CN102930155A - Method and device for acquiring early-warming parameters of power demands - Google Patents

Abstract

The invention discloses a method and device for acquiring early-warming parameters of power demands. The method comprises the following steps of: acquiring data sequences for generating an early-warning indicator; screening the data sequences according to adjustment parameters so as to obtain data sequences; calculating trend indicators of the data sequences containing trend terms and period terms, thus obtaining early-warming indicator sequences; extracting reference indicators and selected indicators in the early-warming indicator sequences; carrying out correlation calculation on the reference indicators and the selected indicators according to a time difference analyzing model and/or a K-L information quantity model so as to acquire a correlation coefficient between the selected indicator and the reference indicator, and screening the selected indicators according to the correlation coefficient so as to acquire leading indicators and coincident indicators; and combining the leading indicators with the coincident indicators so as to obtain leading composite indexes and coincident composite indexes which are taken as the early-warming parameters. The method and the device can achieve an effect of precisely acquiring the early-warming parameters of the power demands so as to take reasonable and scientific solutions according to short-term periodic fluctuation accurately.

Description

Obtain method and the device of the early-warning parameters of electricity needs

Technical field

The present invention relates to power domain, in particular to a kind of method and device that obtains the early-warning parameters of electricity needs.

Background technology

Economic cyclic swing is the phenomenon of outwardness in the socio-economic development, it is objective law independent of man's will in the process of economic growth, it is unpractical that attempt is forced to eliminate cyclic swing by various artificial means, even the artificial elimination cyclic swing of forcing also can aggravate fluctuation under certain conditions.By the research to the economic cycle moving law, can hold the fluctuation pattern of economic cycle, thereby take suitable means to reduce the amplitude of business cycle fluctuation, prolong the cycle of economic fluctuation, thereby realize the purpose of sustained economic growth.

As a key problem of macroeconomic research, the research of economic cycle is subject to national governments and numerous economists' attention always.Along with the development of economic theory and the progress of Econometric technology, lot of domestic and foreign scholar begins to use quantitative method that economic cycle is carried out monitor and predict, purpose is to hold exactly the length of stages duration in cycle, the concrete time of turning point appearance and the dynamics of enlargement and contraction etc. as far as possible, thereby for Government and enterprise is formulated scientific and rational counter-measure for different cycles characteristics and formation mechanism, slow down the amplitude of cyclic swing, reduce the destructiveness that cyclic swing causes economic development.

Economic activity is the expulsive force of electricity needs, therefore the cycle standing wave also can to occur moving for electricity needs, but in the electricity needs field, the means of analyzing the power demand cycle fluctuation are economic growth curve and the power consumption growth curves that contrast for many years, the integrated economics scholar is to the division of China's stage of economic development, be 9-11 the cycle of fluctuation of comparatively subjectively judging electricity needs, also do not study the method for short-term power demand cycle fluctuation.

In order to address the above problem, can with the electricity needs growth rate of the direct predict future of forecasting techniques, analyze its fluctuation situation.But there are a lot of defectives in forecasting techniques:

1, the basic data that adopts is without seasonal adjustment, the Spring Festival, month in two-day weekend fate, festivals or holidays fate, leap year all basic data is had a significant impact, the fluctuation tendency of prediction will be distortion based on this;

2, no matter use the causality models such as recurrence, per capita household electricity consumption, neural network, still adopt ARIMA, logistic equal time series model, all be to do rational extrapolation according to the historical trend of statistics, be equivalent to pay close attention to the Changing Pattern of the medium-term and long-term trend of Fig. 1, can not promptly make the reasonably counter-measure of compound current economy according to the Changing Pattern of secular trend.

At present for adopting forecasting techniques to obtain short-term cyclic fluctuation distortion in the electricity needs field in the correlation technique, can't acquire the early-warning parameters that meets electricity needs, thereby cause to formulate the problem of the measure of rational reply cyclic fluctuation, not yet propose at present effective solution.

Summary of the invention

For adopting forecasting techniques to obtain short-term cyclic fluctuation distortion in the electricity needs field in the correlation technique, can't acquire the early-warning parameters that meets electricity needs, thereby cause to formulate the problem of the measure of rational reply cyclic fluctuation, effective solution is not yet proposed at present, for this reason, fundamental purpose of the present invention is to provide a kind of method and device that obtains the early-warning parameters of electricity needs, to address the above problem.

To achieve these goals, according to an aspect of the present invention, provide a kind of method of obtaining the early-warning parameters of electricity needs, the method comprises: obtain for the data sequence that generates warning index; According to adjusting parameter the data sequence is screened, to obtain the data sequence that includes trend term and periodic term; Calculating includes the trend index of the data sequence of trend term and periodic term, and according to the trend index data sequence that includes trend term and periodic term is filtered, to obtain the warning index sequence, the warning index sequence is the data of trend index for increasing in the data sequence; The generated energy that extracts in the warning index sequence is reference index, and the index of extracting except generated energy is selected index; According to step-out time analysis model and/or K-L quantity of information model warning index is carried out correlation calculations, to obtain the relative coefficient between each selected index and the reference index, and according to relative coefficient selected index is screened, to obtain leading indicators and coincidence indicator; According to the composite index number model leading indicators and coincidence indicator are synthesized, to obtain in advance composite index number and the coincident composite Index as early-warning parameters.

Further, utilize step-out time analysis model and/or K-L quantity of information model that warning index is carried out correlation calculations, with the relative coefficient between each selected index and the reference index, and according to relative coefficient selected index is screened, comprise with the step of obtaining leading indicators and coincidence indicator: obtain relative coefficient between each selected index and the reference index according to following formula _rL:

$r_l=\frac{\underset{t=1}{\overset{n_l}{\mathrm{\Σ}}}(x_{t+l}-\stackrel{\‾}{x})(y_t-\stackrel{\‾}{y})}{\sqrt{\underset{t=1}{\overset{n_l}{\mathrm{\Σ}}}{(x_{t+l}-\stackrel{\‾}{x})}^2{(y_t-\stackrel{\‾}{y})}^2}},$

Wherein, l=0, ± 1, ± 2 ..., ± 12, Y=(y ₁, y ₂..., y _n) be reference index, X=(x ₁, x ₂..., x _n) being selected index, l is the time difference, n _lBe the number of all indexs, t=1,2 ..., n is a month umber; With time difference value in the first span and relative coefficient r _lGreater than the selected index of first threshold as leading indicators, and with time difference value in the second span and relative coefficient r _lGreater than the selected index of Second Threshold as coincidence indicator.

Further, according to step-out time analysis model and/or K-L quantity of information model warning index is carried out correlation calculations, with the relative coefficient between each selected index and the reference index, and according to relative coefficient selected index is screened, comprise with the step of obtaining leading indicators and coincidence indicator: obtain relative coefficient r between each selected index and the reference index according to following formula _l:

Wherein, l=0, ± 1, ± 2 ..., ± 12, Y=(y ₁, y ₂..., y _n) be reference index, X=(x ₁, x ₂..., x _n) being selected index, l is the time difference, n _lBe the number of all indexs, t=1,2 ..., n is a month umber; With time difference value in the first span and relative coefficient r _lGreater than the selected index of the first threshold Raw performance as leading indicators, and with time difference value in the second span and relative coefficient r _lGreater than the selected index of the Second Threshold Raw performance as coincidence indicator; The Raw performance of reference index, leading indicators and the Raw performance of coincidence indicator are carried out standardization, to obtain standard basis index series p _t, the selected index of standard sequence q _t, wherein, the selected index of standard comprises standard leading indicators and standard coincidence indicator; Obtain K-L quantity of information between the selected index of each standard and the standard basis index by following formula _k: k _l=∑ p _tLn (p _t/ q _T+1), wherein, l=0, ± 1 ..., ± 12,

T=1,2 ..., n is a month umber, l is the time difference, n _lNumber for all indexs; With time difference value in the 3rd span and K-L quantity of information k _lLess than the selected index of the standard of the 3rd threshold value as leading indicators, and with time difference value in the 4th span and K-L quantity of information k _lLess than the selected index of the 4th threshold value as coincidence indicator.

Further, according to the composite index number model leading indicators and coincidence indicator are synthesized, comprise to obtain as the in advance composite index number of early-warning parameters and the step of coincident composite Index: leading indicators and coincidence indicator are carried out respectively symmetrical change process, to obtain the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t), wherein, by following formula leading indicators is carried out symmetrical change process, to obtain the symmetrical rate of change C of leading indicators _{W, i}(t):

Wherein,

Be i (i=1,2 ..., k _w) individual leading indicators, t=2,3 ..., n, k _wNumber for leading indicators; By following formula coincidence indicator is carried out symmetrical change process, to obtain the symmetrical rate of change C of coincidence indicator _{Z, i}(t):

Wherein,

Be i (i=1,2 ..., k _z) individual coincidence indicator, t=2,3 ..., n is a month umber, k _zIt is the number of coincidence indicator; To the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t) result who carries out obtaining after standardization and the trend adjustment synthesize calculating, to obtain go ahead of the rest composite index number and coincident composite Index.

Further, to the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t) result who carries out obtaining after standardization and the trend adjustment synthesize calculating, comprises with the step of obtaining go ahead of the rest composite index number and coincident composite Index: obtain normalization factor A by following formula _{W, i}And A _{Z, i}:

T=2,3 ..., n; Adopt normalization factor A _{W, i}And A _{Z, i}Respectively with the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t) carry out standardization, to obtain standardization rate of change S _{W, i}(t) and S _{Z, i}(t), wherein,

T=2,3 ..., n; To standardization rate of change S _{W, i}(t) and S _{Z, i}(t) average rate of change and process, to obtain the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t); Standardization average rate of change V according to leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t) synthesize calculating, to obtain in advance composite index number I _w(t) and coincident composite Index I _z(t), wherein,

$I_z\left(t\right)=I_z(t-1)\×\frac{200+V_z\left(t\right)}{200-V_z\left(t\right)},$ And I _w(1)=100, I _z(1)=100.

Further, to standardization rate of change S _{W, i}(t) and S _{Z, i}(t) average rate of change and process, to obtain the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t) step comprises: by following formula respectively with the standardization rate of change S of leading indicators _{W, i}(t) and the standardization rate of change S of coincidence indicator _{Z, i}(t) average rate of change and process, to obtain the average rate of change R of leading indicators _w(t) and the average rate of change R of coincidence indicator _z(t):

Wherein, λ _{W, i}And λ _{Z, i}It is respectively the weight of i index of leading indicators and coincidence indicator; Obtain index normalization factor F by following formula _w: $F_w=\left[\underset{t=2}{\overset{n}{\mathrm{\Σ}}}\right|R_w\left(t\right)|/(n-1)]/\left[\underset{t=2}{\overset{n}{\mathrm{\Σ}}}\right|R_z\left(t\right)|/(n-1)];$ According to index normalization factor F _wCarry out the standardization average rate of change and process, to obtain the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t), wherein, V _w(t)=R _w(t)/F _w, V _z(t)=R _z(t).

Further, after obtaining for the data sequence that generates warning index, method also comprises: the data in the data sequence are carried out pre-service, and pre-service comprises: fill up missing data and process, revise noise data processing, data smoothing processing and data normalization processing.

To achieve these goals, according to an aspect of the present invention, provide a kind of device that obtains the early-warning parameters of electricity needs, this device comprises: the first acquisition module is used for obtaining the data sequence for generating warning index; The first processing module is used for according to adjusting parameter the data sequence being screened, to obtain the data sequence that includes trend term and periodic term; The first computing module, be used for calculating the trend index of the data sequence that includes trend term and periodic term, and according to the trend index data sequence that includes trend term and periodic term is filtered, to obtain the warning index sequence, the warning index sequence is the data of trend index for increasing in the data sequence; The first extraction module, the generated energy that is used for extraction warning index sequence is reference index, and the index of extracting except generated energy is selected index; The second computing module, be used for according to step-out time analysis model and/or K-L quantity of information model warning index being carried out correlation calculations, to obtain the relative coefficient between each selected index and the reference index, and according to relative coefficient selected index is screened, to obtain leading indicators and coincidence indicator; The second processing module is used for according to the composite index number model leading indicators and coincidence indicator being synthesized, to obtain in advance composite index number and the coincident composite Index as early-warning parameters.

Further, the second computing module comprises: the first sub-computing module, and for the relative coefficient r that obtains according to following formula between each selected index and the reference index _l:

Wherein, l=0, ± 1, ± 2 ..., ± 12, Y=(y ₁, y ₂..., y _n) be reference index, X=(x ₁, x ₂..., x _n) being selected index, l is the time difference, n _lBe the number of all indexs, t=1,2 ..., n is a month umber; The first sub-processing module is used for time difference value in the first span and relative coefficient r _lGreater than the selected index of first threshold as leading indicators, and with time difference value in the second span and relative coefficient r _lGreater than the selected index of Second Threshold as coincidence indicator.

