CN108519989A - The reduction retroactive method and device of a kind of day electricity missing data - Google Patents
The reduction retroactive method and device of a kind of day electricity missing data Download PDFInfo
- Publication number
- CN108519989A CN108519989A CN201810163770.6A CN201810163770A CN108519989A CN 108519989 A CN108519989 A CN 108519989A CN 201810163770 A CN201810163770 A CN 201810163770A CN 108519989 A CN108519989 A CN 108519989A
- Authority
- CN
- China
- Prior art keywords
- data
- electricity
- day
- day electricity
- value
- 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
- 230000005611 electricity Effects 0.000 title claims abstract description 160
- 238000000034 method Methods 0.000 title claims abstract description 47
- 230000009467 reduction Effects 0.000 title claims abstract description 45
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 claims abstract description 22
- 238000013480 data collection Methods 0.000 claims abstract description 13
- 230000008569 process Effects 0.000 claims description 19
- 238000012417 linear regression Methods 0.000 claims description 12
- 238000010276 construction Methods 0.000 claims description 6
- 238000002372 labelling Methods 0.000 claims description 6
- 238000007477 logistic regression Methods 0.000 claims description 6
- 238000012795 verification Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 230000000295 complement effect Effects 0.000 claims 1
- 238000011160 research Methods 0.000 abstract description 7
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000009471 action Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000007812 deficiency Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The reduction retroactive method of a kind of day electricity missing data, includes the following steps:Obtain user information and its electric energy meter information and moon electricity data;The day electricity data of each user is obtained according to day electricity rule;Day electricity monthly is checked, the day electricity for needing restoring data per month is found out;Restored respectively using two kinds of multiple interpolation models based on algorithms of different needs the day electricity of restoring data per month, obtains restoring data collection;It is compared using the moon electricity moon electricity corresponding with restoring data collection for having obtained user, the restoring data of deviation ratio minimum is taken to integrate as optimal result collection.The present invention has the following advantages:The present invention improves the quality of data of day electricity, the true truth for reflecting the daily electricity consumption of user provides true and reliable data basis for the relevant research of user power utilization behavior by being restored to day electricity missing values;The present invention carries out data convert using two kinds of multiple interpolation models based on algorithms of different, improves the accuracy of data convert.
Description
Technical field
The present invention relates to power marketing electric quantity monitoring analysis technical fields, and in particular to a kind of day electricity missing data also
Former retroactive method and device.
Background technology
With the development of national grid intelligent information technology in recent years and the raising of precision service level, electricity consumption
Data especially day electricity data also increasingly improves the importance of electric power enterprise and other industries, and Accurate Analysis industry is used
Electric feature, estimation trade power consumption trend, development scale and regional economic development trend, instruct electric power enterprise to formulate corresponding war
Slightly plan and promote customer service quality etc. is all played the role of vital.
Day electricity data is mainly derived from power information acquisition system, passes through (same day electric energy meter indicating value-in intelligent electric energy meter
Day before yesterday electric energy meter indicating value) acquisition of * current/voltage multiplying powers, but due to power network fluctuation, electric energy meter replacement, equipment fault etc., electricity
Energy expression value often cannot be obtained normally, lead to problems such as a day electricity missing, the confidence level of data relatively low.Currently, electric power is looked forward to
The analysis that industry is generally basede on electricity still uses moon electricity data (manual meter reading, monthly segmentation interception) or interception part-time section
Day electricity, simple delete or the method for mean value is used as the data of correlation statistical analysis so that data statistic analysis
As a result there is relatively large deviation, the confidence level of result of study substantially reduces.It is retrieved by technology and finds that the country is lacked about day electricity
The research of data convert retrospect is less, is to carry out seasonal or tendency prediction from macroscopic perspective mostly, does not provide
Specific solution.
Invention content
Present invention aims to overcome that the shortcomings that prior art and deficiency, provide the reduction of a kind of day electricity missing data
Retroactive method improves the quality of data of day electricity, true reflection user is per daily by being restored to day electricity missing values
The truth of electricity provides true and reliable data basis for the relevant research of user power utilization behavior.
