CN108053094A - A kind of weight grade evaluation method and system - Google Patents
A kind of weight grade evaluation method and system Download PDFInfo
- Publication number
- CN108053094A CN108053094A CN201711132881.2A CN201711132881A CN108053094A CN 108053094 A CN108053094 A CN 108053094A CN 201711132881 A CN201711132881 A CN 201711132881A CN 108053094 A CN108053094 A CN 108053094A
- Authority
- CN
- China
- Prior art keywords
- data
- weight
- sample
- mrow
- opinion rating
- 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
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression 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
- 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
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Strategic Management (AREA)
- Data Mining & Analysis (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Mathematical Physics (AREA)
- General Business, Economics & Management (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Mathematics (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Computational Biology (AREA)
- Algebra (AREA)
- Life Sciences & Earth Sciences (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of weight grade evaluation method and systems, and the sample data transferred from sampling memory is transferred to netscape messaging server Netscape the described method includes data collection station；Netscape messaging server Netscape pre-processes sample data；Weight analysis is carried out to the sample data by pretreatment；It is analyzed by weight and obtains opinion rating model；Opinion rating model is stored in data storage；Data collection station collects pending data, and pending data is inputted opinion rating model, obtains the opinion rating of pending data.A kind of weight grade evaluation method of the present invention and system are acquiring magnanimity government department data, on the basis of the relevant informations such as inside data of enterprise and public sentiment data, with weight analytic approach, multiple evaluation indexes are converted into a few overall target, quick succinct evaluation can be achieved to calculate, compensate for the deficiency that traditional evaluation method ignores incidence relation between index in evaluation.
Description
Technical field
The present invention relates to the method and system that network data management field more particularly to a kind of weight grade are evaluated.
Background technology
Big data technology be using data as essence a kind of information technology, during data are taped the latent power, can drive theory,
The innovation of pattern, technology and application practice.The improvement of the highlighting of data value, data acquisition means and data processing technique is
The root of " big data " outburst.And with data production factors, data science, the continuous development of data science and technology and data value
Depth excavate and application, by driving, Regional Economic Development, smart city construction, enterprise transformation upgrading, social management etc. are each
The innovation and development in field.Big data will change innovation mode and management philosophy under the guidance of data science theory, and development is big
Data technique, the application of in-depth big data and practice.And industry big data will be the maximum optimal application field of big data.
The application of big data is applied in business administration, statistical desired realization can be precipitated in the data of magnanimity
As a result, training counting statistics model, helps company manager to grasp accurate company information, corporate behavior and the guiding public is instructed to disappear
Take.But current big data apply there is it is many shortcomings that, for example, the analysis of data needs the data accumulation based on magnanimity, mesh
Preceding big data needs are analyzed according to millions of mass datas, and the platform of the overwhelming majority lacks big data and relies on,
Often small data or middle data, therefore, the scarcity of data can cause result of calculation to lack persuasion.In addition, existing conventional enterprise
Industry credit crisis evaluation model mainly has multivariate statistical model, the Credit Risk Assessment Model based on artificial intelligence, based on city
Credit Risk Assessment Model of field value etc., these models are harsher for the precondition of sample, the sample number in reality
These hypothesis have been run counter to according to serious.And the variable of model selection can not avoid mutual Problems of Multiple Synteny.Manually
Although the foundation of intelligent Credit Model solves the problems, such as these to a certain extent, knot that the training of this "black box" formula is drawn
By lacking interpretability to a certain extent.Meanwhile if using business standing crisis as a Multivariable, variable is too many
Calculation amount can be increased and increase the complexity of problem analysis.In order to increase data validity, should quantitatively divided as much as possible
The variable being related to during analysis is less, and obtained information content is more.
The content of the invention
In order to solve the above-mentioned problems in the prior art, the present invention proposes a kind of weight grade evaluation method
And system, on the basis of the information such as magnanimity government department data, inside data of enterprise and the public sentiment data of acquisition, by enterprise
Multiple evaluation indexes of credit crisis are converted into a few overall target, it can be achieved that quick succinct evaluation calculating, compensates for
Traditional evaluation method ignores the deficiency of incidence relation between index in evaluation.
The technical solution used in the present invention is：
A kind of weight grade evaluation method, including being stored in sample data in sampling memory, data collection station
Sample data is transferred from sampling memory,
The sample data transferred from sampling memory is transferred to netscape messaging server Netscape by data collection station；
Netscape messaging server Netscape pre-processes sample data；
Netscape messaging server Netscape carries out weight analysis to the sample data by pretreatment；
It is analyzed by weight and obtains opinion rating model；
Opinion rating model is stored in data storage；
Data collection station collects pending data, and pending data is inputted opinion rating model, obtains pending number
According to opinion rating.
Preferably, sample data includes government department's database data, inside data of enterprise and public sentiment data.
In any of the above-described scheme preferably, weight analysis includes：
Sample data comprehensive score is calculated, according to data distribution state demarcation classification in comprehensive score and sample data, if
Determine data level；
By parameter coefficient, comprehensive score function is built, calculates weight score.
In any of the above-described scheme preferably, sample data comprehensive score is calculated to further comprise：
Test samples data are if appropriate for using weight factor approach；
Weight chooses the method using factor of the keeping characteristics value of SPSS software defaults more than 1；
Variance is rotated with orthogonal very big method, carries out factorial analysis；
S weight before selection calculates comprehensive score by the weight coefficient of each factor according to the Return Law.