Further, the second computing module comprises: the second sub-computing module, and for the relative coefficient r that obtains according to following formula between each selected index and the reference index _l:

Wherein, l=0, ± 1, ± 2 ..., ± 12, Y=(y ₁, y ₂..., y _n) be reference index, X=(x ₁, x ₂..., x _n) being selected index, l is the time difference, n _lBe the number of all indexs, t=1,2 ..., n is a month umber; The second sub-processing module is used for time difference value in the first span and relative coefficient r _lGreater than the selected index of the first threshold Raw performance as leading indicators, and with time difference value in the second span and relative coefficient r _lGreater than the selected index of the Second Threshold Raw performance as coincidence indicator; The 3rd sub-processing module is used for the Raw performance of reference index, leading indicators and the Raw performance of coincidence indicator are carried out standardization, to obtain standard basis index series p _t, the selected index of standard sequence q _t, wherein, the selected index of standard comprises standard leading indicators and standard coincidence indicator; The 3rd sub-computing module is for the K-L quantity of information of obtaining by following formula between the selected index of each standard and the standard basis index _kL:k _l=∑ p _tLn (p _t/ q _T+1), wherein, l=0, ± 1 ..., ± 12,

T=1,2 ..., n is a month umber, l is the time difference, n _lNumber for all indexs;

The 4th sub-processing module is used for time difference value in the 3rd span and K-L quantity of information k _lLess than the selected index of the standard of the 3rd threshold value as leading indicators, and with time difference value in the 4th span and K-L quantity of information k _lLess than the selected index of the 4th threshold value as coincidence indicator.

Further, the second processing module comprises: the 5th sub-processing module is used for leading indicators and coincidence indicator are carried out respectively symmetrical change process, to obtain the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t), the 5th sub-processing module comprises: the 4th sub-computing module is used for by following formula leading indicators being carried out symmetrical change process, to obtain the symmetrical rate of change C of leading indicators _{W, i}(t):

Wherein, Be i (i=1,2 ..., k _w) individual leading indicators, t=2,3 ..., n is a month umber, k _wNumber for leading indicators; The 5th sub-computing module is used for by following formula coincidence indicator being carried out symmetrical change process, to obtain the symmetrical rate of change C of coincidence indicator _{Z, i}(t):

Wherein,

Be i (i=1,2 ..., k _z) individual coincidence indicator, t=2,3 ..., n, k _zIt is the number of coincidence indicator; The 6th sub-processing module is used for the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t) result who carries out obtaining after standardization and the trend adjustment synthesize calculating, to obtain go ahead of the rest composite index number and coincident composite Index.

Further, the 6th sub-processing module comprises: the 6th sub-computing module is used for obtaining normalization factor A by following formula _{W, i}And A _{Z, i}: T=2,3 ..., n; The 7th sub-processing module is used for adopting normalization factor A _{W, i}And A _{Z, i}Respectively with the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t) carry out standardization, to obtain standardization rate of change S _{W, i}(t) and S _{Z, i}(t), wherein,

T=2,3 ..., n;

The 8th sub-processing module is used for standardization rate of change S _{W, i}(t) and S _{Z, i}(t) average rate of change and process, to obtain the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t); The 7th sub-computing module is used for the standardization average rate of change V according to leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t) synthesize calculating, to obtain in advance composite index number I _w(t) and coincident composite Index I _z(t), wherein, $I_z\left(t\right)=I_z(t-1)\×\frac{200+V_z\left(t\right)}{200-V_z\left(t\right)},$ And I _w(1)=100, I _z(1)=100.

Further, the 8th sub-processing module comprises: the 9th sub-processing module is used for by following formula respectively with the standardization rate of change S of leading indicators _{W, i}(t) and the standardization rate of change S of coincidence indicator _{Z, i}(t) average rate of change and process, to obtain the average rate of change R of leading indicators _w(t) and the average rate of change R of coincidence indicator _z(t):

$R_w\left(t\right)=\frac{\underset{i=1}{\overset{k_w}{\mathrm{\Σ}}}{S}_{w,i}\left(t\right){\mathrm{\λ}}_{w,i}}{\underset{i=1}{\overset{k_w}{\mathrm{\Σ}}}{\mathrm{\λ}}_{w,i}},$ $R_z\left(t\right)=\frac{\underset{i=1}{\overset{k_z}{\mathrm{\Σ}}}{S}_{z,i}\left(t\right){\mathrm{\λ}}_{z,i}}{\underset{i=1}{\overset{k_Z}{\mathrm{\Σ}}}{\mathrm{\λ}}_{z,i}},$ Wherein, λ _{W, i}And λ _{Z, i}It is respectively the weight of i index of leading indicators and coincidence indicator; The 8th sub-computing module is used for obtaining index normalization factor F by following formula _w: $F_w=\left[\underset{t=2}{\overset{n}{\mathrm{\Σ}}}\right|R_w\left(t\right)|/(n-1)]/\left[\underset{t=2}{\overset{n}{\mathrm{\Σ}}}\right|R_z\left(t\right)|/(n-1)];$ The 9th sub-computing module is used for according to index normalization factor F _wCarry out the standardization average rate of change and process, to obtain the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t), wherein, V _w(t)=R _w(t)/F _w, V _z(t)=R _z(t).

Further, after carrying out acquisition module, install and also comprise: the tenth sub-processing module, be used for the data of data sequence are carried out pre-service, pre-service comprises: fill up missing data processing, the processing of correction noise data, data smoothing processing and data normalization and process.

Method and the device of the early-warning parameters that obtains electricity needs by the application, after the trend term and periodic term in obtaining original data sequence, by data sequence screening and calculating are obtained leading indicators and coincidence indicator, then above-mentioned leading indicators and coincidence indicator are synthesized the acquisition early-warning parameters with the index synthetic model, and according to the fluctuation of early-warning parameters analysis power demand cycle, having solved in the prior art adopts forecasting techniques to obtain short-term cyclic fluctuation distortion in the electricity needs field, can't acquire the early-warning parameters that meets electricity needs, thereby cause and to formulate the reasonably measure of reply cyclic fluctuation according to the power cycle fluctuation, realized accurately obtaining the early-warning parameters of electricity needs, thereby the effect of the counter-measure of reasonable science is formulated in cyclic fluctuation according to short-term accurately, and then the amplitude of having slowed down cyclic swing, reduce the destructiveness that cyclic swing causes power industry and economic development.

Description of drawings

Accompanying drawing described herein is used to provide a further understanding of the present invention, consists of the application's a part, and illustrative examples of the present invention and explanation thereof are used for explaining the present invention, do not consist of improper restriction of the present invention.In the accompanying drawings:

Fig. 1 is the synoptic diagram according to the cyclic fluctuation of electricity needs in the prior art;

Fig. 2 is the structural representation according to the device of the early-warning parameters that obtains electricity needs of the present invention;

Fig. 3 is the process flow diagram according to the method for the early-warning parameters that obtains electricity needs of the embodiment of the invention;

Fig. 4 is the detail flowchart according to the method for the early-warning parameters that obtains electricity needs of the embodiment of the invention;

Fig. 5 is according to the method synoptic diagram that obtains coincidence indicator and leading indicators embodiment illustrated in fig. 4; And

Fig. 6 is according to trend adjustment synoptic diagram embodiment illustrated in fig. 4.

Embodiment

Need to prove that in the situation of not conflicting, embodiment and the feature among the embodiment among the application can make up mutually.Describe below with reference to the accompanying drawings and in conjunction with the embodiments the present invention in detail.

Fig. 2 is the structural representation according to the device of the early-warning parameters that obtains electricity needs of the present invention.As shown in Figure 2, this device comprises: acquisition module 10 is used for obtaining the data sequence for generating warning index; The first processing module 30 is used for according to adjusting parameter the data sequence being screened, to obtain the data sequence that includes trend term and periodic term; The first computing module 50, be used for calculating the trend index of the data sequence that includes trend term and periodic term, and according to the trend index data sequence that includes trend term and periodic term is filtered, to obtain the warning index sequence, the warning index sequence is the data of trend index for increasing in the data sequence; The first extraction module 70, the generated energy that is used for extraction warning index sequence is reference index, and the index of extracting except power consumption is selected index; The second computing module 90, be used for according to step-out time analysis model and/or K-L quantity of information model warning index being carried out correlation calculations, to obtain the relative coefficient between each selected index and the reference index, and according to relative coefficient selected index is screened, to obtain leading indicators and coincidence indicator; The second processing module 110 is used for according to the composite index number model leading indicators and coincidence indicator being synthesized, to obtain in advance composite index number and the coincident composite Index as early-warning parameters.

Adopt the device that obtains the early-warning parameters of electricity needs of the present invention, to obtain for the data sequence that generates warning index by acquisition module, then the first processing module is screened the data sequence, and obtain the data sequence that includes trend term and periodic term, the first computing module calculates the trend index of above-mentioned data sequence afterwards, and according to the trend index above-mentioned data sequence is filtered, to obtain the warning index sequence of trend index as increasing in the data sequence, then the generated energy that extracts in the warning index sequence by the first extraction module is reference index, and the index of extracting except power consumption is selected index, the second computing module carries out correlation calculations according to step-out time analysis model and/or K-L quantity of information model to warning index afterwards, to obtain the relative coefficient between each selected index and the reference index, and according to relative coefficient selected index is screened, to obtain leading indicators and coincidence indicator, the second last processing module is synthesized leading indicators and coincidence indicator according to the composite index number model, to obtain in advance composite index number and the coincident composite Index as early-warning parameters.The device of the early-warning parameters that obtains electricity needs by the application, obtain first trend term and periodic term in the data sequence, then the data sequence is processed and obtained leading indicators and coincidence indicator, and with above-mentioned leading indicators and the coincidence indicator synthetic early-warning parameters that obtains of index synthetic model, having solved in the prior art adopts forecasting techniques to obtain short-term cyclic fluctuation distortion in the electricity needs field, can't acquire the early-warning parameters that meets electricity needs, thereby cause to formulate according to the power cycle fluctuation problem of rational counter-measure, realized accurately obtaining the early-warning parameters of electricity needs, thereby the effect of rational counter-measure is formulated in cyclic fluctuation according to short-term accurately, and then the amplitude of having slowed down cyclic swing, reduce the destructiveness that cyclic swing causes power industry and economic development.

Particularly, in the electricity needs field, the data sequence that be used for to generate warning index comprises 4 parts: secular trend item, circulation item that the industry expansion-the flourishing contraction-periodic regularities such as decline-further expansion cause, spring, summer, autumn and winter or the per month item in season that causes of fate difference, the random entry that ignorance factor causes.By the device that obtains the early-warning parameters of electricity needs of the present invention, at first reject the item and random entry in season in the original data sequence, obtain wherein trend term and the data sequence of periodic term, then with season gross domestic product (GDP) or monthly industrial added value as reference index, the time spent difference correlation is analyzed or K-L quantity of information technology is screened several leading indicators and coincidence indicator from other a large amount of economic targets, respectively several leading indicators are synthesized beforehand index with the index synthetic model at last, several coincidence indicators are synthesized same index, get access to early-warning parameters, just can realize utilizing according to early-warning parameters the purpose of future trend of the advanced judgement coincidence indicator of beforehand index.

In the above embodiment of the present invention, screen by 30 pairs of data sequences of the first processing module, original data sequence is done seasonal adjustment, namely seasonally adjust and the impact of enchancement factor, circulate (being the periodic term in above-described embodiment) as the basis take secular trend item (being the trend term in above-described embodiment) and short-term, synthetic by index screening and index, fluctuation tendency with prosperous early warning technology research electricity needs, and, above-described embodiment is with economic early warning electricity needs, screening leading indicators and coincidence indicator from a large amount of economic targets, be used for judging the cyclic fluctuation of electricity needs, so that judged result is more accurate, so that for the cyclic fluctuation of formulating reasonable counter-measure more accurately rationally.

In the application's above-mentioned practical work example, the second computing module 90 comprises: the first sub-computing module, and for the relative coefficient r that obtains according to following formula between each selected index and the reference index _l:

Wherein, l=0, ± 1, ± 2 ..., ± 12, Y=(y ₁, y ₂..., y _n) be reference index, X=(x ₁, x ₂..., x _n) being selected index, l is the time difference, n _lNumber for all indexs; The first sub-processing module is used for time difference value in the first span and relative coefficient r _lGreater than the selected index of first threshold as leading indicators, and with time difference value in the second span and relative coefficient r _lGreater than the selected index of Second Threshold as coincidence indicator.

