CN105005623A - Power demand prediction method based on keyword retrieval index correlation analysis - Google Patents

Power demand prediction method based on keyword retrieval index correlation analysis Download PDF

Info

Publication number
CN105005623A
CN105005623A CN201510445907.3A CN201510445907A CN105005623A CN 105005623 A CN105005623 A CN 105005623A CN 201510445907 A CN201510445907 A CN 201510445907A CN 105005623 A CN105005623 A CN 105005623A
Authority
CN
China
Prior art keywords
need
electricity
keyword
retrieval
growth
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
Application number
CN201510445907.3A
Other languages
Chinese (zh)
Inventor
吴金蔚
李璐
张凯锋
袁堃
范海虹
陈浩
田明
杨争林
耿建
薛必克
郑亚先
程海花
史新红
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southeast University
China Electric Power Research Institute Co Ltd CEPRI
State Grid Jiangsu Electric Power Co Ltd
Original Assignee
Southeast University
China Electric Power Research Institute Co Ltd CEPRI
State Grid Jiangsu Electric Power Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Southeast University, China Electric Power Research Institute Co Ltd CEPRI, State Grid Jiangsu Electric Power Co Ltd filed Critical Southeast University
Priority to CN201510445907.3A priority Critical patent/CN105005623A/en
Publication of CN105005623A publication Critical patent/CN105005623A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/95Retrieval from the web
    • G06F16/951Indexing; Web crawling techniques
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06QDATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/04Forecasting or optimisation, e.g. linear programming, "travelling salesman problem" or "cutting stock problem"
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06QDATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06Electricity, gas or water supply

Abstract

The invention discloses a power demand prediction method based on keyword retrieval index correlation analysis, and belongs to the technical field of the power demand prediction of a power system. The power demand prediction method comprises the following steps: firstly, utilizing an Internet search engine tool to obtain keyword retrieval indexes; then, utilizing historical power data to screen a keyword with high correlation of the retrieval indexes and a power growth rate; establishing a neutral network model of the keyword retrieval indexes and the power growth rate; and finally, establishing a correction model used for power demand prediction on the basis of a traditional prediction model. The power demand prediction method can analyze more power demand influence factors so as to improve the precision of power demands.

