CN109993370A - A kind of electric power sale day cash flow projections method based on nonstationary time series - Google Patents
A kind of electric power sale day cash flow projections method based on nonstationary time series Download PDFInfo
- Publication number
- CN109993370A CN109993370A CN201910282578.3A CN201910282578A CN109993370A CN 109993370 A CN109993370 A CN 109993370A CN 201910282578 A CN201910282578 A CN 201910282578A CN 109993370 A CN109993370 A CN 109993370A
- Authority
- CN
- China
- Prior art keywords
- day
- month
- days
- prediction
- cash flow
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The present invention relates to big data processing, it is desirable to provide a kind of electric power sale day cash flow projections method based on nonstationary time series.Include: to read power department historical sales pipelined data and pre-processed, checks missing data and apparent abnormal data, then arrange payment amount of money information and time cycle feature；Stationary test is carried out to per total value month in and month out, then does ARIMA fitting prediction, and reserve inspection set and be tested for the property to obtained prediction model；All total value time serieses ARIMA fitting prediction will be done to it after difference and by unit root test weekly；It establishes the day cash flow regression model an of middle of the month surrounding respectively with the method for regression analysis, and carries out the prediction of electric power sale day cash flow using it.Structure of the present invention can classify to different times, period, and market survey, marketing plan is facilitated to formulate, and it is simply more accurate, reasonable based on the prediction of stationary time series curvilinear trend to compare, and accuracy is promoted.
Description
Technical field
The present invention relates to big data technology, in particular to a kind of electric power based on nonstationary time series sells day cash flow
Prediction technique.
Background technique
Electric power sale day cash flow projections refer in the following specific time, carry out to cash flow in daily power grid account pre-
It surveys.On the basis of day cash forecast is intended to fully consider the following various influence factors, helps enterprise to carry out financial budget, strengthening
In terms of storage fund Efficient Operation, help gradually to carry out unit storage fund " zero remaining sum " management, in refinement flow fund lean
Management aspect, help implement cash flow budget and monthly work out per diem scheduling.Development plan and strategic plan of its result to power grid
Suffer from very important meaning.
However it carries out high-precision being per diem that unit prediction is not easy to.It is presently available for the data mining work of sales forecast
Have main linear or nonlinear regression model (NLRM), Grey System Model, maximum entropy Markov model etc..Although these models are
It is widely used in economic forecasting field, but reason results in these prediction techniques in electric power sale day cash in terms of following two
Accuracy in stream prediction not can guarantee, and limit their practical application yet.On the one hand influence daily power selling income because
Element is very more.The factor for influencing daily power selling income not only has the average electricity consumption of user, also with season, festivals or holidays, user
Pay charge way and behavior pattern etc. are related.The electricity consumption of resident often has apparent seasonal effect, winter and summer
Since air-conditioning frequency of use is high, cause the electricity consumption in the two seasons can be relatively more, the paying behaviors at weekend and non-weekend
Have apparent difference；On the other hand the user volume being related to is very big.By taking Zhejiang Province's grid company as an example (including company home office,
11 prefecture-level companies, 64 companies, county and all subordinate units), share the user more than 26,000,000.If each client is seen
At being a dimension, then here it is the big data analysis problems of typical big dimension (p) small sample (n).
In view of the foregoing, traditional statistical analysis model is difficult to be fitted multifactor interaction, can not solve electric power sale
The accurate prediction of day cash flow.
Summary of the invention
The technical problem to be solved by the present invention is to overcome deficiency in the prior art, provide a kind of based on the non-stationary time
The electric power of sequence sells day cash flow projections method.
In order to solve the technical problem, solution of the invention is:
A kind of electric power sale day cash flow projections method based on nonstationary time series is provided, comprising the following steps:
(1) power department historical sales pipelined data is read；
(2) sales data is pre-processed, missing data and apparent abnormal data is checked, then according to year, month, day
Arrange payment amount of money information and time cycle feature；
(3) it carries out moon prediction: the every total value (total amount of monthly paying the fees) month in and month out sorted out in step (2) being carried out flat
Stability is examined, and ARIMA fitting prediction is then done to it, and reserve inspection set and be tested for the property to obtained prediction model；
(4) carry out weekly forecasting: the total value time series all weekly that will be sorted out in step (2) is by difference and passes through list
After position root is examined, ARIMA fitting prediction is done to it；
(5) it carries out day prediction: establishing the day cash flow regression model an of middle of the month surrounding respectively with the method for regression analysis,
And the prediction of electric power sale day cash flow is carried out using it.