The present invention also provides the reduction retrospective devices of a kind of day electricity missing data.
To achieve the above object, the technical solution adopted by the present invention is as follows:
The reduction retroactive method of a kind of day electricity missing data, includes the following steps:
Obtain user information and its electric energy meter information and moon electricity data;
The day electricity data of each user is obtained according to day electricity rule;
Day electricity monthly is checked, the day electricity for needing restoring data per month is found out;
Restored respectively using two kinds of multiple interpolation models based on algorithms of different needs the day electricity of restoring data per month, obtains
To restoring data collection;
It is compared using the moon electricity moon electricity corresponding with restoring data collection for having obtained user, takes deviation ratio minimum
Restoring data integrates as optimal result collection.
The action principle of the present invention:User information and its electric energy meter information are obtained in advance, and calculate day by preset rules
Data are measured, and day electricity abnormal data is handled, the data area that need to restore retrospect are obtained, to the day electricity data
Measure is restored using two kinds of missing values, and data set is obtained to two kinds of measures and is compared, to obtain optimal selection.The present invention
By being restored to day electricity missing values, the quality of data of day electricity, the true true feelings for reflecting the daily electricity consumption of user are improved
Condition provides true and reliable data basis for the relevant research of user power utilization behavior;The present invention is based on difference using two kinds simultaneously
The multiple interpolation model of algorithm carries out data convert, improves the accuracy of data convert.
Further, the user information includes subscriber-coded and user's name.
Further, the electric energy meter information include electric energy meter indicating value, current transformer multiplying power, voltage transformer multiplying power and
Electric energy meter acquires the date.
Further, it is described by day electricity rule obtain day of each user and be in the way of electricity data:Day electricity=
(next day electric energy meter indicating value-same day electric energy meter indicating value) × current transformer multiple × voltage transformer multiple.
As an improvement of the present invention, day electricity monthly is checked, finds out the day for needing restoring data per month
Electricity, including following sub-step:
It is vacancy and the data less than 0 to find out day electricity, and is identified as the day electricity for needing restoring data, and its value is respectively provided with
For vacancy;
Remaining data to monthly removing AFR control carries out quartile method classification, and the maximum for acquiring remaining data is critical
Value;
It will be greater than being identified as equal to the day electricity of maximum critical value and need restoring data, and its value is disposed as vacancy.
Preferably, described two multiple interpolation models based on algorithms of different are that Bayes's linear prediction regression algorithm is multiple
Interpolation model and propensity score fado weight interpolation model.
Further, the multiple interpolation model reduction day electricity data process of Bayes's linear prediction regression algorithm includes
Following sub-step:
It sets multiple interpolation and generates the quantity of data set as K;
Bayes's linear regression algorithm generates the interpolation data of K data set, and interpolation data is updated to multiple interpolation
In model, K are obtained without missing data set;
By generalized linear model into the verification of line data set and its linear regression is obtained without missing data set to K successively
Equation judges whether its data set is qualified, when the p value of generalized linear model>When 0.05, the data set is qualified;Otherwise the data
Collect unqualified, re-executes the multiple interpolation model reduction day electricity data process of this linear prediction regression algorithm of leaf.
Further, the propensity score fado weight interpolation model reduction day electricity data process includes following sub-step:
Firstly generate variable Ri, work as RiWhen=0, variable x is representedm;Work as RiWhen=1, variable x is representeda, wherein xmTo contain
There are the variable of missing values, xaFor complete variable;
Then construction logic regression equation:Logit(Pi)=β0+β1x1+β2x2+......+βi-1xi-1
Wherein, x1, x2, x3......xi-1For xmCovariant,
Pi=PT(Ri=0 | x1, x2, x3......xi-1), logit (P)=log (P/ (1-P));
Trend score, wherein Logit (P are generated to missing data using logistic regression equationi) it is each missing number
According to opposite reserved portion;
On the basis of trend score, a pair reference group number corresponding with observation group is grouped, and is divided into K groups;
The missing data in K group data sets is filled up using Bootstrap methods, in group m, is usedRepresenting should
Containing n0 without missing X in group miObserved value number,It represents and contains ni X in this group of miMissing values number, then fromIn randomly select n numerical value, the n value randomly selected as Filling power be used for pairIn missing values be filled;
Above step is computed repeatedly, value is filled up until each value in xm obtains one.