Description

Based on the need for electricity Forecasting Methodology that keyword retrieval exponential dependence is analyzed
Technical field
The invention belongs to electric system need for electricity electric powder prediction, is the need for electricity Forecasting Methodology introducing keyword retrieval index analysis.
Background technology
In need for electricity prediction, academia has carried out long-term and research fully.Traditional medium-term and long-term need for electricity Forecasting Methodology is more, and main theory is divided into two large classes: based on the regression prediction method of economic factors, based on seasonal effect in time series Forecasting Methodology.Theoretical based on this two class, in conjunction with various mathematical method, occurred abundant Forecasting Methodology, traditional algorithm has: elastic coefficient method, trend extrapolation, Regression Forecast, time series method etc.Modern algorithm has: gray theory, fuzzy prediction method, expert system, preferred compositions, artificial neural network and wavelet analysis method etc.
But current need for electricity prediction, limit by technological means, the factor of consideration is less, is mainly GDP, weather and historical load etc., does not fully take into account the property difference of each department electrical energy demands, the precision of prediction of heavy dependence input data itself.In fact, characterize the factor of quantity of electricity demand, except the correlative factor represented with structural data that Classical forecast is considered, also comprise a large amount of destructurings, semi-structured influence factor, as industry industry restructuring, new energy environment policy, great social event etc.
In recent years, along with the fast development of internet and social media, the propagation of dependent event and information and public response have become mark in a network and can follow.User, for some event or policy, when especially relevant with oneself interests event occurs, can have the motivation continuing to understand and explore.And internet search engine can record faithfully the focus of people.Therefore, based on search-engine tool, to the analysis of related keyword word and search situation, the situation of stakeholder can be understood to a certain extent.This relation for analysis relevant policies and need for electricity provides basic data supporting.
Summary of the invention
Technical matters to be solved by this invention is for traditional need for electricity Forecasting Methodology Consideration few, the not high deficiency of precision, proposes the need for electricity forecast value revision method based on correlation analysis between internet search engine keyword retrieval index and need for electricity data.
The technical solution adopted in the present invention is: for predicted area, analyzes the correlativity between its need for electricity rate of growth and search engine keywords retrieval index.Select the keyword having high correlation with need for electricity rate of growth, set up the neural network model of these keyword retrieval exponential sum need for electricity rate of growth, and traditional need for electricity forecast model is revised.Its step is as follows:
Steps A, based on the correlation analysis of internet search engine keyword retrieval exponential sum need for electricity rate of growth.
Steps A-1, selects the keyword of need for electricity influence factor, obtains Internal retrieval index.
Steps A-2, calculates the correlativity between predicted regional need for electricity rate of growth and each keyword retrieval index, finds out the keyword that correlativity is higher.
Step B, sets up the need for electricity forecast value revision model considering search engine keywords retrieval index.
Step B-1, the keyword retrieval index utilizing correlativity higher, sets up the backpropagation neural network model of keyword retrieval exponential sum need for electricity rate of growth.
Step B-2, draws need for electricity rate of growth according to above neural network model and keyword retrieval index, revises predicting the outcome of traditional need for electricity forecast model.
The invention has the beneficial effects as follows the keyword retrieval index considered in internet, more need for electricity influence factor can be analyzed, and improve the precision of need for electricity prediction.
Accompanying drawing explanation
Fig. 1 is system flow block diagram of the present invention;
Fig. 2 is the neural network model that the present invention sets up;
Embodiment
Below the need for electricity Forecasting Methodology based on the analysis of search engine keywords retrieval exponential dependence that the present invention proposes is described in detail:
Be illustrated in figure 1 FB(flow block) of the present invention:
Its specific implementation step is as follows:
Steps A, based on the correlation analysis of internet search engine keyword retrieval exponential sum need for electricity rate of growth.
Steps A-1, selects the keyword of need for electricity influence factor, obtains Internal retrieval index.
Consider the impact on need for electricity such as national energy policy, industry restructuring, select the keyword such as " energy-saving and emission-reduction ", " electricity price downward ", " closing high energy-consuming enterprises ", utilize existing search-engine tool " Baidu's index ", obtain the retrieval index of keyword in particular locality, special time period.Concrete operations are, open link " index.baidu.com ", input keyword, between selection area is timely, clicks " using Baidu.com ", can obtain the retrieval index of keyword.
Steps A-2, calculates the correlativity between predicted regional need for electricity rate of growth and each keyword retrieval index, finds out the keyword that correlativity is higher.
Utilize the correlativity between Calculation of correlation factor formula analysis power consumption rate of growth and search engine keywords retrieval index.
Correlativity between Two Variables can be measured by related coefficient.The related coefficient number-letter relation table r of variable x and variable y represents.In the present invention, x is the retrieval index of keyword, and y is the rate of growth of need for electricity.Computing formula is as follows:
In formula, r is the related coefficient of keyword retrieval exponential sum need for electricity rate of growth; Hop count when n is for analyzing; x ifor the retrieval index of a certain keyword in period i, for this keyword is at whole retrieval index analyzed in the period; y ifor the need for electricity rate of growth of predicted area in period i, for this area is in whole average need for electricity rate of growth analyzed in the period.
The span of correlation coefficient r is-1≤r≤1, is negative correlation during r<0, and be positive correlation during r>0, absolute value is larger, represents that the degree of correlation between two variablees is higher.As r=0, claim two variablees uncorrelated.
After calculating the related coefficient between each keyword and need for electricity rate of growth, filter out the higher N number of keyword of related coefficient (the keyword number N of selection is depending on particular problem and correlation coefficient value).
Step B, sets up the need for electricity forecast value revision model considering search engine keywords retrieval index.
Step B-1, the search engine keywords utilizing correlativity higher retrieval index, sets up the backpropagation neural network model of keyword retrieval exponential sum need for electricity rate of growth.
Step B-1-1, sets up neural network model.
The multilayer neural network model that the present invention sets up is divided into three layers: input layer, hidden layer, output layer, as shown in Figure 2.If N is input layer unit number, L is hidden layer unit number, M is output layer unit number.Input layer of the present invention is the retrieval index of keyword 1 to keyword N.Hidden layer unit number is greater than input layer 2 times, gets L>2N.Output layer unit one is the rate of growth of need for electricity.
Step B-1-2, carries out training study to neural network model.
Error backpropagation algorithm is utilized to train above neural network model.
Step B-2, draws need for electricity rate of growth according to above neural network model and keyword retrieval index, revises predicting the outcome of traditional need for electricity forecast model.
The N number of keyword retrieval index utilizing steps A to determine, input neural network model, show that the kth month need for electricity rate of growth same period is Δ nN(k).If the rate of growth that Classical forecast model draws is Δ 0(k), then revised need for electricity rate of growth is:
Δ(k)=Δ NN(k)+Δ 0(k)
Rate of growth Δ (k) is utilized to replace Δ 0k (), can revise Classical forecast model.