Particularly, with " mark post " of reference index as screening, time spent difference correlation analytical model preliminary screening coincidence indicator and leading indicators.The selected index of except reference index all leading or lag period l(is time difference in above-described embodiment) (l=0, ± 1, ± 2 ... ± 12), the first sub-computing module calculates respectively the relative coefficient r of each selected index and reference index according to following formula _l:

$r_l=\frac{\underset{t=1}{\overset{n_l}{\mathrm{\Σ}}}(x_{t+l}-\stackrel{\‾}{x})(y_t-\stackrel{\‾}{y})}{\sqrt{\underset{t=1}{\overset{n_l}{\mathrm{\Σ}}}{(x_{t+l}-\stackrel{\‾}{x})}^2{(y_t-\stackrel{\‾}{y})}^2}},$ l=0，±1，±2，…，±12，

Y=(y in the following formula ₁, y ₂..., y _n) be reference index, X=(x ₁, x ₂..., x _n) being selected index, r is related coefficient, and l represents leading or lag period (being the time difference), represents when l gets negative in advance, represents when getting positive number to lag behind, and l is called as the time difference or postpones number,

With Be respectively the mean value of sequence X and Y.n _lIt is the data amount check of all indexs.Then maximum time difference related coefficient has been considered to reflect the time difference correlationship of selected index and reference index, postpones accordingly number l and represents leading or lag period, namely so that time difference relative coefficient r _lLeading or the lag period that l is exactly this selected index and reference index is counted in maximum delay.

According to above-described embodiment of the application, the second computing module 90 can also comprise: the second sub-computing module, and for the relative coefficient r that obtains according to following formula between each selected index and the reference index _l:

Wherein, l=0, ± 1, ± 2 ..., ± 12, Y=(y ₁, y ₂..., y _n) be reference index, X=(x ₁, x ₂..., x _n) being selected index, l is the time difference, n _lNumber for all indexs; The second sub-processing module is used for time difference value in the first span and relative coefficient r _lGreater than the selected index of the first threshold Raw performance as leading indicators, and with time difference value in the second span and relative coefficient r _lGreater than the selected index of the Second Threshold Raw performance as coincidence indicator; The 3rd sub-processing module is used for the Raw performance of reference index, leading indicators and the Raw performance of coincidence indicator are carried out standardization, to obtain standard basis index series p _t, the selected index of standard sequence q _t, wherein, the selected index of standard comprises standard leading indicators and standard coincidence indicator; The 3rd sub-computing module is for the K-L quantity of information k that obtains by following formula between the selected index of each standard and the standard basis index _l:

k _l=∑ p _tLn (p _t/ q _T+1), wherein, l=0, ± 1 ..., ± 12,

T=1,2 ..., n, l are the time difference, n _lNumber for all indexs; The 4th sub-processing module is used for time difference value in the 3rd span and K-L quantity of information k _lLess than the selected index of the standard of the 3rd threshold value as leading indicators, and with time difference value in the 4th span and K-L quantity of information k _lLess than the selected index of the 4th threshold value as coincidence indicator.Wherein, the first span can be less than-3, and the second span can be more than or equal to-2 and be less than or equal to 2, and first threshold can be 0.7, and Second Threshold also can be 0.7.Wherein, t is a month umber.

Particularly, get access to relative coefficient r by the second sub-computing module calculating _lAfterwards, the second sub-processing module with time difference value in the first span and relative coefficient r _lGreater than the selected index of the first threshold Raw performance as leading indicators, and with time difference value in the second span and relative coefficient r _lGreater than the selected index of the Second Threshold Raw performance as coincidence indicator, the 3rd sub-processing module is done standardization with reference index afterwards, and the standard basis index series after the processing is designated as p _t:

$p_t=y_t/\underset{t=1}{\overset{n}{\mathrm{\Σ}}}{y}_t,$ t＝1，2，…，n，

The 3rd sub-processing module is also done standardization with leading indicators and the coincidence indicator of primary election, and the sequence of the selected index of the standard after the processing is designated as q _t:

$q_t=x_t/\underset{t=1}{\overset{n}{\mathrm{\Σ}}}{x}_t,$ t＝1,2，…，n，

Then, the 4th sub-processing module is calculated as follows behind each primary election beacon delay l the K-L quantity of information k about reference index _l:

k _l=∑p _tln(p _t/q _t+1)，l＝0，±1，…,±12

Wherein, l represents leading or lag period, represents when l gets negative in advance, represents when getting positive number to lag behind, and l is called as the time difference, n _lBe the data amount check (i.e. the number of all indexs) after data are evened up, the t in the above-mentioned formula represents month, and j represents year.

More specifically, after calculating 2L+1 K-L quantity of information, from k _lSelect a minimum value k in the value _{L '}As the K-L quantity of information of selected index x about reference index y, namely

L is counted in its corresponding delay ^*It is exactly the optimal leading or month number (season) that lags behind of selected index.The K-L quantity of information illustrates that more close to 0 index x and reference index y are more approaching.And leading indicators and the coincidence indicator that screens be designated as respectively W ⁽³⁾And Z ⁽³⁾

According to above-described embodiment of the application, the second processing module 110 can comprise: the 5th sub-processing module is used for leading indicators and coincidence indicator are carried out respectively symmetrical change process, to obtain the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t), the 5th sub-computing module comprises: the 4th sub-computing module is used for by following formula leading indicators being carried out symmetrical change process, to obtain the symmetrical rate of change C of leading indicators _{W, i}(t):

Wherein, Be i (i=1,2 ..., k _w) individual leading indicators, t=2,3 ..., n, k _wThe number of leading indicators; And the 5th sub-computing module, be used for by following formula coincidence indicator being carried out symmetrical change process, to obtain the symmetrical rate of change C of coincidence indicator _{Z, i}(t):

Wherein, Be i (i=1,2 ..., k _z) individual coincidence indicator, t=2,3 ..., n, k _zIt is the number of coincidence indicator; The 6th sub-processing module is used for the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t) result who carries out obtaining after standardization and the trend adjustment synthesize calculating, to obtain go ahead of the rest composite index number and coincident composite Index.

Particularly, the 5th sub-processing module ask the symmetrical rate of change of index with the index synthetic model and by the 6th sub-processing module with its standardization, by the 4th sub-computing module in the 5th sub-processing module according to following formula pair

Ask symmetrical rate of change C _{W, i}(t):

T=2,3 ..., n, wherein,

Be i (i=1,2 ..., k _w) individual leading indicators, k _wIt is the number of leading indicators.

By the 5th sub-computing module according to following formula pair Ask symmetrical rate of change C _{Z, i}(t):

T=2,3 ..., n, wherein,

Be i (i=1,2 ..., k _z) individual coincidence indicator, k _zIt is the number of coincidence indicator.

In above-described embodiment of the application, the 6th sub-processing module can comprise: the 6th sub-computing module is used for obtaining normalization factor A by following formula _{W, i}And A _{Z, i}:

T=2,3 ..., n; The 7th sub-processing module is used for adopting normalization factor A _{W, i}And A _{Z, i}Respectively with the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t) carry out standardization, to obtain standardization rate of change S _{W, i}(t) and S _{Z, i}(t), wherein,

$S_{w,i}\left(t\right)=\frac{C_{w,i}\left(t\right)}{A_{w,i}},$ $S_{z,i}\left(t\right)=\frac{C_{z,i}\left(t\right)}{A_{z,i}},$ t=2,3,…,n；

The 8th sub-processing module is used for standardization rate of change S _{W, i}(t) and S _{Z, i}(t) average rate of change and process, to obtain the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t); The 7th sub-computing module is used for the standardization average rate of change V according to leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t) synthesize calculating, to obtain in advance composite index number I _w(t) and coincident composite Index I _z(t), wherein, $I_z\left(t\right)=I_z(t-1)\×\frac{200+V_z\left(t\right)}{200-V_z\left(t\right)},$ And I _w(1)=100, I _z(1)=100.

Particularly, the 6th sub-computing module according to following formula to the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t) do standardized calculation, make its average absolute value equal 1, obtain normalization factor A _{W, i}And A _{Z, i}:

$A_{w,i}=\underset{t=2}{\overset{n}{\mathrm{\Σ}}}\frac{\left|C_{w,i}\left(t\right)\right|}{n-1},$ $A_{z,i}=\underset{t=2}{\overset{n}{\mathrm{\Σ}}}\frac{\left|C_{z,i}\left(t\right)\right|}{n-1},$ t=2,3,…，n，

Then the 7th sub-processing module is according to A _{W, i}And A _{Z, i}Respectively with C _{W, i}(t) and C _{Z, i}(t) standardization obtains standardization rate of change S _{W, i}(t) and S _{Z, i}(t):

T=2,3 ..., n, the 8th sub-processing module is according to standardization rate of change S afterwards _{W, i}(t) and S _{Z, i}(t) average rate of change and process, to obtain the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t), and by the 7th sub-computing module according to the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t) synthesize calculating, to obtain in advance composite index number I _w(t) and coincident composite Index I _z(t), particularly: make I _w(1)=100, I _z(1)=100, then

$I_w\left(t\right)=I_w(t-1)\×\frac{200+V_w\left(t\right)}{200-V_{w}t},$ $I_z\left(t\right)=I_z(t-1)\×\frac{200+V_z\left(t\right)}{200-V_z\left(t\right)}.$

More specifically, getting access in advance composite index number I _w(t) and coincident composite Index I _z(t) afterwards, can adjust electricity needs trend, method is as follows:

According to following compound interest formula each sequence of coincidence indicator group is obtained respectively separately average growth rate:

$r_i=(\sqrt[m_i]{C_{\mathrm{Li}}/C_{\mathrm{Ii}}}-1)\×100,$ i=1,2,…,k _z，

Wherein, Fig. 6 is according to trend adjustment synoptic diagram embodiment illustrated in fig. 4, and as shown in Figure 6, t is month,

With

Be respectively i index of coincidence indicator group at first with the most metacyclic mean value, m _IiWith m _LiBe respectively i index of coincidence indicator group at first with the most metacyclic month number, k ₂The coincidence indicator number, m _iThat the center of circulating at first is to the moon number between the most metacyclic center.

Then obtain the average growth rate G of coincidence indicator group _r, and with it as target trend: Afterwards in advance and the initial composite index number I of coincidence indicator _w(t) and I _z(t) obtain their average growth rate r ' separately with compound interest formula respectively _wAnd r ' _z:

$r_w^{\′}(\sqrt[m_w]{C_{\mathrm{Lw}}/C_{\mathrm{Iw}}}-1)\×100,$ $r_z^{\′}(\sqrt[m_z]{C_{\mathrm{Lz}}/C_{\mathrm{Iz}}}-1)\×100,$

Wherein,

Again to the standardization average rate of change V of leading indicators group and coincidence indicator group _w(t) and V _z(t) do trend adjustment:

V′ _w(t)=V _w(t)+(G _r-r′ _w)，V′ _z(t)=V _z(t)+(G _r-r′ _z)，t=2,3，...，n。Then calculate composite index number according to the method in above-described embodiment: make I ' _w(1)=100, I ' _z(1)=100, then

${I^{\′}}_w\left(t\right)={I^{\′}}_w(t-1)\×\frac{200+{V^{\′}}_w\left(t\right)}{200-{V^{\′}}_w\left(t\right)},$ ${I^{\′}}_z\left(t\right)={I^{\′}}_z(t-1)\×\frac{200+{V^{\′}}_z\left(t\right)}{200-{V^{\′}}_z\left(t\right)},$

Generation is take the in advance composite index number CI of benchmark time as 100 _w(t) and coincident composite Index CI _z(t):

${\mathrm{CI}}_w\left(t\right)=(I_w^{\′}\left(t\right)/\stackrel{\‾}{I_w^{\′}}\×100),$ ${\mathrm{CI}}_z\left(t\right)=(I_z^{\′}\left(t\right)/\stackrel{\‾}{I_z^{\′}})\×100,$

Wherein

With

Respectively I ' _w(t) and I ' _z(t) at the mean value in benchmark time.

