CN107451834A - A kind of method improved based on principal component and K L information Contents Methods structure economic index - Google Patents
A kind of method improved based on principal component and K L information Contents Methods structure economic index Download PDFInfo
- Publication number
- CN107451834A CN107451834A CN201610371884.0A CN201610371884A CN107451834A CN 107451834 A CN107451834 A CN 107451834A CN 201610371884 A CN201610371884 A CN 201610371884A CN 107451834 A CN107451834 A CN 107451834A
- Authority
- CN
- China
- Prior art keywords
- trend
- uniformity
- initial
- data
- consistency
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
Landscapes
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention relates to a kind of method improved based on principal component and K L information Contents Methods structure economic index, it includes:It is determined that the data item with initial trend correlation connection;At least two data parameters are determined according to the data item, and weighted value is determined for each data parameters at least two data parameters;Each data parameters and corresponding weighted value at least two data parameters calculate at least one initial trend point, to form initial trend;Initial trend is fitted to generate the timing diagram of initial trend;Determined according to the timing diagram of initial trend for building at least two parameters of consistency of uniformity trend, and determine the weighted value of each parameter of consistency at least two parameter of consistency;Each parameter of consistency and weighted value at least two parameters of consistency calculate at least one uniformity trend point, to form uniformity trend;And uniformity trend is fitted to generate the timing diagram of uniformity trend.
Description
Technical field
The present invention relates to field of information processing, and relate more specifically to a kind of K-L information Contents Methods that are based on and carry out at data
The method and apparatus of reason.
Background technology
At present, periodic expansion and contraction, and periodically flourishing and decline is always presented in economic development.Generally,
In the different phase of economic cycle, different industries shows different vitality.This rule is referred to as economic cycle.It is economical
Scholar reflects whether economy is flourishing with consumer confidence index, and divides economic cycle by consumer confidence index, so as to understand country, industry warp
Ji state of development, people are instructed to carry out economic activity.Although country has promulgated consumer confidence index, economist, statistics
Scholar and relevant speciality personage etc. would generally construct consumer confidence index based on data, so as to understand national economic development situation.
High ferro plum, Gu Yu and Wang Zhe construct Chinese exports consumer confidence index using principal component analysis.Chinese exports boom refers to
Number calculates growth sequence for the economic data of in August, -2006 in January nineteen ninety-five, and carries out seasonal adjustment, then utilizes the time difference
A variety of methods such as relevant function method, K-L information Contents Methods, peak valley correspondent method filter out 15 menstrual period indexs, to respectively constitute China
Leading, consistent, lagging indicator group.Due to having the preparation of the financial knowledge of abundance, data and the integrality of scheme, acquired results
It is ideal.But data analysis and actual conditions have inconsistent place unavoidably.Since the shaping that make use of country to promulgate
Beforehand index, also to give a strained interpretation can bring a little contradiction from data.In principal component implementation process, beforehand index and
Threshold value is inconsistent used by the variance contribution ratio of the interception principal component of same index, and does not also provide solution to this scheme
Release.
Chen Kejia, Liu Sifeng remain Consumer Prices index, retail sales index, stock price index 3
Index, and remaining 21 index is subjected to seasonal adjustment using Tramo/Seats methods, with the variation of industrial added value and boom
It is foundation that circulation is consistent, and leading, consistent and lagging indicator is screened using K-L information Contents Methods.Economy is referred to using K-L information Contents Methods
Mark carries out step-out time analysis, and Tramo/Seats methods take into account the factor such as national conditions, the Spring Festival of China, preferably extract
Boom circulation.
Basic thoughts of the Ren Rongrong in reference National Bureau of Economic Research in 2011 to Economic Climate Analysis, using diffusion
The method that index is combined with composite index number, the operation of Chinese real-estate market is analyzed.Wherein, diffusion index is used to judge
Economic activity is in prosperous or stagnant state, and composite index number is used to judge that economic activity is in prosperous rising or lower depression of order
Section.Acquired results have also proved the actual effect that the national stable room rate policy used is brought.
But currently existing scheme does not have the data Seasonal for considering economy, and it therefore can not truly reflect economy
Moving law.
The content of the invention
In order to solve the above problems, consumer confidence index is built present invention improves over principal component and K-L information Contents Methods is used before
Method.Technical scheme is based on data and after standard is set, and relaxes the meaning of data in itself, values last
As a result it is whether accurate.The criterion of screening index of the present invention is to utilize K-L information Contents Methods, and strictly from data, based on x-
11 model separation seasonal factors are circulated using (HP filtering methods make reference) separation long-term trend are returned with prosperous, are obtained boom and are referred to
While number, the long-term trend that are also quantified are convenient to recycle.K-L information Contents Methods are uncomplicated, but amount of calculation is larger,
The automation that the present invention passes through programming realization K-L information Contents Methods.
The present invention is based on x-11 addition models, and seasonality has been verified whether using variance analysis, then decides whether to eliminate warp
Help data Seasonal, circulates 2 months as standard using CPI (consumer price index) leading economic, utilizes K-L information content
Method calculates the economic time difference, calculates consumer confidence index with principal component, long-term economic trend is finally calculated in a manner of recurrence, pulls out scape
Gas circulates, and the consumer confidence index correlation that the consumer confidence index drawn is announced with associated companies reaches 0.85.