In any of the above-described scheme preferably, according to the distribution character of sample scores, classification point is selected, setting is each
A data level.
In any of the above-described scheme preferably, the distribution character of sample scores for linear distribution or is just distributed very much.
In any of the above-described scheme preferably, comprehensive score function is built, is further comprised：Build anticipation function mould
Type selects s weight, then the score function of sample estimates is：
F=a1z1+a2z2+…+aszs (1)
Wherein, F：Historical sample comprehensive score；
ai：The contribution rate of i-th of weight；
zi：I-th of weight.
In any of the above-described scheme preferably,
Parameter coefficient is：
Wherein：I, l=1,2 ... s；
KiFor the ith feature value arranged from big to small；
bilFor weight ziTo zlAccumulation contribution rate.
In any of the above-described scheme preferably, i-th of weight can show as linear group of n original index
It closes, then：
zi=bi1x1+ai2x2+…+ainxn (2)
(2) are substituted into (1)：
F=c1x1+c2x2+…+cnxn (3)
Wherein：
I=1,2 ..., s；
J=1,2 ..., n.
In any of the above-described scheme preferably, the weight for obtaining pending data is evaluated by weight grade
Grade includes the following steps：
Build information collecting platform, the pending data of receive information acquisition platform transmission, using pending data as new
Sample pre-processes new samples data for standard data format by the processing mode identical with sample data；
The function model completed using formula (3) structure calculates comprehensive score, show that the weight of new samples data obtains
Point；
By the weight score of new samples data compared with sample data weight grade, show that new samples data are weighed
The grade evaluation of weight factor determines the weight grade of pending data.
In any of the above-described scheme preferably, by KMO and Barrelett sphericity test samples data if appropriate for making
With weight analysis method, when KMO values are more than 0.5, when sphericity verifies as notable, judge be using weight factor approach
Suitably.
In any of the above-described scheme preferably, sample data is randomly divided into two groups, and one group is training sample group, the instruction
The data for practicing sample group are used for structure forecast model；Another group is test sample group, and the data of the test sample group are used to test
Demonstrate,prove the significant degree of prediction model.
A kind of weight grade evaluation system, including：Sample data memory module：It is configured and adapted in samples storage
Sample data is stored in device, is further included with lower module：
Data collection station：Its sample data for being configured and adapted to transfer from sampling memory is transmitted to information processing
Server；
Sample data preprocessing module：It is configured and adapted to pre-process sample data；
Weight analysis module：It is configured and adapted to carry out weight analysis to the sample data by pretreatment；
Opinion rating model determining module：It, which is configured and adapted to analyze by weight, obtains opinion rating model；
Opinion rating model memory module：It is configured and adapted to opinion rating model being stored in data storage；
Opinion rating acquisition module：It is configured and adapted to after data collection station collects pending data, will be pending
Data input opinion rating model, obtain the opinion rating of pending data.
Preferably, weight analysis module includes：
Sample data comprehensive score computing module is configured to calculate sample data comprehensive score, according to comprehensive score and
Data distribution state demarcation classification in sample data sets data level；
Comprehensive score function builds module：It is configured to, by parameter coefficient, build comprehensive score function, calculates power
Weight factor score.
In any of the above-described scheme preferably, sample data comprehensive score computing module execution is calculated as below：
Test samples data are if appropriate for using weight factor approach；
Weight chooses the method using factor of the keeping characteristics value of SPSS software defaults more than 1；
Variance is rotated with orthogonal very big method, carries out factorial analysis；
S weight before selection calculates comprehensive score by the weight coefficient of each factor according to the Return Law.
In any of the above-described scheme preferably, sample data comprehensive score computing module dividing according to sample scores
Cloth characteristic selectes classification point, sets each data level.
In any of the above-described scheme preferably, the distribution character of sample scores for linear distribution or is just distributed very much.
In any of the above-described scheme preferably, comprehensive score function structure module construction anticipation function model, specially：
S weight is selected, then the score function of sample estimates is：
F=a1z1+a2z2+…+aszs (1)
Wherein, F：Historical sample comprehensive score；
ai：The contribution rate of i-th of weight；
zi：I-th of weight.
In any of the above-described scheme preferably, parameter coefficient is：
Wherein：I, l=1,2 ... s；
KiFor the ith feature value arranged from big to small；
bilFor weight ziTo zlAccumulation contribution rate.
In any of the above-described scheme preferably, i-th of weight can show as linear group of n original index
It closes, then：
zi=bi1x1+ai2x2+…+ainxn (2)
(2) are substituted into (1)：
F=c1x1+c2x2+…+cnxn (3)
Wherein：
I=1,2 ..., s；
J=1,2 ..., n.
In any of the above-described scheme preferably, weight level determination module determines pending number by the following method
According to weight grade：
Build information collecting platform, the pending data of receive information acquisition platform transmission, using pending data as new
Sample pre-processes new samples data for standard data format by the processing mode identical with sample data；
The function model completed using formula (3) structure calculates comprehensive score, show that the weight of new samples data obtains
Point；
By the weight score of new samples data compared with sample data weight grade, show that new samples data are weighed
The grade evaluation of weight factor obtains the weight grade of pending data.