In the present invention, in step (1) in the historical sales pipelined data of reading, including user's industry, reality arrive account day
Phase section, pay charge way and payment amount of money information, while according to practical Payouts to Dat section differentiate what day and whether save vacation
Day (a variety of integrated informations can effectively help to model).
In the present invention, in the step (2), step is specifically included:
(2.1) to overcome influence of the different calendar month number of days differences to time series models, for as unit of one month
12 months are cycle fit time series models, the number of days of each month is adjusted to 30 days；It, will for there is 31 days months
As the 30th day assignment after 31st day value and the 30th day value are average；For there was only 29 days months, make tax in the 30th day
It is worth equal with the 29th day value；For there was only 28 days months, the value of the 28th day and the 1st day next month is assigned to the 29th respectively
It and the 30th day；
(2.2) for as unit of one week, 4 weeks be cycle fit time series models when, the number of days of each month is equal
It is adjusted to 28 days；Number of days is once removed into average value to the previous day more than or equal to 28 days parts, be adapted to the 28th day data it
In.
In the present invention, in the step (3), after to stationary test is carried out per total value month in and month out, first-order difference is first carried out
And unit root test, then it is fitted with ARIMA sequence.
In the present invention, in the step (4), using 28 days one month put in order in step (2) and with all information
Data, annual average weekly is calculated to obtain a new sequence, then carry out the building of week series model.
It include: first in " day cash flow valuve=moon total value × week ratio × day ratio " in the present invention, in the step (5)
Principle framework proposes day cash flow tentative prediction model, and the accounting of every day in one week is then gone out with test set data regression.
The principle of the present invention description:
It gives a forecast for the daily consumption sum of auxiliary power department statistics and to the following charge trend and amount of money range, this hair
It is bright to be extracted previous user information, time series feature, as input, make full use of the moon, week, three free periods of day
Rule grasps its feature, excavates they and the non-future to contacting between the account amount of money, to obtain being predicted not according to previous data
The model of future cash flow.
Compared with prior art, the beneficial effects of the present invention are:
The present invention constructs a set of day cash flow budget model based on power department, can to different times, the period into
Row classification facilitates market survey, marketing plan to formulate, bent based on stationary time series (using ARIMA as representative) compared to simple
Line trend prediction is more accurate, reasonable, and accuracy is promoted.
Detailed description of the invention
Fig. 1 is the building process schematic diagram of nonstationary time series model of the present invention.
Specific embodiment
Present invention is further described in detail with specific embodiment with reference to the accompanying drawing.
Electric power in the present invention based on nonstationary time series sells day cash flow projections method, comprising the following steps:
(1) power department historical sales pipelined data, including user's industry, practical Payouts to Dat section, payment side are read
The information of formula and payment amount of money, while what day is differentiated according to practical Payouts to Dat section and whether festivals or holidays are (a variety of comprehensive to believe
Breath can effectively help to model).
(2) sales data is pre-processed, missing data and apparent abnormal data is checked, then according to year, month, day
Arrange payment amount of money information and time cycle feature；Specifically include step:
(2.1) to overcome influence of the different calendar month number of days differences to time series models, for as unit of one month
12 months are cycle fit time series models, the number of days of each month is adjusted to 30 days；It, will for there is 31 days months
As the 30th day assignment after 31st day value and the 30th day value are average；For there was only 29 days months, make tax in the 30th day
It is worth equal with the 29th day value；For there was only 28 days months, the value of the 28th day and the 1st day next month is assigned to the 29th respectively
It and the 30th day；
(2.2) for as unit of one week, 4 weeks be cycle fit time series models when, the number of days of each month is equal
It is adjusted to 28 days；Number of days is once removed into average value to the previous day more than or equal to 28 days parts, be adapted to the 28th day data it
In.
(3) it carries out moon prediction: the every total value (total amount of monthly paying the fees) month in and month out sorted out in step (2) being carried out flat
Stability is examined, and first-order difference and unit root test are first carried out, and ARIMA fitting prediction is then done to it, and reserve inspection set to institute
Prediction model is obtained to be tested for the property；
(4) carry out weekly forecasting: the total value time series all weekly that will be sorted out in step (2) is by difference and passes through list
After position root is examined, ARIMA fitting prediction is done to it；
28 days one month and the data with all information put in order in specific usable step (2), by day weekly
Mean value computation comes out and obtains a new sequence, then carries out the building of week series model.
(5) it carries out day prediction: establishing the day cash flow regression model an of middle of the month surrounding respectively with the method for regression analysis；
Day cash flow tentative prediction model first is proposed in the principle framework of " day cash flow valuve=moon total value × week ratio × day ratio ", so
Go out the accounting of every day in one week with test set data regression afterwards.Finally, carrying out electricity using the day cash flow projections model obtained
The prediction of power sale day cash flow.