A kind of reduction retrospective device of day electricity missing data, including:
Data acquisition module, for obtaining user information and its electric energy meter information and moon electricity data;
Day electricity computing module, for obtaining the day electricity data of each user according to day electricity rule;
Data processing module for checking day electricity monthly, and finds out the day for needing restoring data per month
Amount;
Data restoring module, restored respectively using two kinds of multiple interpolation models based on algorithms of different needs reduction number per month
According to day electricity, obtain restoring data collection;
Comparing module is compared using the moon electricity moon electricity corresponding with restoring data collection for having obtained user, is taken partially
The restoring data of rate minimum integrates as optimal result collection.
Further, the user information includes subscriber-coded and user's name.
Further, the electric energy meter information include electric energy meter indicating value, current transformer multiplying power, voltage transformer multiplying power and
Electric energy meter acquires the date.
Further, it is described by day electricity rule obtain day of each user and be in the way of electricity data:Day electricity=
(next day electric energy meter indicating value-same day electric energy meter indicating value) × current transformer multiple × voltage transformer multiple.
As an improvement of the present invention, the data processing module includes following submodule:
AFR control labeling submodule needs reduction number for finding out data of day electricity for vacancy and less than 0, and being identified as
According to day electricity, and its value is disposed as vacancy;
Maximum critical value seeks submodule, returns for carrying out quartile method to the remaining data for monthly removing AFR control
Class, and acquire the maximum critical value of remaining data;
AFR control labeling submodule again, the day electricity for will be greater than equal to maximum critical value, which is identified as, needs reduction number
According to, and its value is disposed as vacancy.
Preferably, described two multiple interpolation models based on algorithms of different are that Bayes's linear prediction regression algorithm is multiple
Interpolation model and propensity score fado weight interpolation model;
The multiple interpolation model reduction day electricity data process of Bayes's linear prediction regression algorithm is as follows:
It sets multiple interpolation and generates the quantity of data set as K;
Bayes's linear regression algorithm generates the interpolation data of K data set, and interpolation data is updated to multiple interpolation
In model, K are obtained without missing data set;
By generalized linear model into the verification of line data set and its linear regression is obtained without missing data set to K successively
Equation judges whether its data set is qualified, when the p value of generalized linear model>When 0.05, the data set is qualified;Otherwise the data
Collect unqualified, re-executes the multiple interpolation model reduction day electricity data process of this linear prediction regression algorithm of leaf.
The propensity score fado weight interpolation model reduction day electricity data process includes as follows:
Firstly generate variable Ri, work as RiWhen=0, variable x is representedm;Work as RiWhen=1, variable x is representeda, wherein xmTo contain
There are the variable of missing values, xaFor complete variable;
Then construction logic regression equation:Logit(Pi)=β0+β1x1+β2x2+......+βi-1xi-1
Wherein, x1, x2, x3......xi-1For xmCovariant,
Pi=PT(Ri=0 | x1, x2, x3......xi-1), logit (P)=log (P/ (1-P));
Trend score, wherein Logit (P are generated to missing data using logistic regression equationi) it is each missing number
According to opposite reserved portion;
On the basis of trend score, a pair reference group number corresponding with observation group is grouped, and is divided into K groups;
The missing data in K group data sets is filled up using Bootstrap methods, in group m, is usedRepresenting should
Containing n0 without missing X in group miObserved value number,It represents and contains ni X in this group of miMissing values number, then fromIn randomly select n numerical value, the n value randomly selected as Filling power be used for pairIn missing values be filled;
Above step is computed repeatedly, until xmIn each value obtain one fill up value.
Compared with prior art, the present invention has the following advantages:
1, the present invention improves the quality of data of day electricity, really reflects user by being restored to day electricity missing values
The truth of daily electricity consumption provides true and reliable data basis for the relevant research of user power utilization behavior.