According to an aspect of the present invention, there is provided one kind is improved based on principal component and K-L information Contents Methods structure economic index
Method, methods described includes:
It is determined that the data item with initial trend correlation connection;
At least two data parameters are determined according to the data item, and are each at least two data parameters
Data parameters determine weighted value;
Each data parameters and corresponding weighted value at least two data parameters are at least one to calculate
Initial trend point, to form initial trend;
Initial trend is fitted to generate the timing diagram of initial trend;
At least two parameters of consistency for building uniformity trend are determined according to the timing diagram of initial trend, and really
The weighted value of each parameter of consistency in fixed at least two parameter of consistency;
Each parameter of consistency and corresponding weighted value at least two parameters of consistency calculate at least
One uniformity trend point, to form uniformity trend;And
Uniformity trend is fitted to generate the timing diagram of uniformity trend.
Preferably, wherein the weighted value is the variance contribution of each data parameters at least two data parameters
Rate.
Preferably, wherein each data parameters include at least one data item.
Preferably, it is described initial trend is fitted using generate the timing diagram of initial trend as:Returned using simple linear
Return or log-linear regression be fitted to initial trend to generate the timing diagram of initial trend.
Preferably, in addition to according to the timing diagram of uniformity trend determine the rise sections of data, stagflation section and under
Section drops.
According to another aspect of the present invention, there is provided one kind is improved based on principal component and K-L information Contents Methods structure economic index
Equipment, the equipment includes:
Data item determining unit, it is determined that the data item with initial trend correlation connection;
Data parameters determining unit, at least two data parameters are determined according to the data item, and be described at least two
Each data parameters in individual data parameters determine weighted value;
Trend point computing unit, each data parameters and corresponding weighted value at least two data parameters
To calculate at least one initial trend point;
Initial trend generation unit, initial trend is formed according at least one initial trend point, and to initial
Trend is fitted to generate the timing diagram of initial trend;
Parameter of consistency generation unit, according to the timing diagram of initial trend determine for build uniformity trend at least two
Individual parameter of consistency, and determine the weighted value of each parameter of consistency at least two parameter of consistency;
Uniformity trend point generation unit, each parameter of consistency at least two parameters of consistency and relative
The weighted value answered calculates at least one uniformity trend point;And
Uniformity trend generation unit, uniformity trend is formed according at least one uniformity trend point, and it is right
Uniformity trend is fitted to generate the timing diagram of uniformity trend.
Preferably, wherein the weighted value is the variance contribution of each data parameters at least two data parameters
Rate.
Preferably, wherein each data parameters include at least one data item.
Preferably, it is described initial trend is fitted using generate the timing diagram of initial trend as:Returned using simple linear
Return or log-linear regression be fitted to initial trend to generate the timing diagram of initial trend.
Preferably, in addition to according to the timing diagram of uniformity trend determine the rise sections of data, stagflation section and under
Section drops.
The advantages of technical scheme is that the acquisition of data is convenient first, relaxes anticipate economic in itself to initial data
The requirement of justice, but focus on the accuracy of end product, so as to which available data are more.What technical scheme used
Data are the macroscopical Finance data announced in ripe database and ripe software, and attention builds boom with principal component
Accessory substance-long-term trend of index, it is finally to realize to calculate the data time difference with K-L information Contents Methods by c Programming with Pascal Language.
Brief description of the drawings
By reference to the following drawings, the illustrative embodiments of the present invention can be more fully understood by:
Fig. 1 is the flow chart according to the data processing method of the preferred embodiment for the present invention;
Fig. 2 is the schematic diagram according to the data item trend of the preferred embodiment for the present invention;
Fig. 3 is the timing diagram according to the initial trend of the preferred embodiment for the present invention;
Fig. 4 is the schematic diagram according to the original uniformity trend of the preferred embodiment for the present invention;
Fig. 5 is the residual error that trend is eliminated according to the linear model of the uniformity trend of the synthesis of the preferred embodiment for the present invention
Figure;
Fig. 6 is the uniformity trend according to synthesized by the preferred embodiment for the present invention and the comparison signal of existing tendency chart
Figure;
Fig. 7 is the trend schematic diagram that mask data is filtered according to the HP of the preferred embodiment for the present invention;And
Fig. 8 is the structural representation according to the data processing equipment of the preferred embodiment for the present invention.
Embodiment
The illustrative embodiments of the present invention are introduced with reference now to accompanying drawing, however, the present invention can use many different shapes
Formula is implemented, and is not limited to embodiment described herein, there is provided these embodiments are to disclose at large and fully
The present invention, and fully pass on the scope of the present invention to person of ordinary skill in the field.Show for what is be illustrated in the accompanying drawings
Term in example property embodiment is not limitation of the invention.In the accompanying drawings, identical cells/elements are attached using identical
Icon is remembered.
Unless otherwise indicated, term (including scientific and technical terminology) used herein has to person of ordinary skill in the field
It is common to understand implication.Further it will be understood that the term limited with usually used dictionary, be appreciated that and its
The linguistic context of association area has consistent implication, and is not construed as Utopian or overly formal meaning.
Fig. 1 is the flow chart according to the data processing method 100 of the preferred embodiment for the present invention.As shown in figure 1, at data
The method that reason method 100 builds consumer confidence index before improving with principal component and K-L information Contents Methods.Data processing method 100 is logical
Cross and circulated using recurrence (HP filtering methods) separation long-term trend and boom, while obtaining consumer confidence index, and also obtained
The long-term trend of quantization, recycled so as to convenient.Generally, K-L information Contents Methods are uncomplicated, but amount of calculation is larger, data processing
Method 100 realizes the calculation automation of K-L information Contents Methods by computer program.