Beneficial effects of the present invention：
A kind of weight grade evaluation method of the present invention and system are acquiring magnanimity government department data, enterprises
On the basis of the business crisis relevant information such as data and public sentiment data, with weight analysis method, evaluation of training grade
Multiple evaluation indexes of business standing crisis are converted into a few overall target, it can be achieved that quick succinct by computation model
Evaluation calculates, and compensates for the deficiency that traditional evaluation method ignores incidence relation between index in evaluation, is obtained so as to accurate
Go out weight grade evaluation result, and effective early warning and management are carried out to business standing crisis, the exhibition of all-dimensional multi-angle ground
Show the credit situation of enterprise, so that government provides reference for business standing deciding grade and level, perfect credit system for next step each place and build
If realizing " combine and discipline as a warning ", service basic and technical foundation have been established.
A kind of weight grade evaluation method of the present invention and system introduce weight analytic approach and come to weight etc.
Grade evaluation index carries out dimensionality reduction.In weight Grade, how the index coefficient of core composite evaluation function obtains
To being key factor, which must at least is fulfilled for the following conditions, could preferably draw accurate credit crisis evaluation
As a result:
1st, can preferably change with sample changed；
2nd, objectivity；
3rd, the reasonability of evaluation result can be ensured.
For these conditions, the present invention is in itself based entirely on data using weight grade evaluation method
Number, so as to accurately draw weight grade evaluation result.
A kind of weight grade evaluation method of the present invention and system are not required to artificially estimate weight, reduce artificial in evaluation
Factor influences, better than traditional evaluation method in accuracy.In addition, the present invention can be according to the Character adjustment index body of different industries
System has certain scalability.
Description of the drawings
Fig. 1 is according to a kind of flow chart of a preferred embodiment of weight grade evaluation method of the present invention；
Fig. 2 is according to a kind of flow chart of the weight analytical procedure of weight grade evaluation method of the present invention；
Fig. 3 is according to a kind of weight grade of the definite pending data of weight grade evaluation method of the present invention
The flow chart of step；
Fig. 4 is according to a kind of flow chart of the another preferred embodiment of weight grade evaluation method of the present invention.
Specific embodiment
Referring to the drawings and embodiment the present invention will be described in detail：
Embodiment 1
It is a kind of weight grade evaluation method as shown in attached drawing 1-4, including：It is stored in sample in sampling memory
Data, data collection station transfer sample data from sampling memory, further include following steps：
Step 1：The sample data transferred from sampling memory is transferred to information processing services by data collection station
Device, sample data include government department's database data, inside data of enterprise and public sentiment data.
Step 2：Netscape messaging server Netscape pre-processes sample data；Test samples data are if appropriate for the right to use
Weight factor approach；By KMO and Barrelett sphericity test samples data if appropriate for using weight factor analysis side
Method, when KMO values be more than 0.5, when sphericity verifies as notable, judge that using weight factor approach be suitable.Sample data
Two groups are randomly divided into, one group is training sample group, and the data of training sample group are used for structure forecast model；Another group is test specimens
This group, the data of test sample group are used to verify the significant degree of prediction model.
Step 3：Netscape messaging server Netscape carries out weight analysis to the sample data by pretreatment；Weight
Analysis includes：
1st, sample data comprehensive score is calculated, according to data distribution state demarcation classification in comprehensive score and sample data,
Set data level；Sample data comprehensive score is calculated to further comprise：Weight chooses the guarantor using SPSS software defaults
The method for staying factor of the characteristic value more than 1；Variance is rotated with orthogonal very big method, carries out factorial analysis；Before selection s weight because
Element calculates comprehensive score by the weight coefficient of each factor according to the Return Law.According to the distribution character of sample scores, choosing
Surely classify a little, set each data level.The distribution character of sample scores for linear distribution or is just distributed very much.
2nd, by parameter coefficient, comprehensive score function is built, calculates weight score.Build comprehensive score letter
Number, further comprises：Anticipation function model is built, selects s weight, then the score function of sample estimates is：
F=a1z1+a2z2+…+aszs (1)
Wherein, F：Historical sample comprehensive score；
ai：The contribution rate of i-th of weight；
zi：I-th of weight.
Parameter coefficient is：
Wherein：I, l=1,2 ... s；
KiFor the ith feature value arranged from big to small；
bilFor weight ziTo zlAccumulation contribution rate.
I-th of weight can show as the linear combination of n original index, then：
zi=bi1x1+ai2x2+…+ainxn (2)
(2) are substituted into (1)：
F=c1x1+c2x2+…+cnxn (3)
Wherein：
I=1,2 ..., s；
J=1,2 ..., n.
Step 4：It is analyzed by weight and obtains opinion rating model；
Step 5：Opinion rating model is stored in data storage；
Step 6：Data collection station collects pending data, and pending data is inputted opinion rating model, is treated
Handle the opinion rating of data.
The weight grade that acquisition pending data is evaluated by weight grade includes the following steps：
Build information collecting platform, the pending data of receive information acquisition platform transmission, using pending data as new
Sample pre-processes new samples data for standard data format by the processing mode identical with sample data；
The function model completed using formula (3) structure calculates comprehensive score, show that the weight of new samples data obtains
Point；
By the weight score of new samples data compared with sample data weight grade, show that new samples data are weighed
The grade evaluation of weight factor obtains the weight grade of pending data.