Claims (5)

1., based on the need for electricity Forecasting Methodology that search engine keywords retrieval exponential dependence is analyzed, it is characterized in that the method includes the steps of:
Steps A, based on the correlation analysis of internet search engine keyword retrieval exponential sum need for electricity rate of growth, comprising:
Steps A-1, selects the keyword of need for electricity influence factor, obtains Internal retrieval index;
Steps A-2, calculates the correlativity between predicted regional need for electricity rate of growth and each keyword retrieval index, finds out the keyword that correlativity is higher;
Step B, sets up the need for electricity forecast value revision model considering search engine keywords retrieval index, comprising:
Step B-1, the keyword retrieval index utilizing correlativity higher, sets up the backpropagation neural network model of keyword retrieval exponential sum need for electricity rate of growth;
Step B-2, draws need for electricity rate of growth according to above neural network model and keyword retrieval index, revises predicting the outcome of traditional need for electricity forecast model.
2., as claimed in claim 1 based on the need for electricity Forecasting Methodology that search engine keywords retrieval exponential dependence is analyzed, it is characterized in that steps A-1 concrete steps are as follows:
Utilize existing search-engine tool " Baidu's index ", obtain the retrieval index of keyword in particular locality, special time period, described keyword comprises " energy-saving and emission-reduction ", " electricity price downward ", " closing high energy-consuming enterprises ".
3., as claimed in claim 1 based on the need for electricity Forecasting Methodology that search engine keywords retrieval exponential dependence is analyzed, it is characterized in that steps A-2 concrete steps are as follows:
Utilize the correlativity between formula of correlation coefficient calculating power consumption rate of growth and search engine keywords retrieval index, filter out the keyword that related coefficient is higher.
4. as claimed in claim 1 based on the need for electricity Forecasting Methodology that search engine keywords retrieval exponential dependence is analyzed, it is characterized in that: the multilayer neural network model that step B-1 sets up, its input layer is the retrieval index of each keyword that correlativity is higher, and output layer is the rate of growth of need for electricity.
5., as claimed in claim 1 based on the need for electricity Forecasting Methodology that search engine keywords retrieval exponential dependence is analyzed, it is characterized in that step B-2 concrete steps are as follows:
The keyword retrieval index that the correlativity utilizing steps A to determine is higher, as the input of neural network model, draws need for electricity rate of growth, the rate of growth that Classical forecast model draws is added with it, is the need for electricity rate of growth revising forecast model.
CN201510445907.3A 2015-07-27 2015-07-27 Power demand prediction method based on keyword retrieval index correlation analysis Pending CN105005623A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510445907.3A CN105005623A (en) 2015-07-27 2015-07-27 Power demand prediction method based on keyword retrieval index correlation analysis

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510445907.3A CN105005623A (en) 2015-07-27 2015-07-27 Power demand prediction method based on keyword retrieval index correlation analysis

Publications (1)

Publication Number Publication Date
CN105005623A true CN105005623A (en) 2015-10-28

Family

ID=54378299

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510445907.3A Pending CN105005623A (en) 2015-07-27 2015-07-27 Power demand prediction method based on keyword retrieval index correlation analysis

Country Status (1)