As shown in figure 1, data processing method 100 is since step 101 place.At step 101, it is determined that with initial trend phase
The data item of association.Generally, the major parameter for influenceing economy is, for example, generated energy Power, output of steel Ste, narrow definition of money
Measure M1, total import value Inp, raw coal output Coal, total retail sales of consumer goods Retail and manufacturing industry PMI etc..Preferably,
Data processing method 100 can be by generated energy Power, output of steel Ste, narrow money supply M1, total import value Inp, raw coal
One or more of yield Coal, total retail sales of consumer goods Retail and manufacturing industry PMI are used as data item, can also claim
For beforehand index.Preferably, these data item and the initial trend correlation to be obtained join.Generally, determine that data item (refers in advance
Number) it is the basis for analyzing economic trend.If selected data item (beforehand index) is relatively more accurate, believe by K-L
Result after breath amount method is handled can preferably reflect Economic Operation.
Preferably, at step 102, at least two data parameters are determined according to the data item, and for described at least
Each data parameters in two data parameters determine weighted value.Preferably, data processing method 100 is made with two data parameters
Illustrated for example, i.e. illustrated using two principal components.Table 1 is the Correlation Matrix list of feature values of beforehand index.Such as table 1
It is shown, the characteristic values of five principal components be respectively 5.89570355,0.94183987,0.08583071,0.0445344 and
0.01903714.The variance of five principal components be 4.95386368,0.85600916,0.04129631,0.02549726 with
And 0.010775.The variance contribution ratio of five principal components is 0.8422,0.1345,0.0123,0.0064 and 0.0027.
Table 1
Preferably, table 2 is characterized vector table.Table 2 shows generated energy Power, output of steel Ste, narrow money supply
M1, total import value Inp, raw coal output Coal, total retail sales of consumer goods Retail and manufacturing industry PMI characteristic vector
Value.For example, in first principal component, generated energy Power characteristic vector value is 0.408981, output of steel Ste characteristic vector value
Characteristic vector value for 0.405487, narrow money supply M1 is 0.405659, total import value Inp characteristic vector value is
0.39249th, raw coal output Coal characteristic vector value is 0.403539, total retail sales of consumer goods Retail characteristic vector
It is 0.151395 to be worth for 0.404899 and manufacturing industry PMI characteristic vector value.For example, in Second principal component, generated energy Power
Characteristic vector value be -0.023181, output of steel Ste characteristic vector value is -0.10433, narrow money supply M1 spy
Sign vector value is -0.130746, total import value Inp characteristic vector value is 0.160181, raw coal output Coal characteristic vector
The characteristic vector value being worth for -0.14448, total retail sales of consumer goods Retail is the spy of -0.109649 and manufacturing industry PMI
It is 0.955473 to levy vector value.
Table 2
Prin1 | Prin2 | Prin3 | Prin4 | Prin5 | Prin6 | Prin7 | |
Power | 0.408981 | -0.023181 | 0.114113 | -0.417112 | -0.053799 | -0.622972 | -0.504186 |
Ste | 0.405487 | -0.10433 | 0.292129 | -0.311413 | -0.621277 | 0.214561 | 0.458647 |
M1 | 0.405659 | -0.130746 | 0.08917 | 0.447054 | -0.166233 | 0.481845 | -0.592225 |
Inp | 0.39249 | 0.160181 | -0.879422 | -0.154803 | -0.000182 | 0.124825 | 0.085824 |
Coal | 0.403539 | -0.14448 | 0.255298 | -0.254342 | 0.759574 | 0.282831 | 0.171665 |
Retail | 0.404899 | -0.109649 | 0.013224 | 0.658865 | 0.074425 | -0.48631 | 0.384342 |
PMI | 0.151395 | 0.955473 | 0.234421 | 0.080154 | 0.031538 | 0.040282 | 0.012486 |
Preferably, calculated by taking two principal components as an example.Wherein
First principal component:Variance contribution ratio 84.22%
y1=0.408981*power+0.405487*ste+0.405659*M1+0.392490*inp+0.4 03539*
coal+0.404899*retail+0.151395*PMI
Second principal component,:Variance contribution ratio 13.45%
y2=-0.023181*power-0.10433*ste-0.130746*M1+0.160181*inp-0.14 448*
coal-0.109649*retail+0.955473PMI
It is comprehensive into an index, the index (beforehand index) prosperous as reflection that both are assigned into power.It typically, there are a variety of taxes
Power method, such as subjective multiple power, analytic hierarchy process (AHP) are weighed again.The variance contribution ratio obtained using principal component analysis is very suitable
Weight, therefore the present invention illustrates using variance contribution ratio as weight.
Preferably, at step 103, each data parameters and corresponding power at least two data parameters
Weight values calculate at least one initial trend point, to form initial trend.Data processing method 100 is counted using following formula
Calculate initial trend point.
Preferably, the tendency chart of data item is obtained according to initial trend point, as shown in Figure 2.Fig. 2 is excellent according to the present invention
Select the schematic diagram of the data item trend of embodiment.It is often necessary to by being further processed data item trend to obtain
Initial trend.