Embodiment 2
A kind of weight grade evaluation method of the present invention, which is particularly applicable in evaluation business standing crisis grade, to be included：Sample
This selection, sample data pretreatment, weight are analyzed, build anticipation function model, division credit grade, utilize comprehensive function
Prediction new samples score, evaluation business standing crisis grade and etc..
Sample is chosen：Choose target zone in (certain industry or a certain region) a certain number of credit crisis enterprises and
The historical sample of the good enterprise's composition of credit, wherein good enterprise of credit crisis enterprise and credit need predefined.Form history
The credit line that the enterprise of sample determines according to standard data format is divided into credit crisis enterprise or the good enterprise of credit.It will
Total sample is randomly divided into two groups, and one group is training sample group, and another set is test sample group.The data of training sample group are used for
Structure forecast model, and the data that test sample is rented are used to verify the significant degree of prediction model.
Sample data pre-processes：For the accuracy that guarantee weight rank method uses, initial data will be carried out in advance
Processing mainly ensures that data have same directionality, and achievement data is made to have economic meanings.It, will according to the thought
Sample data is first classified according to default current standard, then good using the credit crisis enterprise in a certain concrete kind and credit
Input data of the ratio of enterprise as weight rating calculation.This not only ensure that each achievement data has same direction
Property, and have economic implications.For example, for industry index, farming, forestry, husbandary and fishing, medical and health, building materials, metallurgical ore deposit are classified as
It produces, 5 specific categories of other grades, creditable crisis enterprise 60 in farming, forestry, husbandary and fishing, the good enterprise of credit 34 then should in sample
Item input data is set to 1.765 (60/34), remaining similar process.
Weight is analyzed：When carrying out weight rating calculation, inspection data first is if appropriate for using weight
Analysis method, the instrument for examining this method are mainly that KMO and Barrelett sphericitys are examined, when KMO values are examined more than 0.5 sphericity
For it is notable when, the use of weight rank method is suitable.KMO：Kaiser-Meyer-Olkin, test statistics are to be used for
The index of simple correlation coefficient and partial correlation coefficient between comparison variable.KMO statistics be value between zero and one.When all variables
Between simple correlation coefficient quadratic sum when being far longer than partial correlation coefficient quadratic sum, KMO values close to 1.KMO values closer to 1, meaning
The correlation that taste between variable is stronger, and original variable is more suitable as factorial analysis；When the simple correlation coefficient between all variables is put down
During just and close to 0, KMO values are close to 0.KMO values closer to 0, it is meant that the correlation between variable is weaker, and original variable is more uncomfortable
Cooperation factorial analysis.It is appropraite condition judging historical sample data using weight rank method, on the one hand, to calculate and go through
History sample comprehensive score, the method that weight of the keeping characteristics value more than 1 of SPSS software defaults may be employed in ingredient selection,
Variance is rotated with orthogonal very big method, carries out factorial analysis, s weight before selection by the weight coefficient of each factor and is pressed
Last comprehensive score is calculated according to the Return Law.(may be linear distribution or just according to the distribution characters of sample scores
It is distributed very much), classification point is selected, sets each credit crisis grade.On the other hand, we set structure anticipation function model as：Choosing
S weight is selected, then the score function of sample estimates is：
F=a1z1+a2z2+…+aszs (1)
Wherein, F：Historical sample comprehensive score；
ai：The contribution rate of i-th of weight；
zi：I-th of weight.
Parameter coefficient is：
Wherein：I, l=1,2 ... s；
KiFor the ith feature value arranged from big to small；
bilFor weight ziTo zlAccumulation contribution rate.
I-th of weight can show as the linear combination of n original index, then：
zi=bi1x1+ai2x2+…+ainxn (2)
(2) are substituted into (1)：
F=c1x1+c2x2+…+cnxn (3)
Wherein：
I=1,2 ..., s；
J=1,2 ..., n.
It is evaluated by business standing crisis grade and determines that business standing crisis grade includes the following steps：
Using Pre-Evaluation enterprise data as new samples, pre-processed by the mode identical with historical sample as normal data lattice
Formula；
The function model completed using formula (3) structure calculates comprehensive score, draws business standing crisis score；
By business standing crisis score compared with credit crisis grade, show that business standing crisis grade is evaluated, obtain enterprise
Industry credit crisis grade.
Divide credit grade：According to the distribution of the sample comprehensive score of calculating and credit crisis enterprise and the good enterprise of credit
Situation sets credit crisis grade according to certain standard and method；
New samples score is predicted using comprehensive function, evaluates business standing crisis grade.
Embodiment 3
For a kind of weight grade evaluation system, including：Sample data memory module, data collection station, sample number
It Data preprocess module, weight analysis module, opinion rating model determining module, opinion rating model memory module and comments
Valency grade acquisition module.Specifically,
First, sample data memory module：It is configured and adapted to store sample data in sampling memory；
2nd, data collection station：Its sample data for being configured and adapted to transfer from sampling memory is transmitted to information
Processing server；
3rd, sample data preprocessing module：It is configured and adapted to pre-process sample data；
4th, weight analysis module：It is configured and adapted to carry out weight point to the sample data by pretreatment
Analysis；
5th, opinion rating model determining module：It, which is configured and adapted to analyze by weight, obtains opinion rating model；
6th, opinion rating model memory module：It is configured and adapted to opinion rating model being stored in data storage；
7th, opinion rating acquisition module：It is configured and adapted to after data collection station collects pending data, will wait to locate
Data input opinion rating model is managed, obtains the opinion rating of pending data.