Finally it should be noted that the above enumerated are only specific embodiments of the present invention.It is clear that the invention is not restricted to
Above embodiments can also have many variations.Those skilled in the art can directly lead from present disclosure
Out or all deformations for associating, it is considered as protection scope of the present invention.
Claims (6)
1. a kind of electric power based on nonstationary time series sells day cash flow projections method, which is characterized in that including following step
It is rapid:
(1) power department historical sales pipelined data is read；
(2) sales data is pre-processed, checks missing data and apparent abnormal data, is then arranged according to year, month, day
Payment amount of money information and time cycle feature；
(3) it carries out moon prediction: carrying out stationary test per total value month in and month out to what is sorted out in step (2), then it is done
ARIMA fitting prediction, and reserve inspection set and obtained prediction model is tested for the property；
(4) carry out weekly forecasting: the total value time series all weekly that will be sorted out in step (2) is by difference and passes through unit root
After inspection, ARIMA fitting prediction is done to it；
(5) it carries out day prediction: establishing the day cash flow regression model an of middle of the month surrounding, and benefit respectively with the method for regression analysis
The prediction of electric power sale day cash flow is carried out with it.
2. the method according to claim 1, wherein in the historical sales pipelined data read in the step (1),
Including user's industry, practical Payouts to Dat section, pay charge way and payment amount of money information, while according to practical Payouts to Dat area
Between differentiate what day and whether festivals or holidays.
3. the method according to claim 1, wherein specifically including step in the step (2):
(2.1) to overcome influence of the different calendar month number of days differences to time series models, for one month for unit 12
The moon is cycle fit time series models, the number of days of each month is adjusted to 30 days；For there is 31 days months, by the 31st
As the 30th day assignment after it value and the 30th day value are average；For there was only 29 days months, make the 30th day assignment with
29th day value is equal；For there was only 28 days months, by the value of the 28th day and the 1st day next month be assigned to respectively the 29th day and
30th day；
(2.2) for as unit of one week, 4 weeks be cycle fit time series models when, the number of days of each month is adjusted
It is 28 days；Average value is once removed to the previous day in part by number of days more than or equal to 28 days, is adapted among the 28th day data.
4. the method according to claim 1, wherein being carried out steadily in the step (3) to per total value month in and month out
Property examine after, first carry out first-order difference and unit root test, then be fitted to it with ARIMA sequence.
5. the method according to claim 1, wherein using what is put in order in step (2) in the step (4)
28 days one month and the data with all information, annual average weekly are calculated to obtain a new sequence, then carry out
The building of week series model.
6. the method according to claim 1, wherein include: in the step (5) first by " day cash flow valuve=
The principle framework of month total value × week ratio × day ratio " proposes day cash flow tentative prediction model, then with test set data time
Return the accounting of every day in one week out.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910282578.3A CN109993370A (en) | 2019-04-10 | 2019-04-10 | A kind of electric power sale day cash flow projections method based on nonstationary time series |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910282578.3A CN109993370A (en) | 2019-04-10 | 2019-04-10 | A kind of electric power sale day cash flow projections method based on nonstationary time series |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109993370A true CN109993370A (en) | 2019-07-09 |
Family
ID=67132718
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910282578.3A Pending CN109993370A (en) | 2019-04-10 | 2019-04-10 | A kind of electric power sale day cash flow projections method based on nonstationary time series |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109993370A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112580880A (en) * | 2020-12-23 | 2021-03-30 | 中国交通信息科技集团有限公司 | Capacity prediction method, computer processor and terminal equipment |
| CN113128693A (en) * | 2019-12-31 | 2021-07-16 | 中国移动通信集团北京有限公司 | Information processing method, device, equipment and storage medium |
| CN113256036A (en) * | 2021-07-13 | 2021-08-13 | 国网浙江省电力有限公司 | Power supply cost analysis and prediction method based on Prophet-LSTNet combined model |
| CN113313529A (en) * | 2021-06-15 | 2021-08-27 | 大唐软控(青岛)科技有限公司 | Finished oil sales amount prediction method based on time regression sequence |