Preferably, at step 104, initial trend is fitted to generate the timing diagram of initial trend.Preferably, just
The timing diagram of beginning trend is as shown in figure 3, Fig. 3 is the timing diagram according to the initial trend of the preferred embodiment for the present invention.From Fig. 3
It can be seen that there is an obvious ascendant trend in index in 2009.This ascendant trend should be that country 4,000,000,000,000 invests band
The influence come, it can be seen that powerful driving power of government's macro policy to economy., can for the long-term trend in economic problems
To be handled using multi-method.The present invention is intended data item trend using simple linear regression and log-linear regression
Close.For example, using 2 months 2005 as 1 in abscissa, and using in March, 2005 as 2, by that analogy.
Preferably, if obtaining trend T using linear regression1:
Y=0.082193x-3.53428
If obtain trend T using logarithmic model2:
Y=e0.017671x+0.755009-5
Preferably, the data item constructed by data processing method 100 (that is, consumer confidence index) in data processing stage just
Remove seasonal factor S, then subtract this trend Ti, gained residual error is exactly economic cyclical factor C and random fluctuation I.Such as Fig. 3 institutes
Show, compare the timing diagram of residual error obtained by two methods, it can be found that difference is smaller.
Preferably, at step 105, according to the timing diagram of initial trend determine for build uniformity trend at least two
Individual parameter of consistency, and determine the weighted value of each parameter of consistency at least two parameter of consistency.Preferably, root
At least two parameters of consistency for building uniformity trend, i.e. coincident indicator are selected according to the timing diagram of initial trend.
Preferably, the index of the selection of data processing method 100 is:The sum of investments in fixed assets used adds up year-on-year x1, fixed assets are newly opened
Project of that month x2, light industry value added on year-on-year basis of that month x3, consumer price index CPI and index of consumer confidence CCI on year-on-year basis.
Preferably, data processing method 100 is illustrated using four parameters of consistency as example, i.e. using four uniformity it is main into
Divide and illustrate.Table 3 is the Correlation Matrix list of feature values of consistent row parameter.As shown in table 3, the characteristic value of four uniformity principal components
It is 2.23635697,1.21449337,0.66367198 and 0.52862867 respectively.The variance of four principal components is
1.0218636,0.55082138,0.13504331 and 0.17177966.The variance contribution ratio of four principal components is
0.4473rd, 0.2429,0.1327 and 0.1057.
Table 3
Preferably, table 4 is the characteristic vector table of parameter of consistency.Show that the sum of investments in fixed assets used adds up in table 4
Year-on-year x1, fixed assets newly open project of that month x2, light industry value added on year-on-year basis of that month x3, consumer price index CPI on year-on-year basis with
And index of consumer confidence CCI characteristic vector value.For example, in first principal component, the sum of investments in fixed assets used is accumulative same
Than x1 be -0.559658, fixed assets newly open project it is of that month x2 on year-on-year basis be -0.425855, light industry value added this month it is year-on-year x3
For 0.302168, consumer price index CPI be 0.547739 and index of consumer confidence CCI is 0.337798.Second
In principal component, the sum of investments in fixed assets used adds up that year-on-year x1 is 0.174462, fixed assets newly open the of that month x2 on year-on-year basis of project and be
0.450538th, the of that month x3 on year-on-year basis of light industry value added is 0.653526, consumer price index CPI is -0.174768 and is disappeared
The person's of expense confidence index CCI is 0.555822.In the 3rd principal component, the sum of investments in fixed assets used add up year-on-year x1 for-
0.281676th, fixed assets newly open the of that month x2 on year-on-year basis of project be 0.433395, the of that month x3 on year-on-year basis of light industry value added be
0.434026th, consumer price index CPI is 0.240019 and index of consumer confidence CCI is -0.697739.In the 4th master
In composition, it is that -0.152778, fixed assets newly open the of that month x2 on year-on-year basis of project and are that the sum of investments in fixed assets used, which adds up year-on-year x1,
0.640985th, the of that month x3 on year-on-year basis of light industry value added is -0.525553, consumer price index CPI is 0.45392 and consumption
Person's confidence index CCI is 0.289046.
Table 4
Prin1 | Prin2 | Prin3 | Prin4 | Prin5 | |
x1 | -0.559658 | 0.174462 | -0.281676 | -0.152778 | 0.744086 |
x2 | -0.425855 | 0.450538 | 0.433395 | 0.640985 | -0.130267 |
x3 | 0.302168 | 0.653526 | 0.434026 | -0.525553 | 0.130438 |
CPI | 0.547739 | -0.174768 | 0.240019 | 0.45392 | 0.637014 |
CCI | 0.337798 | 0.555822 | -0.697739 | 0.289046 | -0.081032 |
Preferably, calculated by taking four uniformity principal components as an example.Wherein
y1=-0.559568x1-0.425855x2+0.302168x3+0.547739*CPI+0.337798*CCI
y2=0.174462x1+0.450538x2+0.653526x3+-0.174768*CPI+0.555822*CCI
y3=-0.281676x1+0.433395x2+0.434026x3+0.240019*CPI-0.697739*CCI
y4=-0.152778x1+0.640985x2-0.525553x3+0.45392*CPI-0.289046*CCI
Preferably, at step 106, each parameter of consistency at least two parameters of consistency and corresponding
Weighted value calculate at least one uniformity trend point, to form uniformity trend.