Weight analysis module includes：
1st, sample data comprehensive score computing module is configured to calculate sample data comprehensive score, according to comprehensive score
And data distribution state demarcation classification in sample data, set data level；Sample data comprehensive score computing module performs such as
Lower calculating：Test samples data are if appropriate for using weight factor approach；Weight is chosen using SPSS software defaults
Keeping characteristics value more than 1 factor method；Variance is rotated with orthogonal very big method, carries out factorial analysis；S weight before selection
Factor calculates comprehensive score by the weight coefficient of each factor according to the Return Law.Sample data comprehensive score computing module root
According to the distribution character of sample scores, classification point is selected, sets each data level.The distribution character of sample scores,
For linear distribution or just it is being distributed very much.
2nd, comprehensive score function structure module：It is configured to, by parameter coefficient, build comprehensive score function, calculates
Weight score.Comprehensive score function builds module construction anticipation function model, is specially：S weight is selected, then is estimated
Meter sample score function be：
F=a1z1+a2z2+…+aszs (1)
Wherein, F：Historical sample comprehensive score；
ai：The contribution rate of i-th of weight；
zi：I-th of weight.
Parameter coefficient is：
Wherein：I, l=1,2 ... s；
KiFor the ith feature value arranged from big to small；
bilFor weight ziTo zlAccumulation contribution rate.
I-th of weight can show as the linear combination of n original index, then：
zi=bi1x1+ai2x2+…+ainxn (2)
(2) are substituted into (1)：
F=c1x1+c2x2+…+cnxn (3)
Wherein：
I=1,2 ..., s；
J=1,2 ..., n.
Weight level determination module determines the weight grade of pending data by the following method：
Build information collecting platform, the pending data of receive information acquisition platform transmission, using pending data as new
Sample pre-processes new samples data for standard data format by the processing mode identical with sample data；
The function model completed using formula (3) structure calculates comprehensive score, show that the weight of new samples data obtains
Point；
By the weight score of new samples data compared with sample data weight grade, show that new samples data are weighed
The grade evaluation of weight factor obtains the weight grade of pending data.
Embodiment 4
A kind of weight grade evaluation system of the present invention is applied in business standing crisis evaluation system, is specifically included：
Sample chooses module, sample data preprocessing module, weight analysis module, business standing crisis grade evaluation module, enterprise
Industry credit crisis level determination module.
Sample chooses module, and it is good to be configured and adapted to choose a certain number of credit crisis enterprises and credit in target zone
Good enterprise forms historical sample；Choose (certain industry or a certain region) a certain number of credit crisis enterprises in target zone
The historical sample of enterprise's composition good with credit, wherein good enterprise of credit crisis enterprise and credit need predefined, sample choosing
The credit line that modulus block determines the enterprise for forming historical sample according to standard data format, be divided into credit crisis enterprise or
The good enterprise of credit.Total sample is randomly divided into two groups, one group is training sample group, and another set is test sample group.Training
The data of sample group are used for structure forecast model, and the data that test sample is rented are used to verify the significant degree of prediction model.
Sample data preprocessing module：It is configured and adapted to pre-process the initial data of historical sample；To ensure
The accuracy that weight method uses will pre-process initial data, mainly ensure that data have same direction
Property, and make achievement data that there are economic meanings.According to the thought, sample data is first classified according to default current standard,
Then the input that the ratio of the credit crisis enterprise in a certain concrete kind and the good enterprise of credit is calculated as weight is used
Data.This not only ensure that each achievement data has same directionality, and have economic implications.For example, for industry index,
It is classified as farming, forestry, husbandary and fishing, medical and health, building materials, metallurgical mineral products, 5 specific categories of other grades, farming, forestry, husbandary and fishing in sample
In creditable crisis enterprise 60, the good enterprise of credit 34, then this input data be set to 1.765 (60/34), remaining is similar
Processing.
Weight analysis module：It is configured and adapted to carry out weight point to the historical sample data by pretreatment
Analysis；Historical sample data is examined if appropriate for using weight factor approach；When carrying out weight calculating, check number first
According to if appropriate for weight factor approach is used, the instrument for examining this method is mainly that KMO and Barrelett sphericitys are examined,
The use of weight factor scheme is suitable when KMO values verify as notable more than 0.5 sphericity.KMO：Kaiser-Meyer-
Olkin, test statistics are the indexs for simple correlation coefficient between comparison variable and partial correlation coefficient.KMO statistics are to take
Value is between zero and one.When the simple correlation coefficient quadratic sum between all variables is far longer than partial correlation coefficient quadratic sum, KMO values
Close to 1.KMO values closer to 1, it is meant that the correlation between variable is stronger, and original variable is more suitable as factorial analysis；When all
When simple correlation coefficient quadratic sum between variable is close to 0, KMO values are close to 0.KMO values closer to 0, it is meant that the phase between variable
Weaker, the original more uncomfortable cooperation factorial analysis of variable of closing property.It is suitable judging historical sample data using weight factor scheme
Condition, on the one hand, to calculate historical sample comprehensive score, the keeping characteristics value that SPSS software defaults may be employed in ingredient selection is big
In the method for 1 weight, variance is rotated with orthogonal very big method, carries out factorial analysis, s weight, passes through before selection
The weight coefficient of each factor simultaneously calculates last comprehensive score according to the Return Law.According to the distribution character of sample scores
(may be linear distribution or just too distribution), selectes classification point, sets each credit crisis grade.On the other hand, Wo Menshe
Building anticipation function model is：S weight is selected, then the score function of sample estimates is：
F=a1z1+a2z2+…+aszs (1)
Wherein, F：Historical sample comprehensive score；
ai：The contribution rate of i-th of weight；
zi：I-th of weight.