According to above-described embodiment of the application, the 8th sub-processing module can comprise: the 9th sub-processing module is used for by following formula respectively with the standardization rate of change S of leading indicators _{W, i}(t) and the standardization rate of change S of coincidence indicator _{Z, i}(t) average rate of change and process, to obtain the average rate of change R of leading indicators _w(t) and the average rate of change R of coincidence indicator _z(t):

$R_w\left(t\right)=\frac{\underset{i=1}{\overset{k_w}{\mathrm{\Σ}}}{S}_{w,i}\left(t\right){\mathrm{\λ}}_{w,i}}{\underset{i=1}{\overset{k_w}{\mathrm{\Σ}}}{\mathrm{\λ}}_{w,i}},$ $R_z\left(t\right)=\frac{\underset{i=1}{\overset{k_z}{\mathrm{\Σ}}}{S}_{z,i}\left(t\right){\mathrm{\λ}}_{z,i}}{\underset{i=1}{\overset{k_z}{\mathrm{\Σ}}}{\mathrm{\λ}}_{z,i}},$ Wherein, λ _{W, i}And λ _{Z, i}It is respectively the weight of i index of leading indicators and coincidence indicator; The 8th sub-computing module is used for obtaining index normalization factor F by following formula _w:

$F_w=\left[\underset{t=2}{\overset{n}{\mathrm{\Σ}}}\right|R_w\left(t\right)|/(n-1)]/\left[\underset{t=2}{\overset{n}{\mathrm{\Σ}}}\right|R_z\left(t\right)|/(n-1)];$

The 9th sub-computing module is used for according to index normalization factor F _wCarry out the standardization average rate of change and process, to obtain the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator ₂(t), wherein, V _w(t)=R _w(t)/F _w, V _z(t)=R _z(t).

Particularly, by following formula respectively with the standardization rate of change S of leading indicators _{W, i}(t) and the standardization rate of change S of coincidence indicator _{Z, i}(t) average rate of change and process, to obtain the average rate of change R of leading indicators _w(t) and the average rate of change R of coincidence indicator _z(t):

$R_w\left(t\right)=\frac{\underset{i=1}{\overset{k_w}{\mathrm{\Σ}}}{S}_{w,i}\left(t\right){\mathrm{\λ}}_{w,i}}{\underset{i=1}{\overset{k_w}{\mathrm{\Σ}}}{\mathrm{\λ}}_{w,i}},$ $R_z\left(t\right)=\frac{\underset{i=1}{\overset{k_z}{\mathrm{\Σ}}}{S}_{z,i}\left(t\right){\mathrm{\λ}}_{z,i}}{\underset{i=1}{\overset{k_Z}{\mathrm{\Σ}}}{\mathrm{\λ}}_{z,i}},$ λ _{W, i}And λ _{Z, i}Respectively in advance and the weight of i index of coincidence indicator group.

Then according to following formula gauge index normalization factor F _w:

$F_w=\left[\underset{t=2}{\overset{n}{\mathrm{\Σ}}}\right|R_w\left(t\right)|/(n-1)]/\left[\underset{t=2}{\overset{n}{\mathrm{\Σ}}}\right|R_z\left(t\right)|/(n-1)];$

At last according to index normalization factor F _wCarry out the standardization average rate of change and process, to obtain the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t):

V _w(t)=R _w(t)/F _w, V _z(t)=R _z(t), t=2,3 ..., n wherein, goes to adjust the average rate of change of leading indicators sequence and lagging indicator sequence with the amplitude of the average rate of change of coincidence indicator sequence, its objective is for two indexes are used as a harmonious system and uses.

Above-described embodiment according to the application, after carrying out acquisition module 10, device can also comprise: the tenth sub-processing module, be used for the data of data sequence are carried out pre-service, pre-service comprises: fill up missing data processing, the processing of correction noise data, data smoothing processing and data normalization and process.

Fig. 3 is the process flow diagram according to the method for the early-warning parameters that obtains electricity needs of the embodiment of the invention.Fig. 4 is the detail flowchart according to the method for the early-warning parameters that obtains electricity needs of the embodiment of the invention.

As shown in Figure 3 and Figure 4, the method comprises the steps:

Step S102 obtains for the data sequence that generates warning index.

Wherein, this step can realize by the step S202 among Fig. 4: collect macroeconomy and electricity needs monthly data, and do pre-service.

Step S104 screens the data sequence according to adjusting parameter, to obtain the data sequence that includes trend term and periodic term.Wherein, can realize the method by step S204 among Fig. 4: by the data sequence is done seasonal adjustment, obtain trend term and the periodic term of all indexs.

Particularly, in the electricity needs field, the data sequence that generates warning index comprises 4 parts: secular trend item (being trend term), circulation item (being periodic term) that the industry expansion-the flourishing contraction-periodic regularities such as decline-further expansion cause, spring, summer, autumn and winter or the per month item in season that causes of fate difference, the random entry that ignorance factor causes, be that the index of the magnitude of value and physical quantity is done seasonal adjustment to data, reject the item and random entry in season in the data sequence, obtain wherein trend term and the data sequence of periodic term.

Step S106, calculating includes the trend index of the data sequence of trend term and periodic term, and according to the trend index data sequence that includes trend term and periodic term is filtered, to obtain the warning index sequence, the warning index sequence is the data of trend index for increasing in the data sequence.Wherein, this step can realize by step S206 to S208: step S206 judges whether the achievement data in the data sequence is growth indices class achievement data, and in the situation that is, execution in step S208: the growth rate of parameter, achievement data in data sequence is not in the situation of growth indices class achievement data, execution in step S210.

Step S108, the generated energy that extracts in the warning index sequence is reference index, and the index of extracting except generated energy is selected index.Wherein, realize the method by step S210: determine reference index.

Specifically in the electricity needs field, with season gross domestic product (GDP) or monthly industrial added value as reference index.

Step S110, according to step-out time analysis model and/or K-L quantity of information model warning index is carried out correlation calculations, obtaining the relative coefficient between each selected index and the reference index, and according to relative coefficient selected index is screened, to obtain leading indicators and coincidence indicator.Wherein, step S212 can realize the method: with the initial coincidence indicator of step-out time analysis model discrimination and initial leading indicators; Then execution in step S214: with K-L quantity of information model final coincidence indicator and leading indicators of screening from the coincidence indicator of primary election and leading indicators.

Particularly, can use step-out time analysis model preliminary screening coincidence indicator and leading indicators from the indexs such as macroeconomy, the output of industrial product.

Step S112 synthesizes leading indicators and coincidence indicator according to the composite index number model, to obtain in advance composite index number and the coincident composite Index as early-warning parameters.Wherein, realize said method by execution in step S216: respectively coincidence indicator and leading indicators are synthesized same index and beforehand index with the index synthetic model; Then execution in step S218: obtain early-warning parameters according to same index and beforehand index, and use early-warning parameters to analyze the power demand cycle fluctuation.

Particularly, respectively several leading indicators are synthesized beforehand index with the index synthetic model, several coincidence indicators are synthesized same index, get access to early-warning parameters, just can realize utilizing according to early-warning parameters the purpose of future trend of the advanced judgement coincidence indicator of beforehand index.

Adopt the method for obtaining the early-warning parameters of electricity needs of the present invention, screen by obtaining for the data sequence that generates warning index, and obtain the data sequence that includes trend term and periodic term, calculate afterwards the trend index of above-mentioned data sequence, and according to the trend index this data sequence is filtered, to obtain the warning index sequence of trend index as increasing in the data sequence, then the generated energy that extracts in the warning index sequence is reference index, and the index of extracting except power consumption is selected index, according to step-out time analysis model and/or K-L quantity of information model warning index is carried out correlation calculations afterwards, to obtain the relative coefficient between each selected index and the reference index, and according to relative coefficient selected index is screened, to obtain leading indicators and coincidence indicator, according to the composite index number model leading indicators and coincidence indicator are synthesized at last, to obtain in advance composite index number and the coincident composite Index as early-warning parameters.The method of the early-warning parameters that obtains electricity needs by the application, obtain first trend term and periodic term in the data sequence, then obtain leading indicators and coincidence indicator by screening, and with above-mentioned leading indicators and the coincidence indicator synthetic early-warning parameters that obtains of index synthetic model, having solved in the prior art adopts forecasting techniques to obtain short-term cyclic fluctuation distortion in the electricity needs field, can't acquire the early-warning parameters that meets electricity needs, thereby cause to formulate according to the power cycle fluctuation problem of rational counter-measure, realized accurately obtaining the early-warning parameters of electricity needs, thereby formulate the reasonably effect of the measure of reply cyclic fluctuation according to the short-term cyclic fluctuation accurately, and then the amplitude of having slowed down cyclic swing, reduce the destructiveness that cyclic swing causes power industry and economic development.

In the above embodiment of the present invention, at first the data sequence is screened by step S102, also namely original data sequence is done seasonal adjustment, namely seasonally adjust and the impact of enchancement factor, circulate (being the periodic term in above-described embodiment) as the basis take secular trend item (being the trend term in above-described embodiment) and short-term, synthetic by index screening and index, fluctuation tendency with prosperous early warning technology research electricity needs, and, above-described embodiment is with economic early warning electricity needs, screening leading indicators and coincidence indicator from a large amount of economic targets, be used for judging the cyclic fluctuation of electricity needs, so that judged result is more accurate, so that for the cyclic fluctuation of formulating reasonable counter-measure more accurately rationally.

Wherein, the pre-service among the step S202 can comprise: fill up missing data and process, revise noise data processing, data smoothing processing and data normalization processing, and the data sequence after will processing is as Y ⁽⁰⁾

Particularly, in above-described embodiment of the application, before execution in step S204, the method also comprises the steps:

(1) remove given month that festivals or holidays or other reason cause what difference of working days between different year, concrete grammar is as follows:

In this embodiment, establish monthly data sequence to be analyzed

(t=1,2 ..., n) total m, n month, n=m * 12, t represents a month umber, and i represents number of weeks, and j represents a year number, and then each month real work fate is D _t(t=1,2 ..., n), the work fate in mean annual each month of m is:

(L=1,2 ..., 12), can obtain the work fate and adjust coefficient sequence p _t:

T=1,2 ..., then n adjusts coefficient p according to the work fate _tObtain the monthly data sequence Y after month is adjusted ⁽¹⁾: $Y_t^{\left(1\right)}=Y_t^{\left(0\right)}/p_t.$

(2) to the inside of a week adjustment in the sequence: from former sequence, extract out because of the different changes that cause of the inside of a week (week structure) of each month.

Suppose that the inside of a week variable factor is included in the irregular key element, namely the form of irregular key element is ID _r, suppose to go out ID from former Series Decomposition _r, obtain Monday with regretional analysis, two ..., the respective weights of day, thereby with ID _rBe decomposed into real irregular key element I and the inside of a week variable factor D _r

${\mathrm{ID}}_{\mathrm{rt}}\text{-1.0=}\frac{x_{1\mathrm{<figure-callout\; id="1"\; label="t\; B"\; filenames="HDA00002333373000011.png"\; state="\{\{state\}\}">t</figure-callout>}}{B}_{}{}_1+x_{2\mathrm{<figure-callout\; id="2"\; label="t\; B"\; filenames="HDA00002333373000011.png"\; state="\{\{state\}\}">t</figure-callout>}}{B}_{}{}_2+\&CenterDot;\&CenterDot;\&CenterDot;x_{7t}{B}_t}{A_t}+I_t,$

In the following formula: ID _RtFor including t month the inside of a week variable factor D _rIrregular key element; x _ItFor the fate of t week i in the month (t=1 ..., n); B _iFor the weight of week i (

); A _tBe the fate of the t month, got February 28.25 days; I _tBe real irregular key element.

The B that is obtaining thus _iEstimated value be b _iThe time, can calculate according to following formula the inside of a week variable factor D of the t month _r:

D _rt={x _1t(b ₁+1)+x _2t(b ₂+1)+…+x _7t(b ₇+1)}/A _t。

(3) after according to festivals or holidays and the inside of a week factor the data sequence being adjusted, special in the sequence is revised.

During various factors in decomposing economic time series, need to revise in advance the item that in erratic variation, has remarkable exceptional value (being special, such as the impact of strike, awful weather, data error etc.).Its method is:

A. the setting of special boundary value.

Suppose from former sequence, to decomposite irregular key element I.In order to get rid of the exceptional value among the Irregular variation I, need to calculate 5 years moving average standard deviations of I.At first calculate 5 years initial moving average standard deviations

That is:

${\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="0"\; label="j"\; filenames="HDA00002333373000011.png"\; state="\{\{state\}\}">j</figure-callout>}}^0=\sqrt{\frac{1}{60}\underset{t=j\&times;12-36+1}{\overset{j\&times;12+24}{\mathrm{\&Sigma;}}}{(I_t-{\stackrel{\&OverBar;}{I}}_j)}^2},$ j=3,4,…,m-2

In the following formula Be 5 years moving averages of I sequence, m is the year number of I sequence, t=1, and 2 ..., n(t is the moon number of I sequence).Allow

In center year corresponding to during 5 years, calculate one every year

So

It is an annual sequence.Can think satisfied

I _tBe special, remove these I _t, by following formula:

${\mathrm{\&sigma;}}_j=\sqrt{\frac{1}{60-a}\mathrm{\&Sigma;}{(I_t-{\stackrel{\&OverBar;}{I}}_j)}^2},$ j=3,4,…,m-2，

Again calculate 5 years mobile standard deviation { σ _j, t=1 in the formula, 2 ..., n (t is the moon number of I sequence), a is the number of special value.{ σ _jThe sequence two ends respectively lack 2, adopt respectively apart from top and the { σ in the 3rd year of terminal _jReplace the σ in 2 years of two ends shortcomings _jValue.

