021-09-18 | 2021-12-10 | 深圳供电局有限公司 | Component index evaluation method and system based on user daily electric quantity adjustment value |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106372775A (en) | Assessment method and system of comprehensive value of power grid client | |
| Bretschneider et al. | Political and organizational influences on the accuracy of forecasting state government revenues | |
| Waring | Industry differences in the persistence of firm-specific returns | |
| Kourentzes et al. | Improving forecasting by estimating time series structural components across multiple frequencies | |
| Xu | Choquet integrals of weighted intuitionistic fuzzy information | |
| Mutl et al. | The Hausman test in a Cliff and Ord panel model | |
| CN109063945B (en) | Value evaluation system-based 360-degree customer portrait construction method for electricity selling company | |
| CN103258069B (en) | A kind of Forecasting Methodology of steel industry electricity needs | |
| He et al. | Comprehensive evaluation of regional clean energy development levels based on principal component analysis and rough set theory | |
| CN110909983B (en) | Multidimensional assessment method for electric energy quality of active power distribution network | |
| CN103632203A (en) | Distribution network power supply area division method based on comprehensive evaluation | |
| CN104123600B (en) | A kind of electric power manager's exponential trend Forecasting Methodology towards representative row sparetime university data | |
| CN106355518A (en) | Electricity fee payment customer screening method and system | |
| Radojicic et al. | Measuring the efficiency of banks: the bootstrapped I-distance GAR DEA approach | |
| Jánošíková et al. | Location of emergency stations as the capacitated p-median problem | |
| CN106447075B (en) | Trade power consumption needing forecasting method and system | |
| CN109636146A (en) | A kind of user demand response potentiality portrait method | |
| CN106447198A (en) | Power consumption checking method based on business expanding installation data | |
| CN106600146A (en) | Electricity fee collection risk evaluation method and apparatus | |
| CN115496627A (en) | Method and system for evaluating response potential of adjustable resource | |
| CN105913366A (en) | Industrial electric power big data-based regional industry business climate index building method | |
| Morin et al. | Estimating capacity utilization from survey data | |
| Kung et al. | A fuzzy MCDM method to select the best company based on financial report analysis | |
| Jorgenson et al. | Progress on measuring the industry origins of the japan-us productivity gap | |
| Khalili-Damghani et al. | A hybrid approach based on multi-criteria satisfaction analysis (MUSA) and a network data envelopment analysis (NDEA) to evaluate efficiency of customer services in bank branches |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170201 |