2, the present invention carries out data convert using two kinds of multiple interpolation models based on algorithms of different, improves data convert
Accuracy.
Description of the drawings
Fig. 1 is the flow chart of the reduction retroactive method of day electricity missing data of the invention.
Specific implementation mode
The invention will be further described with reference to the accompanying drawings and examples.It is understood that tool described herein
Body embodiment is used only for explaining the present invention rather than limitation of the invention.It also should be noted that for the ease of retouching
It states, only some but not all contents related to the present invention are shown in the drawings.
Embodiment
The present invention restores day electricity missing values using multiple interpolation side, improves the quality of data of day electricity, really
The truth for reflecting the daily electricity consumption of user provides true and reliable data basis for the relevant research of user power utilization behavior, under
Face is described the detailed process of the present invention using specific embodiment.
Referring to FIG. 1, the reduction retroactive method of a kind of day electricity missing data, includes the following steps:
S1. obtain user information and its electric energy meter information and moon electricity data, the user information include it is subscriber-coded and
User's name, the electric energy meter information include that electric energy meter indicating value, current transformer multiplying power, voltage transformer multiplying power and electric energy meter are adopted
Collect the date.
S2. the day electricity data of each user is obtained according to day electricity rule, specifically, the day electricity of each user indicates
Formula is:Day electricity=(next day electric energy meter indicating value-same day electric energy meter indicating value) × current transformer multiple × voltage transformer times
Number.
S3. day electricity monthly is checked, finds out the day electricity for needing restoring data per month;Include specifically following
Sub-step:
It is vacancy and the data less than 0 to find out day electricity, and is identified as the day electricity for needing restoring data, and its value is respectively provided with
For vacancy;
Remaining data to monthly removing AFR control carries out quartile method classification, and the maximum for acquiring remaining data is critical
Value;
It will be greater than being identified as equal to the day electricity of maximum critical value and need restoring data, and its value is disposed as vacancy.
S4. two kinds of multiple interpolation models based on algorithms of different are used to restore the day for needing restoring data per month respectively
Amount, obtains restoring data collection;Wherein described two multiple interpolation models based on algorithms of different return for Bayes's linear prediction
The multiple interpolation model of algorithm and propensity score fado weight interpolation model;
The multiple interpolation model reduction day electricity data process of Bayes's linear prediction regression algorithm includes following sub-step
Suddenly:
It sets multiple interpolation and generates the quantity of data set as K;
Bayes's linear regression algorithm generates the interpolation data of K data set, and interpolation data is updated to multiple interpolation
In model, K are obtained without missing data set;
By generalized linear model into the verification of line data set and its linear regression is obtained without missing data set to K successively
Equation judges whether its data set is qualified, when the p value of generalized linear model>When 0.05, the data set is qualified;Otherwise the data
Collect unqualified, re-executes the multiple interpolation model reduction day electricity data process of this linear prediction regression algorithm of leaf;
The propensity score fado weight interpolation model reduction day electricity data process includes following sub-step:
Firstly generate variable Ri, work as RiWhen=0, variable x is representedm;Work as RiWhen=1, variable x is representeda, wherein xmTo contain
There are the variable of missing values, xaFor complete variable;
Then construction logic regression equation:Logit(Pi)=β0+β1x1+β2x2+......+βi-1xi-1
Wherein, x1, x2, x3......xi-1For xmCovariant,
Pi=PT(Ri=0 | x1, x2, x3......xi-1), logit (P)=log (P/ (1-P));
Trend score, wherein Logit (P are generated to missing data using logistic regression equationi) it is each missing number
According to opposite reserved portion;
On the basis of trend score, a pair reference group number corresponding with observation group is grouped, and is divided into K groups;
The missing data in K group data sets is filled up using Bootstrap methods, in m groups, 1≤m≤K is usedIt represents in the m groups containing n0 without missing XiObserved value number,It represents and contains ni X in the m groupsiLack
Mistake value number, then fromIn randomly select n numerical value, the n value randomly selected as Filling power be used for pairIn lack
Mistake value is filled;
Above step is computed repeatedly, until xmIn each value obtain one fill up value.