Preferably, the timing diagram of uniformity trend is obtained according to uniformity trend point, as shown in Figure 4.Fig. 4 is according to this
The schematic diagram of the original uniformity trend of preferred embodiment of the invention.It is often necessary to by entering to original uniformity trend
Row is further handled to obtain uniformity trend.
Preferably, at step 107, uniformity trend is fitted to generate the timing diagram of uniformity trend.In Fig. 4
Original uniformity trend there is no obvious trend, it is therefore desirable to go out trend with logarithm method and linear regression fit.It can think
As this trend is to be similar to y=0 straight line and curve.Because the coincidence indicator of the inventive method is not specifically to measure,
But on year-on-year basis.
Preferably, trend is obtained using linear regression:
Y=0.082193x-3.53428
Then, trend then with linear regression method is eliminated, so as to obtain figure as shown in Figure 5.Fig. 5 is excellent according to the present invention
The linear model of the uniformity trend of the synthesis of embodiment is selected to eliminate the residual plot of trend.Wherein, residual error e>0 means economy
It is prosperous, residual error e<0 means economic depression.
Preferably, in order to determine that it is no reliable that the present invention is concluded that, by synthesized uniformity trend with it is existing become
Gesture figure is compared, as shown in Figure 6.Fig. 6 be uniformity trend according to synthesized by the preferred embodiment for the present invention with it is existing become
The comparison schematic diagram of gesture figure.By it was found that both approximate trends are identical, finally calculating the beforehand index and phase of oneself structure
The correlation for the beforehand index that Guan companies announce, coefficient correlation reach 0.85.Thus, it is possible to think the uniformity trend of the present invention
In the presence of the higher degree of accuracy.It is if different in detail, it is believed that it is the selection of data or index, and method
Caused by the problems such as reference.Preferably, may finally conclude that:Rise section 2005.3-2006.2,2007.2-
2008.3 and 2010.2-20124, stagflation section 2006.2-2006.3,2008.3-2008.9, last transition 2006.3-
2006.8th, 2008.9-2009.7, and recovery section 2006.8-2007.2,2009.7-2010.2.Preferably, it is more straight in Fig. 7
See ground and trend and the cycle that mask data is filtered with HP are presented with econometrics software kit (eviews).Fig. 7 is according to this
The trend schematic diagram of the HP filtering mask datas of preferred embodiment of the invention.
It shown below is the C language code that the time difference is calculated using K-L information Contents Methods:
Fig. 8 is the structural representation according to the data processing equipment 800 of the preferred embodiment for the present invention.As shown in figure 8, number
With principal component and the method for K-L information Contents Methods structure consumer confidence index before being improved according to processing equipment 800.Data processing equipment
800 are circulated by using (HP filtering methods) separation long-term trend are returned with prosperous, while obtaining consumer confidence index, and also
The long-term trend quantified have been arrived, have been recycled so as to convenient.Generally, K-L information Contents Methods are uncomplicated, but amount of calculation is larger, data
Processing equipment 800 realizes the calculation automation of K-L information Contents Methods by computer program.
As shown in figure 8, data processing equipment 800 includes:Data item determining unit 801, data parameters determining unit 802,
Trend point computing unit 803, initial trend generation unit 804, parameter of consistency generation unit 805, the generation of uniformity trend point
Unit 806 and uniformity trend generation unit 807.Data item determining unit 801 determines the data with initial trend correlation connection
.Generally, the major parameter for influenceing economy is, for example, that generated energy Power, output of steel Ste, narrow money supply M1, import are total
Volume Inp, raw coal output Coal, total retail sales of consumer goods Retail and manufacturing industry PMI etc..Preferably, data processing side
Method 100 can be by generated energy Power, output of steel Ste, narrow money supply M1, total import value Inp, raw coal output Coal, society
Meeting one or more of total volume of retail sales of consumer goods Retail and manufacturing industry PMI are used as data item, alternatively referred to as beforehand index.
Preferably, these data item and the initial trend correlation to be obtained join.Generally, it is that analysis is economical to determine data item (beforehand index)
The basis of development trend.If selected data item (beforehand index) is relatively more accurate, handled by K-L information Contents Methods
Result afterwards can preferably reflect Economic Operation.
Preferably, data parameters determining unit 802 determines at least two data parameters according to the data item, and is institute
The each data parameters stated at least two data parameters determine weighted value.Preferably, data processing equipment 800 is with two data
Parameter illustrates as example, i.e. is illustrated using two principal components.As it was previously stated, the characteristic value of five principal components point
It is not 5.89570355,0.94183987,0.08583071,0.0445344 and 0.01903714.The side of five principal components
Difference is 4.95386368,0.85600916,0.04129631,0.02549726 and 0.010775.The side of five principal components
Poor contribution rate is 0.8422,0.1345,0.0123,0.0064 and 0.0027.
For example, in first principal component, generated energy Power characteristic vector value is 0.408981, output of steel Ste spy
Levy vector value be 0.405487, narrow money supply M1 characteristic vector value is 0.405659, total import value Inp feature to
Value is 0.39249, raw coal output Coal characteristic vector value is 0.403539, total retail sales of consumer goods Retail spy
Sign vector value is 0.404899 and manufacturing industry PMI characteristic vector value is 0.151395.For example, in Second principal component, generate electricity
The characteristic vector value for measuring Power is -0.023181, output of steel Ste characteristic vector value is -0.10433, narrow money supply
M1 characteristic vector value is -0.130746, total import value Inp characteristic vector value is 0.160181, raw coal output Coal spy
Sign vector value is -0.14448, total retail sales of consumer goods Retail characteristic vector value is -0.109649 and manufacturing industry
PMI characteristic vector value is 0.955473.