B. special correction.

According to 5 years mobile standard deviation { σ _jCalculate the flexible strategy w of correction,

Wherein, t=1,2 ..., n(t is the moon number of I sequence), j=1,2 ..., m utilizes above-mentioned inequality can revise the special item of I sequence: corresponding to w _t＜1 It is with this w _tBe flexible strategy, with it each I of two of close front and back _T-2, I _T-1, I _T+1, I _T+2(notice that the corresponding w of item get must equal 1, otherwise get the value on next door) totally 5 make weighted mean, with the value that obtains like this Substitute I _tIf corresponding to w _t＜1 I _tWhen being positioned at two ends, with this w _tBe power, 3 ws close with it _t=1 I value totally 4 make weighted mean, with this mean value of obtaining

Substitute I _tThe I sequence of revising after special is designated as I ^w

(4) according to the monthly data sequence

Carry out initial estimation, obtain initial trend circulating component.

Use the trend circulating component in 12 moving average estimated sequences of centralization

${\mathrm{TC}}_t^{\left(1\right)}=\frac{1}{2}(\frac{y_{t-6}+y_{t-5}+\&CenterDot;\&CenterDot;\&CenterDot;+y_t+\&CenterDot;\&CenterDot;\&CenterDot;+y_{t+5}}{12}+\frac{y_{t-5}+\&CenterDot;\&CenterDot;\&CenterDot;+y_t+\&CenterDot;\&CenterDot;\&CenterDot;+y_{t+5}+y_{t+6}}{12}),$

$=\frac{1}{24}{y}_{t-6}+\frac{1}{12}{y}_{t-5}+\&CenterDot;\&CenterDot;\&CenterDot;+\frac{1}{12}{y}_t+\&CenterDot;\&CenterDot;\&CenterDot;+\frac{1}{12}{y}_{t+5}+\frac{1}{24}{y}_{t+6}$

Wherein, y _T-6, y _T-5..., y _t..., y _T+5, y _T+6It is the monthly data sequence

In element.

(5) according to the trend circulating component

Estimate irregular composition in season

(6) according to irregular composition in season

To use in each in month 3 * 3 moving averages according to a preliminary estimate season composition:

At first according to following formula, obtain seasonal factor

${\hat{S}}_t^{\left(1\right)}=\frac{1}{3}(\frac{{\mathrm{SI}}_{t-24}^{\left(1\right)}+{\mathrm{SI}}_{t-12}^{\left(1\right)}+{\mathrm{SI}}_t^{\left(1\right)}}{3}+\frac{{\mathrm{SI}}_{t-12}^{\left(1\right)}+{\mathrm{SI}}_t^{\left(1\right)}+{\mathrm{SI}}_{t+12}^{\left(1\right)}}{3}+\frac{{\mathrm{SI}}_t^{\left(1\right)}+{\mathrm{SI}}_{t+12}^{\left(1\right)}+{\mathrm{SI}}_{t+24}^{\left(1\right)}}{3});$

$\frac{1}{9}{\mathrm{SI}}_{t-24}^{\left(1\right)}+\frac{2}{9}{\mathrm{SI}}_{t-12}^{\left(1\right)}+\frac{3}{9}{\mathrm{SI}}_t^{\left(1\right)}+\frac{2}{9}{\mathrm{SI}}_{t+12}^{\left(1\right)}+\frac{1}{9}{\mathrm{SI}}_{t+24}^{\left(1\right)}$

Then seasonal factor is carried out standardized calculation, obtain the standard seasonal factor

So that factor sum is approximately zero in each continuous 12 month:

$S_t^{\left(1\right)}={\hat{S}}_t^{\left(1\right)}-(\frac{1}{24}{\hat{S}}_{t-6}^{\left(1\right)}+\frac{1}{12}{\hat{S}}_{t-5}^{\left(1\right)}+\&CenterDot;\&CenterDot;\&CenterDot;+\frac{1}{12}{\hat{S}}_{t+5}^{\left(1\right)}+\frac{1}{24}{\hat{S}}_{t+6}^{\left(1\right)}).$

(7) according to the standard seasonal factor

Sequence after the first estimation seasonal adjustment

(8) further estimate the trend circulating component with 13 moving averages

${\mathrm{TC}}_t^{\left(2\right)}=\frac{1}{16796}(-{375\mathrm{TCI}}_{t-6}^{\left(1\right)}-{468\mathrm{TCI}}_{t-5}^{\left(1\right)}+{1100\mathrm{TCI}}_{t-3}^{\left(1\right)}+2475\mathrm{TC}{I}_{t-2}^{\left(1\right)}+{3600\mathrm{TCI}}_{t-1}^{\left(1\right)}$

$+{4032\mathrm{TCI}}_t^{\left(1\right)}+{3600\mathrm{TCI}}_{t+1}^{\left(1\right)}+{2475\mathrm{TCI}}_{t+2}^{\left(1\right)}+{1100\mathrm{TCI}}_{t+3}^{\left(1\right)}-{468\mathrm{TCI}}_{t+5}^{\left(1\right)}-{325\mathrm{TCI}}_{t+6}^{\left(1\right)}).$

(9) further estimate irregular composition in season:

(10) estimate final composition in season with 3 * 5 moving averages:

Obtain final seasonal factor according to following formula

${\hat{S}}_t^{\left(2\right)}=\frac{1}{15}{\mathrm{SI}}_{t-36}^{\left(2\right)}+\frac{2}{15}{\mathrm{SI}}_{t-24}^{\left(2\right)}+\frac{3}{15}{\mathrm{SI}}_{t-12}^{\left(2\right)}+\frac{3}{15}{\mathrm{SI}}_t^{\left(2\right)}+\frac{3}{15}{\mathrm{SI}}_{t+12}^{\left(2\right)}+\frac{2}{15}{\mathrm{SI}}_{t+24}^{\left(2\right)}+\frac{1}{15}{\mathrm{SI}}_{t+36}^{\left(2\right)}$

Then to seasonal factor

Carry out standardization and obtain standardized final seasonal factor So that factor sum is approximately zero in each continuous 12 month:

$S_t^{\left(2\right)}={\hat{S}}_t^{\left(2\right)}-(\frac{1}{24}{\hat{S}}_{t-6}^{\left(2\right)}+\frac{1}{12}{\hat{S}}_{t-5}^{\left(2\right)}+\&CenterDot;\&CenterDot;\&CenterDot;+\frac{1}{12}{\hat{S}}_{t+5}^{\left(2\right)}+\frac{1}{24}{\hat{S}}_{t+6}^{\left(2\right)}).$

(11) according to standardized final seasonal factor

Sequence after the estimation seasonal adjustment

(12) obtain final trend circulating component, its method is as follows:

At first estimate irregular composition:

Then use irregular

And trend term The ratio of absolute value sum of monthly increment rate weigh the conspicuousness of irregular composition

$\stackrel{\&OverBar;}{I}/\stackrel{\&OverBar;}{C}=\frac{\underset{t=2}{\overset{n}{\mathrm{\&Sigma;}}}|I_t^{\left(2\right)}/{I_{t-1}^{\left(2\right)}}^{-1}|}{\underset{t=2}{\overset{n}{\mathrm{\&Sigma;}}}\left|{\mathrm{TC}}_t^{\left(2\right)}{{\mathrm{TC}}_{t-1}^{\left(2\right)}}^{-1}\right|},$

If

Then estimate final trend circulating component with following 9 moving averages:

${\mathrm{TC}}_t^{\left(3\right)}=\frac{1}{2431}(-{99\mathrm{TCI}}_{t-4}^{\left(2\right)}-{24\mathrm{TCI}}_{t-3}^{\left(2\right)}+{288\mathrm{TCI}}_{t-2}^{\left(2\right)}+{348\mathrm{TCI}}_{t-1}^{\left(2\right)}+{805\mathrm{TCI}}_t^{\left(2\right)};$

$+648{\mathrm{TCI}}_{t+1}^{\left(2\right)}+{288\mathrm{TCI}}_{t+2}^{\left(2\right)}-{24\mathrm{TCI}}_{t+3}^{\left(2\right)}-{99\mathrm{TCI}}_{t+4}^{\left(2\right)})$

If

Then estimate final trend circulating component with following 13 moving averages:

${\mathrm{TC}}_t^{\left(3\right)}=\frac{1}{16796}(-{325\mathrm{TCI}}_{t-6}^{\left(2\right)}-468\mathrm{TC}{I}_{t-5}^{\left(2\right)}+{1100\mathrm{TCI}}_{t-3}^{\left(2\right)}+{2475\mathrm{TCI}}_{t-2}^{\left(2\right)}+{3600\mathrm{TCI}}_{t-1}^{\left(2\right)};$

$+{4032\mathrm{TCI}}_t^{\left(2\right)}+{3600\mathrm{TCI}}_{t+1}^{\left(2\right)}+{2475\mathrm{TCI}}_{t+2}^{\left(2\right)}+{1100\mathrm{TCI}}_{t+3}^{\left(2\right)}-{468\mathrm{TCI}}_{t+5}^{\left(2\right)}-{325\mathrm{TCI}}_{t+6}^{\left(2\right)})$

If

Then estimate final trend circulating component with following 23 moving averages:

${\mathrm{TC}}_t^{\left(3\right)}=\frac{1}{4032015}(-{17250\mathrm{TCI}}_{t-11}^{\left(2\right)}-{44022\mathrm{TCI}}_{t-10}^{\left(2\right)}-{63250\mathrm{TCI}}_{t-9}^{\left(2\right)}-{5757\mathrm{TCI}}_{t-8}^{\left(2\right)}-{19950\mathrm{TCI}}_{t-7}^{\left(2\right)})$

$+{54150\mathrm{TCI}}_{t-6}^{\left(2\right)}+156978{\mathrm{TCI}}_{t-5}^{\left(2\right)}+{275400\mathrm{TCI}}_{t-4}^{\left(2\right)}+{392700\mathrm{TCI}}_{t-3}^{\left(2\right)}+{491700\mathrm{TCI}}_{t-2}^{\left(2\right)}+{557700\mathrm{TCI}}_{t-1}^{\left(2\right)},$

$+580853{\mathrm{TCI}}_t^{\left(2\right)}+{557700\mathrm{TCI}}_{t+1}^{\left(2\right)}+{491700\mathrm{TCI}}_{t+2}^{\left(2\right)}+{392700\mathrm{TCI}}_{t+3}^{\left(2\right)}+{275400\mathrm{TCI}}_{t+4}^{\left(2\right)}+{156978\mathrm{TCI}}_{t+5}^{\left(2\right)}$

$+54150{\mathrm{TCI}}_{t+6}^{\left(2\right)}-{19950\mathrm{TCI}}_{t+7}^{\left(2\right)}-{58575\mathrm{TCI}}_{t+8}^{\left(2\right)}-{63250\mathrm{TCI}}_{t+9}^{\left(2\right)}-{44022\mathrm{TCI}}_{t+10}^{\left(2\right)}-{17250\mathrm{TCI}}_{t+11}^{\left(2\right)})$

Get access to final trend circulating component through above-mentioned calculating

(13) according to the trend circulating component

With the sequence after the seasonal adjustment

Estimate final irregular composition: $I_t^{\left(3\right)}={\mathrm{TCI}}_t^{\left(2\right)}-{\mathrm{TC}}_t^{\left(3\right)}.$

Then get access to the sequence Y that only comprises trend circulation item after seasonal adjustment ⁽²⁾=TC ⁽³⁾

In above-described embodiment of the application, step S108 specifically realizes by the following method among Fig. 4: to physical quantity and the indicator of output value, calculate the development index of seasonal adjustment sequence afterwards:

Be the index of development index, do not do conversion, i.e. Y ⁽³⁾=Y ⁽²⁾

In above-described embodiment of the application, execution in step S110: warning index is carried out correlation calculations according to step-out time analysis model and/or K-L quantity of information model, to obtain the relative coefficient between each selected index and the reference index, and according to relative coefficient selected index is screened, comprise the steps: with the process of obtaining leading indicators and coincidence indicator

Obtain relative coefficient r between each selected index and the reference index according to following formula _l:

Wherein, l=0, ± 1, ± 2 ..., ± 12, Y=(y ₁, y ₂..., y _n) be reference index, X=(x ₁, x ₂..., x _n) being selected index, l is the time difference, is in particular leading or lag period, n _lNumber for all indexs; With time difference value in the first span and relative coefficient r _lGreater than the selected index of first threshold as leading indicators, and with time difference value in the second span and relative coefficient r _lGreater than the selected index of Second Threshold as coincidence indicator.Wherein, the first span can be less than-3, and first threshold can be that 0.7, the second span can be more than or equal to-2 and less than or equal to 2, and Second Threshold can be 0.7.