The present invention carries out data convert using two kinds of multiple interpolation models based on algorithms of different, improves data convert
Accuracy.
S5. it is compared using the moon electricity moon electricity corresponding with restoring data collection for having obtained user, takes deviation ratio most
Small restoring data integrates as optimal result collection.
The action principle of the present invention:User information and its electric energy meter information are obtained in advance, and calculate day by preset rules
Data are measured, and day electricity abnormal data is handled, the data area that need to restore retrospect are obtained, to the day electricity data
Measure is restored using two kinds of missing values, and data set is obtained to two kinds of measures and is compared, to obtain optimal selection.The present invention
By being restored to day electricity missing values, the quality of data of day electricity, the true true feelings for reflecting the daily electricity consumption of user are improved
Condition provides true and reliable data basis for the relevant research of user power utilization behavior;The present invention is based on difference using two kinds simultaneously
The multiple interpolation model of algorithm carries out data convert, improves the accuracy of data convert.
A kind of reduction retrospective device of day electricity missing data, including:
Data acquisition module, for obtaining user information and its electric energy meter information and moon electricity data, the user information
Including subscriber-coded and user's name, the electric energy meter information includes electric energy meter indicating value, current transformer multiplying power, voltage transformer
Multiplying power and electric energy meter acquire the date.
Day electricity computing module, for obtaining the day electricity data of each user, wherein day electricity according to day electricity rule
=(next day electric energy meter indicating value-same day electric energy meter indicating value) × current transformer multiple × voltage transformer multiple.
Data processing module for checking day electricity monthly, and finds out the day for needing restoring data per month
Amount;Specifically include following submodule:
AFR control labeling submodule needs reduction number for finding out data of day electricity for vacancy and less than 0, and being identified as
According to day electricity, and its value is disposed as vacancy;
Maximum critical value seeks submodule, returns for carrying out quartile method to the remaining data for monthly removing AFR control
Class, and acquire the maximum critical value of remaining data;
AFR control labeling submodule again, the day electricity for will be greater than equal to maximum critical value, which is identified as, needs reduction number
According to, and its value is disposed as vacancy.
Data restoring module, restored respectively using two kinds of multiple interpolation models based on algorithms of different needs reduction number per month
According to day electricity, obtain restoring data collection;Described two multiple interpolation models based on algorithms of different are Bayes's linear prediction
The multiple interpolation model of regression algorithm and propensity score fado weight interpolation model;
The multiple interpolation model reduction day electricity data process of Bayes's linear prediction regression algorithm is as follows:
It sets multiple interpolation and generates the quantity of data set as K;
Bayes's linear regression algorithm generates the interpolation data of K data set, and interpolation data is updated to multiple interpolation
In model, K are obtained without missing data set;
By generalized linear model into the verification of line data set and its linear regression is obtained without missing data set to K successively
Equation judges whether its data set is qualified, when the p value of generalized linear model>When 0.05, the data set is qualified;Otherwise the data
Collect unqualified, re-executes the multiple interpolation model reduction day electricity data process of this linear prediction regression algorithm of leaf;
The propensity score fado weight interpolation model reduction day electricity data process includes following sub-step:
Firstly generate variable Ri, work as RiWhen=0, variable x is representedm;Work as RiWhen=1, variable x is representeda, wherein xmTo contain
There are the variable of missing values, xaFor complete variable;
Then construction logic regression equation:Logit(Pi)=β0+β1x1+β2x2+......+βi-1xi-1
Wherein, x1, x2, x3......xi-1For xmCovariant,
Pi=PT(Ri=0 | x1, x2, x3......xi-1), logit (P)=log (P/ (1-P));
Trend score, wherein Logit (P are generated to missing data using logistic regression equationi) it is each missing number
According to opposite reserved portion;
On the basis of trend score, a pair reference group number corresponding with observation group is grouped, and is divided into K groups;
The missing data in K group data sets is filled up using Bootstrap methods, in m groups, 1≤m≤K is usedIt represents in the m groups containing n0 without missing XiObserved value number,It represents and contains ni X in the m groupsiLack
Mistake value number, then fromIn randomly select n numerical value, the n value randomly selected as Filling power be used for pairIn lack
Mistake value is filled;
Above step is computed repeatedly, until xmIn each value obtain one fill up value.