Preferably, calculated by taking two principal components as an example.Wherein
First principal component:Variance contribution ratio 84.22%
y1=0.408981*power+0.405487*ste+0.405659*M1+0.392490*inp+0.4 03539*
coal+0.404899*retail+0.151395*PMI
Second principal component,:Variance contribution ratio 13.45%
y2=-0.023181*power-0.10433*ste-0.130746*M1+0.160181*inp-0.14 448*
coal-0.109649*retail+0.955473PMI
It is comprehensive into an index, the index (beforehand index) prosperous as reflection that both are assigned into power.It typically, there are a variety of taxes
Power method, such as subjective multiple power, analytic hierarchy process (AHP) are weighed again.The variance contribution ratio obtained using principal component analysis is very suitable
Weight, therefore the present invention illustrates using variance contribution ratio as weight.
Preferably, each data parameters of the trend point computing unit 803 at least two data parameters and relative
The weighted value answered calculates at least one initial trend point.Data processing equipment 800 is initially become using following formula to calculate
Gesture point.
Preferably, the tendency chart of data item is obtained according to initial trend point, as shown in Figure 2.It is often necessary to pass through logarithm
It is further processed according to item trend to obtain initial trend.
Preferably, initial trend generation unit 804 forms initial trend according at least one initial trend point, and
And initial trend is fitted to generate the timing diagram of initial trend.Preferably, the timing diagram of initial trend is as shown in Figure 3.
As can be seen from Figure 3 in 2009 there is an obvious ascendant trend in index.This ascendant trend should be country 40,000
The influence that hundred million investments are brought, it can be seen that powerful driving power of government's macro policy to economy.For the length in economic problems
Phase trend, it can be handled using multi-method.The present invention is become using simple linear regression and log-linear regression to data item
Gesture is fitted.For example, using 2 months 2005 as 1 in abscissa, and using in March, 2005 as 2, by that analogy.
Preferably, if obtaining trend T using linear regression1:
Y=0.082193x-3.53428
If obtain trend T using logarithmic model2:
Y=e0.017671x+0.755009-5
Preferably, the data item constructed by data processing equipment 800 (that is, consumer confidence index) in data processing stage just
Remove seasonal factor S, then subtract this trend Ti, gained residual error is exactly economic cyclical factor C and random fluctuation I.Such as Fig. 3 institutes
Show, compare the timing diagram of residual error obtained by two methods, it can be found that difference is smaller.
Preferably, parameter of consistency generation unit 804 determines to become for building uniformity according to the timing diagram of initial trend
At least two parameters of consistency of gesture, and determine the weight of each parameter of consistency at least two parameter of consistency
Value.Preferably, at least two parameters of consistency for building uniformity trend are selected according to the timing diagram of initial trend, i.e.,
Coincident indicator.Preferably, the index of the selection of data processing equipment 800 is:The sum of investments in fixed assets used adds up year-on-year x1, consolidated
Determine assets and newly open project of that month x2, light industry value added on year-on-year basis of that month x3, consumer price index CPI on year-on-year basis and consumer's letter
Cardiac index CCI.Preferably, data processing equipment 800 is illustrated using four parameters of consistency as example, i.e. uses four
Uniformity principal component illustrates.For example, the characteristic value of four uniformity principal components be 2.23635697 respectively,
1.21449337,0.66367198 and 0.52862867.The variance of four principal components be 1.0218636,0.55082138,
0.13504331 and 0.17177966.The variance contribution ratio of four principal components be 0.4473,0.2429,0.1327 and
0.1057。
For example, in first principal component, the sum of investments in fixed assets used adds up year-on-year x1 and is -0.559658, fixes
Assets newly open project this month, and x2 is -0.425855 on year-on-year basis, the of that month x3 on year-on-year basis of light industry value added is 0.302168, consumer price
Index CPI is 0.547739 and index of consumer confidence CCI is 0.337798.In Second principal component, investment in fixed assets
Completion volume add up year-on-year x1 be 0.174462, fixed assets newly open the of that month x2 on year-on-year basis of project be 0.450538, light industry value added
It is of that month that x3 is 0.653526 on year-on-year basis, consumer price index CPI is -0.174768 and index of consumer confidence CCI is
0.555822.In the 3rd principal component, the sum of investments in fixed assets used adds up year-on-year x1 and newly opened for -0.281676, fixed assets
The of that month x2 on year-on-year basis of project is 0.433395, the of that month x3 on year-on-year basis of light industry value added is 0.434026, consumer price index CPI is
0.240019 and index of consumer confidence CCI is -0.697739.In the 4th principal component, the sum of investments in fixed assets used is tired out
Count year-on-year x1 be -0.152778, fixed assets newly open the of that month x2 on year-on-year basis of project be 0.640985, light industry value added it is of that month on year-on-year basis
X3 is -0.525553, consumer price index CPI is 0.45392 and index of consumer confidence CCI is 0.289046.