Parameter coefficient is：
Wherein：I, l=1,2 ... s；
KiFor the ith feature value arranged from big to small；
bilFor weight ziTo zlAccumulation contribution rate.
I-th of weight can show as the linear combination of n original index, then：
zi=bi1x1+ai2x2+…+ainxn (2)
(2) are substituted into (1)：
F=c1x1+c2x2+…+cnxn (3)
Wherein：
I=1,2 ..., s；
J=1,2 ..., n.
It is evaluated by business standing crisis grade and determines that business standing crisis grade includes the following steps：
Using Pre-Evaluation enterprise data as new samples, pre-processed by the mode identical with historical sample as normal data lattice
Formula；
The function model completed using formula (3) structure calculates comprehensive score, draws business standing crisis score；
By business standing crisis score compared with credit crisis grade, show that business standing crisis grade is evaluated, obtain enterprise
Industry credit crisis grade.
Divide credit grade：According to the distribution of the sample comprehensive score of calculating and credit crisis enterprise and the good enterprise of credit
Situation sets credit crisis grade according to certain standard and method；
New samples score is predicted using comprehensive function, evaluates business standing crisis grade.
Business standing crisis grade evaluation module：It is analyzed by weight and obtains the evaluation of business standing crisis grade；
Business standing crisis level determination module：It is evaluated by business standing crisis grade and determines business standing crisis etc.
Grade.
A kind of weight grade evaluation method of the present invention and system are acquiring magnanimity government department data, enterprises
On the basis of the business crisis relevant information such as data and public sentiment data, with weight analytic approach, by business standing crisis
Multiple evaluation indexes be converted into a few overall target, it can be achieved that quick succinct evaluation calculates, compensate for traditional evaluation
Method ignores the deficiency of incidence relation between index in evaluation, so as to accurately draw weight grade evaluation result,
And effective early warning and management are carried out to business standing crisis, the credit situation of enterprise is shown to all-dimensional multi-angle, so as to political affairs
Mansion provides reference for business standing deciding grade and level, perfects Credit System Construction for next step each place, realizes " combine and discipline as a warning ", establish
Service basic and technical foundation.
A kind of weight grade evaluation method of the present invention and system introduce weight analytic approach and come to weight etc.
Grade evaluation index carries out dimensionality reduction.In weight Grade, how the index coefficient of core composite evaluation function obtains
To being key factor, which must at least is fulfilled for the following conditions, could preferably draw accurate credit crisis evaluation
As a result:
1st, can preferably change with sample changed；
2nd, objectivity；
3rd, the reasonability of evaluation result can be ensured.
For these conditions, the present invention obtains coefficient in itself using weight factor scheme based entirely on data, so as to compared with
Accurately to draw weight grade evaluation result.
A kind of weight grade evaluation method of the present invention and system are not required to artificially estimate weight, reduce artificial in evaluation
Factor influences, better than traditional evaluation method in accuracy.In addition, the present invention can be according to the Character adjustment index body of different industries
System has certain scalability.
The above described is only a preferred embodiment of the present invention, not the structure of the present invention is made in any form
Limitation.Any simple modification, equivalent change and modification that every technical spirit according to the invention makees above example,
Belong in the range of technical scheme.
Claims (10)
1. a kind of weight grade evaluation method, including being stored in sample data in sampling memory, data collection station from
Sample data is transferred in sampling memory, which is characterized in that
The sample data transferred from sampling memory is transferred to netscape messaging server Netscape by data collection station；
Netscape messaging server Netscape pre-processes sample data；
Netscape messaging server Netscape carries out weight analysis to the sample data by pretreatment；
It is analyzed by weight and obtains opinion rating model；
Opinion rating model is stored in data storage；
Data collection station collects pending data, and pending data is inputted opinion rating model, obtains pending data
Opinion rating.
2. weight grade evaluation method according to claim 1, which is characterized in that the sample data includes government
Division data storehouse data, inside data of enterprise and public sentiment data.
3. weight grade evaluation method according to claim 1, it is characterised in that：Weight analysis includes：
Sample data comprehensive score is calculated, according to data distribution state demarcation classification in comprehensive score and sample data, sets number
According to grade；
By parameter coefficient, comprehensive score function is built, calculates weight score.
4. weight grade evaluation method according to claim 2, it is characterised in that：Calculate sample data comprehensive score
Further comprise：
Test samples data are if appropriate for using weight factor approach；
Weight chooses the method using factor of the keeping characteristics value of SPSS software defaults more than 1；
Variance is rotated with orthogonal very big method, carries out factorial analysis；
S weight before selection calculates comprehensive score by the weight coefficient of each factor according to the Return Law.
5. weight grade evaluation method according to claim 4, it is characterised in that：According to point of sample scores
Cloth characteristic selectes classification point, sets each data level.
6. weight grade evaluation method according to claim 5, it is characterised in that：The distribution of sample scores is special
Property, for linear distribution or just it is distributed very much.