Comparing module is compared using the moon electricity moon electricity corresponding with restoring data collection for having obtained user, is taken partially
The restoring data of rate minimum integrates as optimal result collection.
The above embodiment is a preferred embodiment of the present invention, but embodiments of the present invention are not by above-described embodiment
Limitation, it is other it is any without departing from the spirit and principles of the present invention made by changes, modifications, substitutions, combinations, simplifications,
Equivalent substitute mode is should be, is included within the scope of the present invention.
Claims (10)
1. the reduction retroactive method of a kind of day electricity missing data, it is characterised in that include the following steps:
Obtain user information and its electric energy meter information and moon electricity data;
The day electricity data of each user is obtained according to day electricity rule;
Day electricity monthly is checked, the day electricity for needing restoring data per month is found out;
Restored respectively using two kinds of multiple interpolation models based on algorithms of different needs the day electricity of restoring data per month, is gone back
Original data set;
It is compared using the moon electricity moon electricity corresponding with restoring data collection for having obtained user, takes the reduction of deviation ratio minimum
Data set is optimal result collection.
2. the reduction retroactive method of according to claim 1 day electricity missing data, it is characterised in that:The electric energy meter letter
Breath includes electric energy meter indicating value, current transformer multiplying power, voltage transformer multiplying power and electric energy meter acquisition date.
3. the reduction retroactive method of according to claim 2 day electricity missing data, it is characterised in that:It is described according to day
Gauge then obtain each user day electricity data mode be:Day electricity=(next day electric energy meter indicating value-same day electric energy indicates
Value) × current transformer multiple × voltage transformer multiple.
4. the reduction retroactive method of according to claim 1 day electricity missing data, it is characterised in that:To day monthly
Amount is checked, the day electricity for needing restoring data per month, including following sub-step are found out:
It is vacancy and the data less than 0 to find out day electricity, and is identified as the day electricity for needing restoring data, and its value is disposed as sky
It lacks;
Remaining data to monthly removing AFR control carries out quartile method classification, and acquires the maximum critical value of remaining data;
It will be greater than being identified as equal to the day electricity of maximum critical value and need restoring data, and its value is disposed as vacancy.
5. the reduction retroactive method of according to claim 1 day electricity missing data, it is characterised in that:It is described two to be based on
The multiple interpolation model of algorithms of different is that the multiple interpolation model of Bayes's linear prediction regression algorithm and propensity score method are multiple slotting
Complementary modulus type;
The multiple interpolation model reduction day electricity data process of Bayes's linear prediction regression algorithm includes following sub-step:
It sets multiple interpolation and generates the quantity of data set as K;
Bayes's linear regression algorithm generates the interpolation data of K data set, and interpolation data is updated to multiple interpolation model
In, K are obtained without missing data set;
By generalized linear model into the verification of line data set and its linear regression side is obtained without missing data set to K successively
Journey judges whether its data set is qualified, when the p value of generalized linear model>When 0.05, the data set is qualified;Otherwise the data set
It is unqualified, re-execute the multiple interpolation model reduction day electricity data process of this linear prediction regression algorithm of leaf;
The propensity score fado weight interpolation model reduction day electricity data process includes following sub-step:
Firstly generate variable Ri, work as RiWhen=0, variable x is representedm;Work as RiWhen=1, variable x is representeda, wherein xmIt is scarce to contain
The variable of mistake value, xaFor complete variable;
Then construction logic regression equation:Logit(Pi)=β0+β1x1+β2x2+......+βi-1xi-1
Wherein, x1, x2, x3......xi-1For xmCovariant,
Pi=PT(Ri=0 | x1, x2, x3......xi-1), logit (P)=log (P/ (1-P));
Trend score, wherein Logit (P are generated to missing data using logistic regression equationi) be each missing data phase
To reserved portion;
On the basis of trend score, a pair reference group number corresponding with observation group is grouped, and is divided into K groups;
The missing data in K group data sets is filled up using Bootstrap methods, in group m, is usedRepresent this group of m
In containing n0 without missing XiObserved value number,It represents and contains ni X in this group of miMissing values number, then from
In randomly select n numerical value, the n value randomly selected as Filling power be used for pairIn missing values be filled;
Above step is computed repeatedly, until xmIn each value obtain one fill up value.