Particularly, in this step with reference index as the screening " mark post ", time spent difference correlation analytical model preliminary screening coincidence indicator and leading indicators.The selected index of except reference index all leading or hysteresis l phase (being the time difference in above-described embodiment) (l=0, ± 1, ± 2 ..., ± 12), calculate respectively the relative coefficient r of each selected index and reference index by following formula _l:

$r_l=\frac{\underset{t=1}{\overset{n_l}{\mathrm{\&Sigma;}}}(x_{t+l}-\stackrel{\&OverBar;}{x})(y_t-\stackrel{\&OverBar;}{y})}{\sqrt{\underset{t=1}{\overset{n_l}{\mathrm{\&Sigma;}}}{(x_{t+l}-\stackrel{\&OverBar;}{x})}^2{(y_t-\stackrel{\&OverBar;}{y})}^2}},$ l=0，±1，±2，…，±12，

Y=(y in the following formula ₁, y ₂..., y _n) be reference index, X=(x ₁, x ₂..., x _n) being selected index, r is related coefficient, and l represents leading or lag period (being the time difference), represents when l gets negative in advance, represents when getting positive number to lag behind, and l is called as the time difference or postpones number,

With

Be respectively the mean value of sequence X and Y.n _lIt is the data amount check of all indexs.Then maximum time difference related coefficient has been considered to reflect the time difference correlationship of selected index and reference index, postpones accordingly number l and represents leading or lag period, namely so that time difference relative coefficient r _lLeading or the lag period that l is exactly this selected index and reference index is counted in maximum delay.

Fig. 5 is according to the method synoptic diagram that obtains coincidence indicator and leading indicators embodiment illustrated in fig. 4, particularly, is getting access to relative coefficient r _lAfterwards, obtain leading indicators and coincidence indicator by method shown in Figure 5:

Step S302: detect achievement data whether the time difference＜-3 and relative coefficient 0.7 or achievement data whether-2≤time difference≤2 and relative coefficient 0.7.Wherein, achievement data the time difference＜-3 and relative coefficient 0.7 or achievement data eligible-2≤time difference≤2 and relative coefficient in 0.7 the situation, execution in step S304, achievement data do not meet the time difference＜-3 and relative coefficient 0.7, and do not meet-2≤time difference≤2 and relative coefficient in 0.7 the situation, execution in step S310: abandon achievement data.

Step S304: obtain initial leading indicators and initial coincidence indicator, wherein, achievement data the time difference＜-3 and relative coefficient be initial leading indicators with this data decimation in 0.7 the situation, achievement data-2≤time difference≤2 and relative coefficient in 0.7 the situation, choosing this achievement data is initial coincidence indicator.

In above-described embodiment of the application, according to step-out time analysis model and/or K-L quantity of information model warning index is carried out correlation calculations, with the relative coefficient between each selected index and the reference index, and according to relative coefficient selected index is screened, comprise with the step of obtaining leading indicators and coincidence indicator: obtain relative coefficient r between each selected index and the reference index according to following formula _l:

Wherein, l=0, ± 1, ± 2 ..., ± 12, Y=(y ₁, y ₂..., y _n) be reference index, X=(x ₁, x ₂..., x _n) being selected index, l is the time difference, n _lNumber for all indexs; With time difference value in the first span and relative coefficient r _lGreater than the selected index of the first threshold Raw performance as leading indicators, and with time difference value in the second span and relative coefficient r _lGreater than the selected index of the Second Threshold Raw performance as coincidence indicator; The Raw performance of reference index, leading indicators and the Raw performance of coincidence indicator are carried out standardization, to obtain standard basis index series p _t, the selected index of standard sequence q _t, wherein, the selected index of standard comprises standard leading indicators and standard coincidence indicator; Obtain K-L quantity of information k between the selected index of each standard and the standard basis index by following formula _l:

k _l=∑ p _tLn (p _t/ q _T+1), wherein, l=0, ± 1 ..., ± 12,

T=1,2 ..., n, l are the time difference, n _lNumber for all indexs; With time difference value in the 3rd span and K-L quantity of information k _lLess than the selected index of the standard of the 3rd threshold value as leading indicators, and with time difference value in the 4th span and K-L quantity of information k _lLess than the selected index of the 4th threshold value as coincidence indicator.Wherein, the 3rd span can be less than-3, and the 3rd threshold value can be that 0.3, the four span can be more than or equal to-2 and can be 0.3 less than or equal to 2, the four threshold values.

Said method can be realized as follows: reference index is done standardization, and the standard basis index series after the processing is designated as p _t:

$p_t=y_t/\underset{t=1}{\overset{n}{\mathrm{\&Sigma;}}}{y}_t,$ t＝1，2，…,n，

Leading indicators and the coincidence indicator of primary election are also done standardization, and the sequence of the selected index of the standard after the processing is designated as q _t:

$q_t=x_t/\underset{t=1}{\overset{n}{\mathrm{\&Sigma;}}}{x}_t,$ t＝1，2，…,n，

Then, be calculated as follows behind each primary election beacon delay l K-L quantity of information k about reference index _l:

k _l=∑p _tln(p _t/q _t+l)＝0，±1，…,±12

Wherein, l represents leading or lag period, represents when l gets negative in advance, represents when getting positive number to lag behind, and l is called as the time difference, n _lBe the data amount check (i.e. the number of all indexs) after data are evened up, the t in the above-mentioned formula represents month, and i represents number of weeks.

After calculating 2L+1 K-L quantity of information, from k _lSelect a minimum value k in the value _{L '}As the K-L quantity of information of selected index x about reference index y, namely

L is counted in its corresponding delay ^*Be exactly the optimal leading or month number (season) that lags behind of selected index, wherein, the K-L quantity of information illustrates that more close to 0 index x and reference index y are more approaching.

Leading indicators can be screened as follows particularly and coincidence indicator is designated as respectively W ⁽³⁾And Z ⁽³⁾:

Step S306: detect initial leading indicators whether the time difference＜-3 and K-L quantity of information＜0.3 or initial consistent data whether-2≤time difference≤2 and K-L quantity of information＜0.3.Wherein, initial leading indicators meet the time difference＜-3 and K-L quantity of information＜0.3 or initial consistent data meet-situation of 2≤time difference≤2 and K-L quantity of information＜0.3 under, execution in step S308, initial leading indicators do not meet the time difference＜-3 and K-L quantity of information＜0.3, execution in step S310: abandon achievement data, do not meet at initial consistent data-situation of 2≤time difference≤2 and K-L quantity of information＜0.3 under, execution in step S310: abandon achievement data.

Step S308: selected leading indicators and coincidence indicator.Wherein, initial leading indicators the time difference＜-3 and the situation of K-L quantity of information＜0.3 under, selected this index is existing index, in initial coincidence indicator data, in the situation of-2≤time difference≤2 and K-L quantity of information＜0.3, selected this index is coincidence indicator.For example: choose delay number l=-2 ,-3 ,-4 index postpones number l=-1 as the leading indicators of primary election, and 0,1 index is as the coincidence indicator of primary election.

Above-described embodiment according to the application, according to the composite index number model leading indicators and coincidence indicator are synthesized, comprise to obtain as the in advance composite index number of early-warning parameters and the step of coincident composite Index: leading indicators and coincidence indicator are carried out respectively symmetrical change process, to obtain the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t), wherein, by following formula leading indicators is carried out symmetrical change process, to obtain the symmetrical rate of change C of leading indicators _{W, i}(t):

Wherein,

Be i (i=1,2 ..., k _w) individual leading indicators, t=2,3 ..., n, k _wThe number of leading indicators; By following formula coincidence indicator is carried out symmetrical change process, to obtain the symmetrical rate of change C of coincidence indicator _{Z, i}(t):

Wherein,

Be i (i=1,2 ..., k _z) individual coincidence indicator, t=2,3 ..., n, k _zIt is the number of coincidence indicator; To the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t) result who carries out obtaining after standardization and the trend adjustment synthesize calculating, to obtain go ahead of the rest composite index number and coincident composite Index.Wherein, the t in the above-mentioned formula represents month.

Particularly, ask the symmetrical rate of change of index and with its standardization with the index synthetic model, right respectively according to following formula

With

Ask symmetrical rate of change C _{W, i}(t) and C _{Z, i}(t):

$C_{w,i}\left(t\right)=200\&times;\frac{W_i^{\left(3\right)}\left(t\right)-W_i^{\left(3\right)}(t-1)}{W_i^{\left(3\right)}\left(t\right)+W_i^{\left(3\right)}(t-1)},$ t=2,3,…,n，

T=2,3 ..., n, wherein,

Be i (i=1,2 ..., k _w) individual leading indicators,

Be i (i=1,2 ..., k _z) individual coincidence indicator, k _wAnd k _zIt is respectively the number of leading indicators and coincidence indicator.

In above-described embodiment of the application, to the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t) result who carries out obtaining after standardization and the trend adjustment synthesize calculating, comprises with the step of obtaining go ahead of the rest composite index number and coincident composite Index: obtain normalization factor A by following formula _{W, i}And A _{Z, i}:

T=2,3 ..., n; Adopt normalization factor A _{W, i}And A _{Z, i}Respectively with the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t) carry out standardization, to obtain standardization rate of change S _{W, i}(t) and S _{Z, i}(t), wherein,

$S_{w,i}\left(t\right)=\frac{C_{w,i}\left(t\right)}{A_{w,i}},$ $S_{z,i}\left(t\right)=\frac{C_{z,i}\left(t\right)}{A_{z,i}},$ t=2,3,…,n；

To standardization rate of change S _{W, i}(t) and S _{Z, i}(t) average rate of change and process, to obtain the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t); Standardization average rate of change V according to leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t) synthesize calculating, to obtain in advance composite index number I _w(t) and coincident composite Index I _z(t), wherein, $I_w\left(t\right)=I_w(t-1)\&times;\frac{200+V_w\left(t\right)}{200-V_w\left(t\right)},$ $I_z\left(t\right)=I_z(t-1)\&times;\frac{200+V_z\left(t\right)}{200-V_z\left(t\right)},$ And I _w(1)=100, I _z(1)=100.

Particularly, according to following formula to the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t) do standardized calculation, make its average absolute value equal 1, obtain normalization factor A _{W, i}And A _{Z, i}:

I _t， $A_{z,i}=\underset{t=2}{\overset{n}{\mathrm{\&Sigma;}}}\frac{\left|C_{z,i}\left(t\right)\right|}{n-1},$ t=2,3,…,n，

Then according to A _{W, i}And A _{Z, i}Respectively with C _{W, i}(t) and C _{Z, i}(t) standardization obtains standardization rate of change S _{W, i}(t) and S _{Z, i}(t):

T=2,3 ..., n is afterwards according to standardization rate of change S _{W, i}(t) and S _{Z, i}(t) average rate of change and process, to obtain the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t), and according to the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t) synthesize calculating, to obtain in advance composite index number I _w(t) and coincident composite Index I _z(t), particularly: I _w(1)=100, I _z(1)=100, then

$I_w\left(t\right)=I_w(t-1)\&times;\frac{200+V_w\left(t\right)}{200-V_w\left(t\right)},$ $I_z\left(t\right)=I_z(t-1)\&times;\frac{200+V_z\left(t\right)}{200-V_z\left(t\right)}.$

More specifically, getting access in advance composite index number I _w(t) and coincident composite Index I _z(t) afterwards, can adjust electricity needs trend, method is as follows:

According to following compound interest formula each sequence of coincidence indicator group is obtained respectively separately average growth rate:

$r_i=(\sqrt[m_i]{C_{\mathrm{Li}}/C_{\mathrm{Ii}}}-1)\&times;100,$ i＝1,2,…,k _z，

Wherein, as shown in Figure 6, t is month,

With

Be respectively i index of coincidence indicator group at first with the most metacyclic mean value, m _IiWith m _LiBe respectively i index of coincidence indicator group at first with the most metacyclic month number, k ₂The coincidence indicator number, m _iThat the center of circulating at first is to the moon number between the most metacyclic center.