Preferably, calculated by taking four uniformity principal components as an example.Wherein
y1=-0.559568x1-0.425855x2+0.302168x3+0.547739*CPI+0.337798*CCI
y2=0.174462x1+0.450538x2+0.653526x3+-0.174768*CPI+0.555822*CCI
y3=-0.281676x1+0.433395x2+0.434026x3+0.240019*CPI-0.697739*CCI
y4=-0.152778x1+0.640985x2-0.525553x3+0.45392*CPI-0.289046*CCI
Preferably, each uniformity ginseng of the uniformity trend point generation unit 806 at least two parameters of consistency
Several and corresponding weighted values calculates at least one uniformity trend point.
Preferably, the timing diagram of uniformity trend is obtained according to uniformity trend point.It is often necessary to by original
Uniformity trend is further processed to obtain uniformity trend.
Preferably, uniformity trend generation unit 807, uniformity is formed according at least one uniformity trend point and become
Gesture, and uniformity trend is fitted to generate the timing diagram of uniformity trend.Original uniformity trend is not obvious
Trend, it is therefore desirable to go out trend with logarithm method and linear regression fit.It is envisioned that this trend is be similar to y=0 straight
Line and curve.Because the coincidence indicator of the inventive method is not specifically to measure, but on year-on-year basis.
Preferably, trend is obtained using linear regression:
Y=0.082193x-3.53428
Then, trend then with linear regression method is eliminated, so as to obtain figure as shown in Figure 5.Fig. 5 is excellent according to the present invention
The linear model of the uniformity trend of the synthesis of embodiment is selected to eliminate the residual plot of trend.Wherein, residual error e>0 means economy
It is prosperous, residual error e<0 means economic depression.
Preferably, in order to determine that it is no reliable that the present invention is concluded that, by synthesized uniformity trend with it is existing become
Gesture figure is compared, as shown in Figure 6.Fig. 6 be uniformity trend according to synthesized by the preferred embodiment for the present invention with it is existing become
The comparison schematic diagram of gesture figure.By it was found that both approximate trends are identical, finally calculating the beforehand index and phase of oneself structure
The correlation for the beforehand index that Guan companies announce, coefficient correlation reach 0.85.Thus, it is possible to think the uniformity trend of the present invention
In the presence of the higher degree of accuracy.It is if different in detail, it is believed that it is the selection of data or index, and method
Caused by the problems such as reference.Preferably, may finally conclude that:Rise section 2005.3-2006.2,2007.2-
2008.3 and 2010.2-20124, stagflation section 2006.2-2006.3,2008.3-2008.9, last transition 2006.3-
2006.8th, 2008.9-2009.7, and recovery section 2006.8-2007.2,2009.7-2010.2.Preferably, it is more straight in Fig. 7
See ground and trend and the cycle that mask data is filtered with HP are presented with econometrics software kit (eviews).Fig. 7 is according to this
The trend schematic diagram of the HP filtering mask datas of preferred embodiment of the invention.
The present invention is described by reference to a small amount of embodiment.However, it is known in those skilled in the art, as
What subsidiary Patent right requirement was limited, except the present invention other embodiments disclosed above equally fall the present invention's
In the range of.
Normally, all terms used in the claims are all solved according to them in the usual implication of technical field
Release, unless clearly being defined in addition wherein.All references " one/described/be somebody's turn to do [device, component etc.] " are all opened ground
At least one example being construed in described device, component etc., unless otherwise expressly specified.Any method disclosed herein
Step need not all be run with disclosed accurately order, unless explicitly stated otherwise.
Claims (10)
1. a kind of method improved based on principal component and K-L information Contents Methods structure economic index, methods described include:
It is determined that the data item with initial trend correlation connection;
At least two data parameters are determined according to the data item, and are each data at least two data parameters
Parameter determines weighted value;
Each data parameters and corresponding weighted value at least two data parameters are at least one initial to calculate
Trend point, to form initial trend;
Initial trend is fitted to generate the timing diagram of initial trend;
Determined according to the timing diagram of initial trend for building at least two parameters of consistency of uniformity trend, and determine institute
State the weighted value of each parameter of consistency at least two parameters of consistency;
Each parameter of consistency and corresponding weighted value at least two parameters of consistency is at least one to calculate
Uniformity trend point, to form uniformity trend;And
Uniformity trend is fitted to generate the timing diagram of uniformity trend.
2. according to the method for claim 1, wherein the weighted value is every number at least two data parameters
According to the variance contribution ratio of parameter.
3. according to the method for claim 1, wherein each data parameters include at least one data item.
4. according to the method for claim 1, it is described initial trend is fitted using generate the timing diagram of initial trend as:
Initial trend is fitted using simple linear regression or log-linear regression to generate the timing diagram of initial trend.
5. according to the method for claim 1, in addition to according to the timing diagram of uniformity trend determine data rise section,
Stagflation section and last transition.
6. a kind of equipment improved based on principal component and K-L information Contents Methods structure economic index, the equipment include:
Data item determining unit, it is determined that the data item with initial trend correlation connection;
Data parameters determining unit, at least two data parameters are determined according to the data item, and be described at least two numbers
Weighted value is determined according to each data parameters in parameter;
Trend point computing unit, each data parameters and corresponding weighted value at least two data parameters are counted
Calculate at least one initial trend point;
Initial trend generation unit, initial trend is formed according at least one initial trend point, and to initial trend
It is fitted to generate the timing diagram of initial trend;
Parameter of consistency generation unit, according to the timing diagram of initial trend determine for build uniformity trend at least two 1
Cause property parameter, and determine the weighted value of each parameter of consistency at least two parameter of consistency;
Uniformity trend point generation unit, each parameter of consistency at least two parameters of consistency and corresponding
Weighted value calculates at least one uniformity trend point;And
Uniformity trend generation unit, uniformity trend is formed according at least one uniformity trend point, and to consistent
Property trend is fitted to generate the timing diagram of uniformity trend.