7. weight grade evaluation method according to claim 3, which is characterized in that structure comprehensive score function, into
One step includes：Anticipation function model is built, selects s weight, then the score function of sample estimates is：
F=a1z1+a2z2+…+aSzS (1)
Wherein, F：Historical sample comprehensive score；
ai：The contribution rate of i-th of weight；
zi：I-th of weight.
8. weight grade evaluation method according to claim 7, it is characterised in that：
Parameter coefficient is：
<mrow>
<msub>
<mi>a</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<mfrac>
<msub>
<mi>K</mi>
<mi>i</mi>
</msub>
<mrow>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>m</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>i</mi>
</msubsup>
<msub>
<mi>K</mi>
<mi>m</mi>
</msub>
</mrow>
</mfrac>
<msub>
<mi>b</mi>
<mrow>
<mi>i</mi>
<mi>l</mi>
</mrow>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>m</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>l</mi>
</msubsup>
<msub>
<mi>K</mi>
<mi>m</mi>
</msub>
</mrow>
<mrow>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>m</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>i</mi>
</msubsup>
<msub>
<mi>K</mi>
<mi>m</mi>
</msub>
</mrow>
</mfrac>
</mrow>
Wherein：I, l=1,2 ... s；
KiFor the ith feature value arranged from big to small；
bi1For weight ziTo z1Accumulation contribution rate.
9. weight grade evaluation method according to claim 8, it is characterised in that：I-th of weight can be with table
It is now the linear combination of n original index, then：
zi=bi1x1+ai2x2+…+ainxn (2)
(2) are substituted into (1)：
F=c1x1+c2x2+…+cnxn (3)
Wherein：
1=1,2 ..., S；
J=1,2 ..., n.
10. a kind of weight grade evaluation system, including：Sample data memory module：It is configured and adapted in samples storage
Sample data is stored in device, which is characterized in that further include with lower module：
Data collection station：Its sample data for being configured and adapted to transfer from sampling memory is transmitted to information processing services
Device；
Sample data preprocessing module：It is configured and adapted to pre-process sample data；
Weight analysis module：It is configured and adapted to carry out weight analysis to the sample data by pretreatment；
Opinion rating model determining module：It, which is configured and adapted to analyze by weight, obtains opinion rating model；
Opinion rating model memory module：It is configured and adapted to opinion rating model being stored in data storage；
Opinion rating acquisition module：It is configured and adapted to after data collection station collects pending data, by pending data
Opinion rating model is inputted, obtains the opinion rating of pending data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711132881.2A CN108053094A (en) | 2017-11-15 | 2017-11-15 | A kind of weight grade evaluation method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711132881.2A CN108053094A (en) | 2017-11-15 | 2017-11-15 | A kind of weight grade evaluation method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108053094A true CN108053094A (en) | 2018-05-18 |
Family
ID=62119186
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711132881.2A Pending CN108053094A (en) | 2017-11-15 | 2017-11-15 | A kind of weight grade evaluation method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108053094A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108770014A (en) * | 2018-07-10 | 2018-11-06 | 广东工业大学 | Calculating appraisal procedure, system, device and the readable storage medium storing program for executing of network server |
| CN108846731A (en) * | 2018-06-28 | 2018-11-20 | 合肥右传媒科技有限公司 | A kind of assessment system of business standing degree |
| CN109168049A (en) * | 2018-09-03 | 2019-01-08 | 广州虎牙信息科技有限公司 | Grade evaluation method, device, storage medium and the server of programme televised live |
| CN109615240A (en) * | 2018-12-13 | 2019-04-12 | 联想(北京)有限公司 | A kind of data processing method, device and evaluation system |
| CN111367980A (en) * | 2020-03-05 | 2020-07-03 | 苏宁云计算有限公司 | Method and system for managing upstream tasks according to e-commerce indexes |
| CN112561277A (en) * | 2020-12-08 | 2021-03-26 | 爱信诺征信有限公司 | City credit index calculation system, city credit index calculation method, electronic device, and storage medium |
| CN112882958A (en) * | 2021-03-30 | 2021-06-01 | 中信银行股份有限公司 | Demand quality evaluation method, device, equipment and storage medium based on system credit system |
| CN117371861A (en) * | 2023-10-30 | 2024-01-09 | 深圳轻喜到家科技有限公司 | Digital-based household service quality intelligent analysis method and system |
| CN117371861B (en) * | 2023-10-30 | 2024-07-02 | 深圳轻喜到家科技有限公司 | Digital-based household service quality intelligent analysis method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102637357A (en) * | 2012-03-27 | 2012-08-15 | 山东大学 | Regional traffic state assessment method |
| CN103514566A (en) * | 2013-10-15 | 2014-01-15 | 国家电网公司 | Risk control system and method |
| KR20160141465A (en) * | 2015-06-01 | 2016-12-09 | 주식회사 에스씨엘 | Risk assessment system for personal information management system |
| CN106326473A (en) * | 2016-08-31 | 2017-01-11 | 国信优易数据有限公司 | Data mining method based on entropy weight algorithm and analytic hierarchy process and system thereof |
-
2017
- 2017-11-15 CN CN201711132881.2A patent/CN108053094A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102637357A (en) * | 2012-03-27 | 2012-08-15 | 山东大学 | Regional traffic state assessment method |