Then obtain the average growth rate G of coincidence indicator group _r, and with it as target trend:

Afterwards in advance and the initial composite index number I of coincidence indicator _w(t) and I _z(t) obtain their average growth rate r ' separately with compound interest formula respectively _wAnd r ' _z:

$r_w^{\&prime;}(\sqrt[m_w]{C_{\mathrm{Lw}}/C_{\mathrm{Iw}}}-1)\&times;100,$ $r_z^{\&prime;}(\sqrt[m_z]{C_{\mathrm{Lz}}/C_{\mathrm{Iz}}}-1)\&times;100,$

Wherein,

Again to the standardization average rate of change V of leading indicators group and coincidence indicator group _w(t) and V _z(t) do trend adjustment:

V′ _w(t)=V _w(t)+(Gr-r′ _w)，V′ _z(t)=V _z(t)+(Gr-r′ _z)，t=2,3,…,n。Then calculate composite index number according to the method in above-described embodiment: make I ' _w(1)=100, I ' _z(1)=100, then

${I^{\&prime;}}_w\left(t\right)={I^{\&prime;}}_w(t-1)\&times;\frac{200+{V^{\&prime;}}_w\left(t\right)}{200-{V^{\&prime;}}_w\left(t\right)},$ ${I^{\&prime;}}_z\left(t\right)={I^{\&prime;}}_z(t-1)\&times;\frac{200+{V^{\&prime;}}_z\left(t\right)}{200-{V^{\&prime;}}_z\left(t\right)},$

Generation is take the in advance composite index number CI of benchmark time as 100 _w(t) and coincident composite Index CI _z(t):

${\mathrm{CI}}_w\left(t\right)=(I_w^{\&prime;}\left(t\right)/\stackrel{\&OverBar;}{I_w^{\&prime;}}\&times;100),$ ${\mathrm{CI}}_z\left(t\right)=(I_z^{\&prime;}\left(t\right)/\stackrel{\&OverBar;}{I_z^{\&prime;}})\&times;100,$

Wherein

With

Respectively I ' _w(t) and I ' _z(t) at the mean value in benchmark time.

According to above-described embodiment of the application, to standardization rate of change S _{W, i}(t) and S _{Z, i}(t) average rate of change and process, to obtain the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t) step comprises:

By following formula respectively with the standardization rate of change S of leading indicators _{W, i}(t) and the standardization rate of change S of coincidence indicator _{Z, i}(t) average rate of change and process, to obtain the average rate of change R of leading indicators _w(t) and the average rate of change R of coincidence indicator _z(t):

$R_w\left(t\right)=\frac{\underset{i=1}{\overset{k_w}{\mathrm{\&Sigma;}}}{S}_{w,i}\left(t\right){\mathrm{\&lambda;}}_{w,i}}{\underset{i=1}{\overset{k_w}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_{w,i}},$ $R_z\left(t\right)=\frac{\underset{i=1}{\overset{k_z}{\mathrm{\&Sigma;}}}{S}_{z,i}\left(t\right){\mathrm{\&lambda;}}_{z,i}}{\underset{i=1}{\overset{k_z}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_{z,i}},$ Wherein, λ _{W, i}And λ _{Z, i}It is respectively the weight of i index of leading indicators and coincidence indicator; Obtain index normalization factor F by following formula _w:

$F_w=\left[\underset{t=2}{\overset{n}{\mathrm{\&Sigma;}}}\right|R_w\left(t\right)|/(n-1)]/\left[\underset{t=2}{\overset{n}{\mathrm{\&Sigma;}}}\right|R_z\left(t\right)|/(n-1)];$ According to index normalization factor F _wCarry out the standardization average rate of change and process, to obtain the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t), wherein, V _w(t)=R _w(t)/F _w, V _z(t)=R _z(t).

Particularly, obtaining initial composite index number I _w(t) and I _z(t) before, by following formula respectively with the standardization rate of change S of leading indicators _{W, i}(t) and the standardization rate of change S of coincidence indicator _{Z, i}(t) average rate of change and process, to obtain the average rate of change R of leading indicators _w(t) and the average rate of change R of coincidence indicator _z(t):

$R_w\left(t\right)=\frac{\underset{i=1}{\overset{k_w}{\mathrm{\&Sigma;}}}{S}_{w,i}\left(t\right){\mathrm{\&lambda;}}_{w,i}}{\underset{i=1}{\overset{k_w}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_{w,i}},$ $R_z\left(t\right)=\frac{\underset{i=1}{\overset{k_z}{\mathrm{\&Sigma;}}}{S}_{z,i}\left(t\right){\mathrm{\&lambda;}}_{z,i}}{\underset{i=1}{\overset{k_Z}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_{z,i}},$ λ _{W, i}And λ _{Z, i}Respectively in advance and the weight of i index of coincidence indicator group.

Then according to following formula gauge index normalization factor F _w:

$F_w=\left[\underset{t=2}{\overset{n}{\mathrm{\&Sigma;}}}\right|R_w\left(t\right)|/(n-1)]/\left[\underset{t=2}{\overset{n}{\mathrm{\&Sigma;}}}\right|R_z\left(t\right)|/(n-1)];$

At last according to index normalization factor F _wCarry out the standardization average rate of change and process, to obtain the standardization average rate of change V of leading indicators _w(t) and the standardization average rate of change V of coincidence indicator _z(t):

V _w(t)=R _w(t)/F _w, V _z(t)=R _z(t), t=2,3 ..., n wherein, goes to adjust the average rate of change of leading indicators sequence and lagging indicator sequence with the amplitude of the average rate of change of coincidence indicator sequence, its objective is for two indexes are used as a harmonious system and uses.

Need to prove, can in the computer system such as one group of computer executable instructions, carry out in the step shown in the process flow diagram of accompanying drawing, and, although there is shown logical order in flow process, but in some cases, can carry out step shown or that describe with the order that is different from herein.

From above description, can find out, the present invention has realized following technique effect: method and the device of the early-warning parameters that obtains electricity needs by the application, after the trend term and periodic term in obtaining original data sequence, by data sequence screening and calculating are obtained leading indicators and coincidence indicator, then above-mentioned leading indicators and coincidence indicator are synthesized the acquisition early-warning parameters with the index synthetic model, and according to the fluctuation of early-warning parameters analysis power demand cycle, having solved in the prior art adopts forecasting techniques to obtain short-term cyclic fluctuation distortion in the electricity needs field, can't acquire the early-warning parameters that meets electricity needs, thereby cause and to formulate the reasonably measure of reply cyclic fluctuation according to the power cycle fluctuation, realized accurately obtaining the early-warning parameters of electricity needs, thereby the effect of the counter-measure of reasonable science is formulated in cyclic fluctuation according to short-term accurately, and then the amplitude of having slowed down cyclic swing, reduce the destructiveness that cyclic swing causes power industry and economic development.

Obviously, those skilled in the art should be understood that, above-mentioned each module of the present invention or each step can realize with general calculation element, they can concentrate on the single calculation element, perhaps be distributed on the network that a plurality of calculation elements form, alternatively, they can be realized with the executable program code of calculation element, thereby, they can be stored in the memory storage and be carried out by calculation element, perhaps they are made into respectively each integrated circuit modules, perhaps a plurality of modules in them or step are made into the single integrated circuit module and realize.Like this, the present invention is not restricted to any specific hardware and software combination.

The above is the preferred embodiments of the present invention only, is not limited to the present invention, and for a person skilled in the art, the present invention can have various modifications and variations.Within the spirit and principles in the present invention all, any modification of doing, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (14)

1. a method of obtaining the early-warning parameters of electricity needs is characterized in that, comprising:

Obtain for the data sequence that generates warning index;

According to adjusting parameter described data sequence is screened, to obtain the data sequence that includes trend term and periodic term;

Calculate the described trend index that includes the data sequence of described trend term and periodic term, and according to described trend index the described data sequence that includes described trend term and periodic term is filtered, to obtain the warning index sequence, described warning index sequence is the data of trend index for increasing in the described data sequence;

The generated energy that extracts in the described warning index sequence is reference index, and the index of extracting except described generated energy is selected index;

According to step-out time analysis model and/or K-L quantity of information model described warning index is carried out correlation calculations, to obtain the relative coefficient between each described selected index and the described reference index, and according to described relative coefficient described selected index is screened, to obtain leading indicators and coincidence indicator;

According to the composite index number model described leading indicators and coincidence indicator are synthesized, to obtain in advance composite index number and the coincident composite Index as early-warning parameters.

2. method according to claim 1, it is characterized in that, utilize step-out time analysis model and/or K-L quantity of information model that described warning index is carried out correlation calculations, with the relative coefficient between each described selected index and the described reference index, and according to described relative coefficient described selected index is screened, comprise with the step of obtaining leading indicators and coincidence indicator:

Obtain relative coefficient r between each described selected index and the described reference index according to following formula _l:

$r_l=\frac{\underset{t=1}{\overset{n_l}{\mathrm{\&Sigma;}}}(x_{t+l}-\stackrel{\&OverBar;}{x})(y_t-\stackrel{\&OverBar;}{y})}{\sqrt{\underset{t=1}{\overset{n_l}{\mathrm{\&Sigma;}}}{(x_{t+l}-\stackrel{\&OverBar;}{x})}^2{(y_t-\stackrel{\&OverBar;}{y})}^2}},$

Wherein, l=0, ± 1, ± 2 ..., ± 12, Y=(y ₁, y ₂..., y _n) be reference index, X=(x ₁, x ₂..., x _n) being selected index, l is the time difference, n _lBe the number of all indexs, t=1,2 ..., n is a month umber;

With described time difference value in the first span and described relative coefficient r _lGreater than the selected index of first threshold as described leading indicators, and with described time difference value in the second span and described relative coefficient r _lGreater than the selected index of Second Threshold as described coincidence indicator.

3. method according to claim 2, it is characterized in that, according to step-out time analysis model and/or K-L quantity of information model described warning index is carried out correlation calculations, with the relative coefficient between each described selected index and the described reference index, and according to described relative coefficient described selected index is screened, comprise with the step of obtaining leading indicators and coincidence indicator:

Obtain relative coefficient r between each described selected index and the described reference index according to following formula _l:

$r_l=\frac{\underset{t=1}{\overset{n_l}{\mathrm{\&Sigma;}}}(x_{t+l}-\stackrel{\&OverBar;}{x})(y_t-\stackrel{\&OverBar;}{y})}{\sqrt{\underset{t=1}{\overset{n_l}{\mathrm{\&Sigma;}}}{(x_{t+l}-\stackrel{\&OverBar;}{x})}^2{(y_t-\stackrel{\&OverBar;}{y})}^2}},$

Wherein, l=0, ± 1, ± 2 ..., ± 12, Y=(y ₁, y ₂..., y _n) be reference index, X=(x ₁, x ₂..., x _n) being selected index, l is the time difference, n _lBe the number of all indexs, t=1,2 ..., n is a month umber;

With described time difference value in the first span and described relative coefficient r _lGreater than the selected index of the first threshold Raw performance as described leading indicators, and with described time difference value in the second span and described relative coefficient r _lGreater than the selected index of the Second Threshold Raw performance as described coincidence indicator;

The Raw performance of described reference index, described leading indicators and the Raw performance of described coincidence indicator are carried out standardization, to obtain standard basis index series p _t, the selected index of standard sequence q _t, wherein, the selected index of described standard comprises standard leading indicators and standard coincidence indicator;

Obtain K-L quantity of information k between the selected index of each standard and the described standard basis index by following formula _l:

k _l=∑ p _tLn (p _t/ q _T+1), wherein, l=0, ± 1 ..., ± 12,

T=1,2 ..., n is a month umber, l is the time difference, n _lNumber for all indexs;

With described time difference value in the 3rd span and described K-L quantity of information k _lLess than the selected index of the standard of the 3rd threshold value as described leading indicators, and with described time difference value in the 4th span and described K-L quantity of information k _lLess than the selected index of the 4th threshold value as described coincidence indicator.

4. according to claim 2 or 3 described methods, it is characterized in that, according to the composite index number model described leading indicators and coincidence indicator synthesized, comprise to obtain as the in advance composite index number of early-warning parameters and the step of coincident composite Index:

Described leading indicators and described coincidence indicator are carried out respectively symmetrical change process, to obtain the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t),

Wherein, by following formula described leading indicators is carried out symmetrical change process, to obtain the symmetrical rate of change C of described leading indicators _{W, i}(t):

Wherein,

Be i (i=1,2 ..., k _w) individual leading indicators, t=2,3 ..., n, k _wNumber for leading indicators;

By following formula described coincidence indicator is carried out symmetrical change process, to obtain the symmetrical rate of change C of coincidence indicator _{Z, i}(t):

Wherein,

Be i (i=1,2 ..., k _z) individual coincidence indicator, t=2,3 ..., n is a month umber, k _zIt is the number of coincidence indicator;

To the symmetrical rate of change C of described leading indicators _{W, i}(t) and the symmetrical rate of change C of described coincidence indicator _{Z, i}(t) result who carries out obtaining after standardization and the trend adjustment synthesize calculating, to obtain described composite index number and the coincident composite Index of going ahead of the rest.