Country Link
CN (1) CN105005623A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105488161A (en) * 2015-11-30 2016-04-13 百度在线网络技术(北京)有限公司 Information pushing method and apparatus
CN108256765A (en) * 2018-01-16 2018-07-06 前海梧桐(深圳)数据有限公司 The computational methods and its system of basic factors interactively between different enterprises
CN108319700A (en) * 2018-02-05 2018-07-24 广东广垦畜牧工程研究院有限公司 A kind of swine plague prediction technique based on web search index

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101625683A (en) * 2008-07-09 2010-01-13 精实万维软件(北京)有限公司 Method for selecting bidding advertisement keyword during release of search engine bidding advertisement
CN101645155A (en) * 2008-08-08 2010-02-10 陈列生 Network marketing method
CN104035994A (en) * 2014-06-11 2014-09-10 华东师范大学 Prediction method of television play on-demand amount based on network data

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101625683A (en) * 2008-07-09 2010-01-13 精实万维软件(北京)有限公司 Method for selecting bidding advertisement keyword during release of search engine bidding advertisement
CN101645155A (en) * 2008-08-08 2010-02-10 陈列生 Network marketing method
CN104035994A (en) * 2014-06-11 2014-09-10 华东师范大学 Prediction method of television play on-demand amount based on network data

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
刘佳霖: "引入互联网搜索量的市场需求预测模型研究", 《中国优秀硕士学位论文全文数据库》 *
王守中等: "基于Web搜索数据的宝马汽车销量预测研究", 《经济师》 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105488161A (en) * 2015-11-30 2016-04-13 百度在线网络技术(北京)有限公司 Information pushing method and apparatus
CN108256765A (en) * 2018-01-16 2018-07-06 前海梧桐(深圳)数据有限公司 The computational methods and its system of basic factors interactively between different enterprises
CN108319700A (en) * 2018-02-05 2018-07-24 广东广垦畜牧工程研究院有限公司 A kind of swine plague prediction technique based on web search index

Similar Documents

Publication Publication Date Title
Wu et al. How does energy consumption affect China's urbanization? New evidence from dynamic threshold panel models
Lin et al. Energy demand in China: Comparison of characteristics between the US and China in rapid urbanization stage
Wang et al. A novel hybridization of nonlinear grey model and linear ARIMA residual correction for forecasting US shale oil production
He et al. Forecasting China's total energy demand and its structure using ADL-MIDAS model
Sözen et al. Prospects for future projections of the basic energy sources in Turkey
Wang et al. The impact and economic cost of environmental regulation on energy utilization in China
CN104537433A (en) Sold electricity quantity prediction method based on inventory capacities and business expansion characteristics
CN105446139A (en) Construction energy consumption analysis method and construction energy consumption analysis system based on BP neural network
CN105005623A (en) Power demand prediction method based on keyword retrieval index correlation analysis
CN104574209A (en) Modeling method of urban electrical network distribution transform weight overload mid-term forewarning model
Pan et al. Grey system theory trends from 1991 to 2018: a bibliometric analysis and visualization
Liu et al. The damping accumulated grey model and its application
CN104050547A (en) Non-linear optimization decision-making method of planning schemes for oilfield development
Ye et al. A new environmental governance cost prediction method based on indicator synthesis and different risk coefficients
Duan et al. A multivariate grey prediction model based on energy logistic equation and its application in energy prediction in China
Wang et al. Multi-criteria comprehensive study on predictive algorithm of heating energy consumption of district heating station based on timeseries processing
Ma et al. An intelligent decision support system for residential energy consumption and renewable energy utilization in rural China
CN104699890A (en) Modeling method of short-term forewarning model for heavy overload of distribution transformer of urban power network
ŞİŞMAN A comparison of ARIMA and Grey models for Electricity consumption demand forecasting: The case of turkey
Liu et al. Optimizing layout of pumping well in irrigation district for groundwater sustainable use in northwest China
Vakulenko et al. A systematic literature review of smart grids
Song et al. Sustainable development of China's regions from the perspective of ecological welfare performance: analysis based on GM (1, 1) and the malmquist index
CN107730025A (en) A kind of power supply enterprise&#39;s Cash flow prospects model changed based on electricity
CN106529742A (en) Method for predicting typical daily load of power grid by utilizing method for searching daily load rates in database
CN112633762A (en) Building energy efficiency obtaining method and equipment

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20151028