| CN103514566A (en) * | 2013-10-15 | 2014-01-15 | 国家电网公司 | Risk control system and method |
| KR20160141465A (en) * | 2015-06-01 | 2016-12-09 | 주식회사 에스씨엘 | Risk assessment system for personal information management system |
| CN106326473A (en) * | 2016-08-31 | 2017-01-11 | 国信优易数据有限公司 | Data mining method based on entropy weight algorithm and analytic hierarchy process and system thereof |
Non-Patent Citations (3)
| Title |
|---|
| 张灵莹: ""基于因子分析和聚类分析的不同行业中小企业信用等级问题研究"", 《2010 ETP/IITA 2010 INTERNATIONAL CONFERENCE ON MANAGEMENT SCIENCE AND ENGINEERING 》 * |
| 董莹，崔瑞雪: ""基于因子分析的红葡萄酒质量评价"", 《大连民族学院学报》 * |
| 陈然等: ""基于统计分析的葡萄酒的评价"", 《滁州学院学报》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108846731A (en) * | 2018-06-28 | 2018-11-20 | 合肥右传媒科技有限公司 | A kind of assessment system of business standing degree |
| CN108770014A (en) * | 2018-07-10 | 2018-11-06 | 广东工业大学 | Calculating appraisal procedure, system, device and the readable storage medium storing program for executing of network server |
| CN108770014B (en) * | 2018-07-10 | 2022-03-25 | 广东工业大学 | Calculation evaluation method, system and device of network server and readable storage medium |
| CN109168049A (en) * | 2018-09-03 | 2019-01-08 | 广州虎牙信息科技有限公司 | Grade evaluation method, device, storage medium and the server of programme televised live |
| CN109615240A (en) * | 2018-12-13 | 2019-04-12 | 联想(北京)有限公司 | A kind of data processing method, device and evaluation system |
| CN111367980A (en) * | 2020-03-05 | 2020-07-03 | 苏宁云计算有限公司 | Method and system for managing upstream tasks according to e-commerce indexes |
| CN111367980B (en) * | 2020-03-05 | 2022-11-11 | 苏宁云计算有限公司 | Method and system for managing upstream tasks according to E-commerce indexes |
| CN112561277A (en) * | 2020-12-08 | 2021-03-26 | 爱信诺征信有限公司 | City credit index calculation system, city credit index calculation method, electronic device, and storage medium |
| CN112561277B (en) * | 2020-12-08 | 2024-02-20 | 爱信诺征信有限公司 | Urban credit index calculation system, method, electronic device and storage medium |
| CN112882958A (en) * | 2021-03-30 | 2021-06-01 | 中信银行股份有限公司 | Demand quality evaluation method, device, equipment and storage medium based on system credit system |
| CN117371861A (en) * | 2023-10-30 | 2024-01-09 | 深圳轻喜到家科技有限公司 | Digital-based household service quality intelligent analysis method and system |
| CN117371861B (en) * | 2023-10-30 | 2024-07-02 | 深圳轻喜到家科技有限公司 | Digital-based household service quality intelligent analysis method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108053094A (en) | A kind of weight grade evaluation method and system | |
| Mahamid et al. | Risks leading to cost overrun in building construction from consultants’ perspective | |
| Chen et al. | Project delivery system selection of construction projects in China | |
| US20170308934A1 (en) | Management method of power engineering cost | |
| Puška et al. | Model for investment decision making by applying the multi-criteria analysis method | |
| CN106372952A (en) | Objective and subjective weight determining multi-model compositional verification-based enterprise credit assessment method and system | |
| CN109993223A (en) | Pavement Condition prediction technique, storage medium, electronic equipment | |
| CN104766215B (en) | A kind of comprehensive, various dimensions owner of cargo selects quantization method | |
| CN104318339A (en) | Power grid investment benefit evaluation method based on comprehensive empowerment method | |
| CN104837184A (en) | Heterogeneous wireless network selection method based on interval triangular fuzzy number | |
| CN107609771A (en) | A kind of supplier's value assessment method | |
| CN108876487A (en) | A kind of industrial plot estimation method based on big data and intelligent decision mechanism | |
| CN108734567A (en) | A kind of asset management system and its appraisal procedure based on big data artificial intelligence air control | |
| CN105224997A (en) | A kind of planning load forecasting method containing multiple uncertain factor impact | |
| CN103839176A (en) | New energy enterprise value assessment model based on analytic hierarchy process | |
| CN108122166A (en) | The method sorted out to stock and scored using intelligent quantization financial technology | |
| Ganiyu et al. | Project cost prediction model using principal component regression for public building projects in Nigeria | |
| CN114881485A (en) | Enterprise fund risk assessment method based on analytic hierarchy process and cloud model | |
| CN108830405B (en) | Real-time power load prediction system and method based on multi-index dynamic matching | |
| CN106203864A (en) | A kind of brand assets appraisal procedure based on big data and system | |
| Aboshady et al. | A fuzzy risk management framework for the Egyptian real estate development projects | |
| Aleksandar et al. | Multiple linear regression model for predicting bidding price | |
| CN115564265A (en) | Power grid enterprise digital transformation evaluation method based on evaluation model | |
| Gao et al. | Bidding strategy with forecast technology based on support vector machine in the electricity market | |
| CN114266478A (en) | Power transformation operation standardization comprehensive evaluation method and related device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180518 |