| CN113538026A (en) * | 2020-04-15 | 2021-10-22 | 北京京东振世信息技术有限公司 | Traffic calculation method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8892409B2 (en) * | 2009-02-11 | 2014-11-18 | Johnathan Mun | Project economics analysis tool |
| CN107766977A (en) * | 2017-10-20 | 2018-03-06 | 江苏电力信息技术有限公司 | A kind of Cash flow prospects method based on ARIMA |
| CN108985646A (en) * | 2018-07-27 | 2018-12-11 | 东莞嘉泰钟表有限公司 | A kind of tracking for financial regulation cash flow in producing |
-
2019
- 2019-04-10 CN CN201910282578.3A patent/CN109993370A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8892409B2 (en) * | 2009-02-11 | 2014-11-18 | Johnathan Mun | Project economics analysis tool |
| CN107766977A (en) * | 2017-10-20 | 2018-03-06 | 江苏电力信息技术有限公司 | A kind of Cash flow prospects method based on ARIMA |
| CN108985646A (en) * | 2018-07-27 | 2018-12-11 | 东莞嘉泰钟表有限公司 | A kind of tracking for financial regulation cash flow in producing |
Non-Patent Citations (2)
| Title |
|---|
| KENNETH S. LOREK ET AL.: "A Multivariate Time-Series Prediction Model for Cash-Flow Data", 《THE ACCOUNTING REVIEW》 * |
| 金翔: "日售电现金流预测模型的研究及应用", 《工业控制计算机》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113128693A (en) * | 2019-12-31 | 2021-07-16 | 中国移动通信集团北京有限公司 | Information processing method, device, equipment and storage medium |
| CN113538026A (en) * | 2020-04-15 | 2021-10-22 | 北京京东振世信息技术有限公司 | Traffic calculation method and device |
| CN113538026B (en) * | 2020-04-15 | 2023-11-03 | 北京京东振世信息技术有限公司 | Service amount calculation method and device |
| CN112580880A (en) * | 2020-12-23 | 2021-03-30 | 中国交通信息科技集团有限公司 | Capacity prediction method, computer processor and terminal equipment |
| CN112580880B (en) * | 2020-12-23 | 2023-12-19 | 中国交通信息科技集团有限公司 | Capacity prediction method, computer processor and terminal equipment |
| CN113313529A (en) * | 2021-06-15 | 2021-08-27 | 大唐软控(青岛)科技有限公司 | Finished oil sales amount prediction method based on time regression sequence |
| CN113256036A (en) * | 2021-07-13 | 2021-08-13 | 国网浙江省电力有限公司 | Power supply cost analysis and prediction method based on Prophet-LSTNet combined model |
| CN113256036B (en) * | 2021-07-13 | 2021-10-12 | 国网浙江省电力有限公司 | Power supply cost analysis and prediction method based on Prophet-LSTNet combined model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | How do tax incentives affect investment and productivity? Firm-level evidence from China | |
| Pless et al. | Pass-through as a test for market power: An application to solar subsidies | |
| CN109993370A (en) | A kind of electric power sale day cash flow projections method based on nonstationary time series | |
| Héricourt et al. | FDI and credit constraints: Firm-level evidence from China | |
| Kim et al. | Biases in static oligopoly models? Evidence from the California electricity market | |
| US20040215529A1 (en) | System and method for energy price forecasting automation | |
| Gersema et al. | An equilibrium pricing model for wind power futures | |
| Reichelstein et al. | Relating product prices to long‐run marginal cost: Evidence from solar photovoltaic modules | |
| Chow et al. | The influence of corporate income taxes on investment location: Evidence from corporate headquarters relocations | |
| US20030101062A1 (en) | Large scale procurement of energy efficiency resources | |
| CN109978230A (en) | A kind of electric power consumption sum intelligent Forecasting based on depth convolutional neural networks | |
| Datta et al. | Shareholder wealth creation following M&A: Evidence from European utility sectors | |
| KR100690088B1 (en) | Web-based demand side management evaluation system for analysing cost effect in the electric power industry | |
| Reichelstein et al. | Cost-and price dynamics of solar PV modules | |
| Lundin | Effects of privatization on price and labor efficiency: The Swedish electricity distribution sector | |
| Hosono et al. | Do financial frictions matter as a source of misallocation? Evidence from Japan | |
| US20140279314A1 (en) | System and Method for Reducing and Monitoring Energy Utility Bills | |
| Busch et al. | Estimation of avoided costs for electric utility demand-side planning | |
| Chakravarthy et al. | Are regulators effective at unraveling accounting manipulation? Evidence from public utility commissions | |
| Jung et al. | Capital Structure Under Imperfect Product Market Competition: Theory and Evidence | |
| Tanrisever et al. | Futures hedging in electricity retailing | |
| CN113269345A (en) | Weighted mean complex correlation analysis model-based cash flow prediction method | |
| CN112001552B (en) | Monthly electricity sales prediction method based on electricity consumption sequence | |
| Dindar et al. | Supply continuity in Turkish electricity distribution grid: electricity interruption cost forecasting with time series analysis and machine learning | |
| Sigrin et al. | Market and Modeling Uncertainty in Distributed Solar Deployment Projections: A Multi-Model Comparison |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190709 |