5. method according to claim 4 is characterized in that, to the symmetrical rate of change C of described leading indicators _{W, i}(t) and the symmetrical rate of change C of described coincidence indicator _{Z, i}(t) result who carries out obtaining after standardization and the trend adjustment synthesize calculating, comprises to obtain described step of going ahead of the rest composite index number and coincident composite Index:

Obtain normalization factor A by following formula _{W, i}And A _{Z, i}:

T=2,3 ..., n;

Adopt described normalization factor A _{W, i}And A _{Z, i}Respectively with the symmetrical rate of change C of described leading indicators _{W, i}(t) and the symmetrical rate of change C of described coincidence indicator _{Z, i}(t) carry out standardization, to obtain standardization rate of change S _{W, i}(t) and S _{Z, i}(t), wherein,

$S_{w,i}\left(t\right)=\frac{C_{w,i}\left(t\right)}{A_{w,i}},$ $S_{z,i}\left(t\right)=\frac{C_{z,i}\left(t\right)}{A_{z,i}},$ t=2,3,…,n；

To described standardization rate of change S _{W, i}(t) and S _{Z, i}(t) average rate of change and process, to obtain the standardization average rate of change V of described leading indicators _w(t) and the standardization average rate of change V of described coincidence indicator _z(t);

Standardization average rate of change V according to described leading indicators _w(t) and the standardization average rate of change V of described coincidence indicator _z(t) synthesize calculating, to obtain described in advance composite index number I _w(t) and coincident composite Index I _z(t), wherein, $I_w\left(t\right)=I_w(t-1)\&times;\frac{200+V_w\left(t\right)}{200-V_{w}t},$ $I_z\left(t\right)=I_z(t-1)\&times;\frac{200+V_z\left(t\right)}{200-V_z\left(t\right)}.$ And I _w(1)=100, I _z(1)=100.

6. method according to claim 5 is characterized in that, to described standardization rate of change S _{W, i}(t) and S _{Z, i}(t) average rate of change and process, to obtain the standardization average rate of change V of described leading indicators _w(t) and the standardization average rate of change V of described coincidence indicator _z(t) step comprises:

By following formula respectively with the standardization rate of change S of described leading indicators _{W, i}(t) and the standardization rate of change S of described coincidence indicator _{Z, i}(t) average rate of change and process, to obtain the average rate of change R of described leading indicators _w(t) and the average rate of change R of described coincidence indicator _z(t):

$R_w\left(t\right)=\frac{\underset{i=1}{\overset{k_w}{\mathrm{\&Sigma;}}}{S}_{w,i}\left(t\right){\mathrm{\&lambda;}}_{w,i}}{\underset{i=1}{\overset{k_w}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_{w,i}},$ $R_z\left(t\right)=\frac{\underset{i=1}{\overset{k_z}{\mathrm{\&Sigma;}}}{S}_{z,i}\left(t\right){\mathrm{\&lambda;}}_{z,i}}{\underset{i=1}{\overset{k_z}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_{z,i}},$ Wherein, λ _{W, i}And λ _{Z, i}It is respectively the weight of i index of leading indicators and coincidence indicator;

Obtain index normalization factor F by following formula _w:

$F_w=\left[\underset{t=2}{\overset{n}{\mathrm{\&Sigma;}}}\right|R_w\left(t\right)|/(n-1)]/\left[\underset{t=2}{\overset{n}{\mathrm{\&Sigma;}}}\right|R_z\left(t\right)|/(n-1)];$

According to described index normalization factor F _wCarry out the standardization average rate of change and process, to obtain the standardization average rate of change V of described leading indicators _w(t) and the standardization average rate of change V of described coincidence indicator _z(t), wherein, V _w(t)=R _w(t)/F _w, V _z(t)=R _z(t).

7. method according to claim 1 is characterized in that, after obtaining for the data sequence that generates warning index, described method also comprises:

Data in the described data sequence are carried out pre-service, and described pre-service comprises: fill up missing data and process, revise noise data processing, data smoothing processing and data normalization processing.

8. a device that obtains the early-warning parameters of electricity needs is characterized in that, comprising:

The first acquisition module is used for obtaining the data sequence for generating warning index;

The first processing module is used for according to adjusting parameter described data sequence being screened, to obtain the data sequence that includes trend term and periodic term;

The first computing module, be used for calculating the described trend index that includes the data sequence of described trend term and periodic term, and according to described trend index the described data sequence that includes described trend term and periodic term is filtered, to obtain the warning index sequence, described warning index sequence is the data of trend index for increasing in the described data sequence;

The first extraction module is reference index for the generated energy that extracts described warning index sequence, and the index of extracting except described generated energy is selected index;

The second computing module, be used for according to step-out time analysis model and/or K-L quantity of information model described warning index being carried out correlation calculations, to obtain the relative coefficient between each described selected index and the described reference index, and according to described relative coefficient described selected index is screened, to obtain leading indicators and coincidence indicator;

The second processing module is used for according to the composite index number model described leading indicators and coincidence indicator being synthesized, to obtain in advance composite index number and the coincident composite Index as early-warning parameters.

9. device according to claim 8 is characterized in that, described the second computing module comprises:

The first sub-computing module is for the relative coefficient r that obtains according to following formula between each described selected index and the described reference index _l:

Wherein, l=0, ± 1, ± 2 ..., ± 12, Y=(y ₁, y ₂..., y _n) be reference index, X=(x ₁, x ₂..., x _n) being selected index, l is the time difference, n _lBe the number of all indexs, t=1,2 ..., n is a month umber;

The first sub-processing module is used for described time difference value in the first span and described relative coefficient r _lGreater than the selected index of first threshold as described leading indicators, and with described time difference value in the second span and described relative coefficient r _lGreater than the selected index of Second Threshold as described coincidence indicator.

10. device according to claim 9 is characterized in that, described the second computing module comprises:

The second sub-computing module is for the relative coefficient r that obtains according to following formula between each described selected index and the described reference index _l:

Wherein, l=0, ± 1, ± 2 ..., ± 12, Y=(y ₁, y ₂..., y _n) be reference index, X=(x ₁, x ₂..., x _n) being selected index, l is the time difference, n _lBe the number of all indexs, t=1,2 ..., n is a month umber;

The second sub-processing module is used for described time difference value in the first span and described relative coefficient r _lGreater than the selected index of the first threshold Raw performance as described leading indicators, and with described time difference value in the second span and described relative coefficient r _lGreater than the selected index of the Second Threshold Raw performance as described coincidence indicator;

The 3rd sub-processing module is used for the Raw performance of described reference index, described leading indicators and the Raw performance of described coincidence indicator are carried out standardization, to obtain standard basis index series p _t, the selected index of standard sequence q _t, wherein, the selected index of described standard comprises standard leading indicators and standard coincidence indicator;

The 3rd sub-computing module is for the K-L quantity of information k that obtains by following formula between the selected index of each standard and the described standard basis index _l:

k _l=∑ p _tLn (p _t/ q _T+1), wherein, l=0, ± 1 ..., ± 12,

T=1,2 ..., n is a month umber, l is the time difference, n _lNumber for all indexs;

The 4th sub-processing module is used for described time difference value in the 3rd span and described K-L quantity of information k _lLess than the selected index of the standard of the 3rd threshold value as described leading indicators, and with described time difference value in the 4th span and described K-L quantity of information k _lLess than the selected index of the 4th threshold value as described coincidence indicator.

11. according to claim 9 or 10 described devices, it is characterized in that described the second processing module comprises:

The 5th sub-processing module is used for described leading indicators and described coincidence indicator are carried out respectively symmetrical change process, to obtain the symmetrical rate of change C of leading indicators _{W, i}(t) and the symmetrical rate of change C of coincidence indicator _{Z, i}(t), the described the 5th sub-processing module comprises:

The 4th sub-computing module is used for by following formula described leading indicators being carried out symmetrical change process, to obtain the symmetrical rate of change C of described leading indicators _{W, i}(t):

Wherein,

Be i (i=1,2 ..., k _w) individual leading indicators, t=2,3 ..., n is a month umber, k _wNumber for leading indicators;

The 5th sub-computing module is used for by following formula described coincidence indicator being carried out symmetrical change process, to obtain the symmetrical rate of change C of coincidence indicator _{Z, i}(t):

Wherein,

Be i (i=1,2 ..., k _z) individual coincidence indicator, t=2,3 ..., n, k _zIt is the number of coincidence indicator;

The 6th sub-processing module is used for the symmetrical rate of change C of described leading indicators _{W, i}(t) and the symmetrical rate of change C of described coincidence indicator _{Z, i}(t) result who carries out obtaining after standardization and the trend adjustment synthesize calculating, to obtain described composite index number and the coincident composite Index of going ahead of the rest.

12. device according to claim 11 is characterized in that, the described the 6th sub-processing module comprises:

The 6th sub-computing module is used for obtaining normalization factor A by following formula _{W, i}And A _{Z, i}:

$A_{z,i}=\underset{t=2}{\overset{n}{\mathrm{\&Sigma;}}}\frac{\left|C_{z,i}\left(t\right)\right|}{n-1},$ t=2,3,…,n；

The 7th sub-processing module is used for adopting described normalization factor A _{W, i}And A _{Z, i}Respectively with the symmetrical rate of change C of described leading indicators _{W, i}(t) and the symmetrical rate of change C of described coincidence indicator _{Z, i}(t) carry out standardization, to obtain standardization rate of change S _{W, i}(t) and S _{Z, i}(t), wherein,

$S_{w,i}\left(t\right)=\frac{C_{w,i}\left(t\right)}{A_{w,i}},$ $S_{z,i}\left(t\right)=\frac{C_{z,i}\left(t\right)}{A_{z,i}},$ t=2,3,…,n；

The 8th sub-processing module is used for described standardization rate of change S _{W, i}(t) and S _{Z, i}(t) average rate of change and process, to obtain the standardization average rate of change V of described leading indicators _w(t) and the standardization average rate of change V of described coincidence indicator _z(t);

The 7th sub-computing module is used for the standardization average rate of change V according to described leading indicators _w(t) and the standardization average rate of change V of described coincidence indicator _z(t) synthesize calculating, to obtain described in advance composite index number I _w(t) and coincident composite Index I _z(t), wherein, $I_w\left(t\right)=I_w(t-1)\&times;\frac{200+V_w\left(t\right)}{200-V_{w}t},$ $I_z\left(t\right)=I_z(t-1)\&times;\frac{200+V_z\left(t\right)}{200-V_z\left(t\right)}.$ And I _w(1)=100, I _z(1)=100.

13. device according to claim 12 is characterized in that, the described the 8th sub-processing module comprises:

The 9th sub-processing module is used for by following formula respectively with the standardization rate of change S of described leading indicators _{W, i}(t) and the standardization rate of change S of described coincidence indicator _{Z, i}(t) average rate of change and process, to obtain the average rate of change R of described leading indicators _w(t) and the average rate of change R of described coincidence indicator _z(t):

$R_w\left(t\right)=\frac{\underset{i=1}{\overset{k_w}{\mathrm{\&Sigma;}}}{S}_{w,i}\left(t\right){\mathrm{\&lambda;}}_{w,i}}{\underset{i=1}{\overset{k_w}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_{w,i}},$ $R_z\left(t\right)=\frac{\underset{i=1}{\overset{k_z}{\mathrm{\&Sigma;}}}{S}_{z,i}\left(t\right){\mathrm{\&lambda;}}_{z,i}}{\underset{i=1}{\overset{k_z}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_{z,i}},$ Wherein, λ _{W, i}And λ _{Z, i}It is respectively the weight of i index of leading indicators and coincidence indicator;

The 8th sub-computing module is used for obtaining index normalization factor F by following formula _w:

$F_w=\left[\underset{t=2}{\overset{n}{\mathrm{\&Sigma;}}}\right|R_w\left(t\right)|/(n-1)]/\left[\underset{t=2}{\overset{n}{\mathrm{\&Sigma;}}}\right|R_z\left(t\right)|/(n-1)];$

The 9th sub-computing module is used for according to described index normalization factor F _wCarry out the standardization average rate of change and process, to obtain the standardization average rate of change V of described leading indicators _w(t) and the standardization average rate of change V of described coincidence indicator _z(t), wherein, V _w(t)=R _w(t)/F _w, V _z(t)=R _z(t).

14. device according to claim 8 is characterized in that, after carrying out acquisition module, described device also comprises:

The tenth sub-processing module is used for the data of described data sequence are carried out pre-service, and described pre-service comprises: fill up missing data processing, the processing of correction noise data, data smoothing processing and data normalization and process.

Images