7. equipment according to claim 6, wherein the weighted value is every number at least two data parameters
According to the variance contribution ratio of parameter.
8. equipment according to claim 6, wherein each data parameters include at least one data item.
9. equipment according to claim 6, it is described initial trend is fitted using generate the timing diagram of initial trend as:
Initial trend is fitted using simple linear regression or log-linear regression to generate the timing diagram of initial trend.
10. equipment according to claim 6, in addition to determine according to the timing diagram of uniformity trend the rise area of data
Between, stagflation section and last transition.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610371884.0A CN107451834A (en) | 2016-05-30 | 2016-05-30 | A kind of method improved based on principal component and K L information Contents Methods structure economic index |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610371884.0A CN107451834A (en) | 2016-05-30 | 2016-05-30 | A kind of method improved based on principal component and K L information Contents Methods structure economic index |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107451834A true CN107451834A (en) | 2017-12-08 |
Family
ID=60484943
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610371884.0A Pending CN107451834A (en) | 2016-05-30 | 2016-05-30 | A kind of method improved based on principal component and K L information Contents Methods structure economic index |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107451834A (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103617548A (en) * | 2013-12-06 | 2014-03-05 | 李敬泉 | Medium and long term demand forecasting method for tendency and periodicity commodities |
CN103985055A (en) * | 2014-05-30 | 2014-08-13 | 西安交通大学 | Stock market investment decision-making method based on network analysis and multi-model fusion |
CN104182800A (en) * | 2013-05-21 | 2014-12-03 | 中国农业科学院棉花研究所 | Intelligent predicting method for time sequence based on trend and periodic fluctuation |
CN104200277A (en) * | 2014-08-12 | 2014-12-10 | 南方电网科学研究院有限责任公司 | Medium-and-long-term power load prediction model establishment method |
-
2016
- 2016-05-30 CN CN201610371884.0A patent/CN107451834A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104182800A (en) * | 2013-05-21 | 2014-12-03 | 中国农业科学院棉花研究所 | Intelligent predicting method for time sequence based on trend and periodic fluctuation |
CN103617548A (en) * | 2013-12-06 | 2014-03-05 | 李敬泉 | Medium and long term demand forecasting method for tendency and periodicity commodities |
CN103985055A (en) * | 2014-05-30 | 2014-08-13 | 西安交通大学 | Stock market investment decision-making method based on network analysis and multi-model fusion |
CN104200277A (en) * | 2014-08-12 | 2014-12-10 | 南方电网科学研究院有限责任公司 | Medium-and-long-term power load prediction model establishment method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Baffigi | Italian national accounts, 1861-2011 | |
Rowthorn et al. | Deindustrialization: causes and implications | |
WO2016101628A1 (en) | Data processing method and device in data modeling | |
CN109344907A (en) | Based on the method for discrimination for improving judgment criteria sorting algorithm | |
Scherer | Lagging productivity growth: Measurement, technology, and shock effects | |
CN112732786A (en) | Financial data processing method, device, equipment and storage medium | |
Jiang et al. | Future growth pattern projections under shared socioeconomic pathways: A municipal city bottom-up aggregated study based on a localised scenario and population projections for China | |
CN107451834A (en) | A kind of method improved based on principal component and K L information Contents Methods structure economic index | |
CN104850711B (en) | A kind of electromechanical product design standard system of selection | |
CN115455341B (en) | Solving method for raw material blanking layout | |
CN116882776A (en) | Method and device for setting energy-saving strategy in steel production, electronic equipment and storage medium | |
CN112241804A (en) | Macroscopic economy leading index construction method and system for energy power data | |
CN103236383B (en) | Dynamically determining and optimization method of plasma panel process parameters importance | |
Wang et al. | Variable selection in macroeconomic forecasting with many predictors | |
Uzyakov | Toolkit for the input–output analysis of the performance of Russian economy in 1991–2013 | |
CN112488410A (en) | Production equipment performance optimization method and device based on equipment comprehensive efficiency | |
CN113112362A (en) | Bulk commodity price fluctuation risk situation evolution analysis method and system based on associated information | |
CN109726850B (en) | Method for establishing resultant yarn quality prediction model | |
Horvath et al. | Core and Periphery in the World Economy: An Empirical Assessment of the Dependence of Third World Growth on the Developed Countries | |
Hui | An empirical analysis on the impact of higher education on income inequality | |
CN110717244A (en) | Data trust degree analysis computer simulation method based on average deviation degree algorithm | |
LU504745B1 (en) | Supply chain quality optimization method and system based on gradient descent | |
Liefu | An Analysis of the Dynamic Relationship between the Maritime Silk Road and the Economic Growth of China's Import and Export Trade | |
Onakoya | The extended balance of payment constrained growth model and Nigeria’s long-run growth performance | |
CN111353638B (en) | Method and device for acquiring raw material quality range based on quality range of rice dumpling product |
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: 20171208 |