6. the reduction retrospective device of a kind of day electricity missing data, it is characterised in that including:
Data acquisition module, for obtaining user information and its electric energy meter information and moon electricity data;
Day electricity computing module, for obtaining the day electricity data of each user according to day electricity rule;
Data processing module for checking day electricity monthly, and finds out the day electricity for needing restoring data per month;
Data restoring module, restored respectively using two kinds of multiple interpolation models based on algorithms of different needs restoring data per month
Day electricity, obtains restoring data collection;
Comparing module is compared using the moon electricity moon electricity corresponding with restoring data collection for having obtained user, takes deviation ratio
Minimum restoring data integrates as optimal result collection.
7. the reduction retrospective device of according to claim 6 day electricity missing data, it is characterised in that:The electric energy meter letter
Breath includes electric energy meter indicating value, current transformer multiplying power, voltage transformer multiplying power and electric energy meter acquisition date.
8. the reduction retrospective device of according to claim 7 day electricity missing data, it is characterised in that:It is described according to day
Gauge then obtain each user day electricity data mode be:Day electricity=(next day electric energy meter indicating value-same day electric energy indicates
Value) × current transformer multiple × voltage transformer multiple.
9. the reduction retrospective device of according to claim 6 day electricity missing data, it is characterised in that:The data processing
Module includes following submodule:
AFR control labeling submodule needs restoring data for finding out data of day electricity for vacancy and less than 0, and being identified as
Day electricity, and its value is disposed as vacancy;
Maximum critical value seeks submodule, for carrying out quartile method classification to the remaining data for monthly removing AFR control, and
Acquire the maximum critical value of remaining data;
AFR control labeling submodule again, the day electricity for will be greater than equal to maximum critical value, which is identified as, needs restoring data, and
Its value is disposed as vacancy.
10. the reduction retrospective device of according to claim 6 day electricity missing data, it is characterised in that:Described two bases
It is multiple for the multiple interpolation model of Bayes's linear prediction regression algorithm and propensity score method in the multiple interpolation model of algorithms of different
Interpolation model;
The multiple interpolation model reduction day electricity data process of Bayes's linear prediction regression algorithm is as follows:
It sets multiple interpolation and generates the quantity of data set as K;
Bayes's linear regression algorithm generates the interpolation data of K data set, and interpolation data is updated to multiple interpolation model
In, K are obtained without missing data set;
By generalized linear model into the verification of line data set and its linear regression side is obtained without missing data set to K successively
Journey judges whether its data set is qualified, when the p value of generalized linear model>When 0.05, the data set is qualified;Otherwise the data set
It is unqualified, re-execute the multiple interpolation model reduction day electricity data process of this linear prediction regression algorithm of leaf;
The propensity score fado weight interpolation model reduction day electricity data process includes as follows:
Firstly generate variable Ri, work as RiWhen=0, variable x is representedm;Work as RiWhen=1, variable x is representeda, wherein xmIt is scarce to contain
The variable of mistake value, xaFor complete variable;
Then construction logic regression equation:Logit(Pi)=β0+β1x1+β2x2+......+βi-1xi-1
Wherein, x1, x2, x3......xi-1For xmCovariant,
Pi=PT(Ri=0 | x1, x2, x3......xi-1), logit (P)=log (P/ (1-P));
Trend score, wherein Logit (P are generated to missing data using logistic regression equationi) be each missing data phase
To reserved portion;
On the basis of trend score, a pair reference group number corresponding with observation group is grouped, and is divided into K groups;
The missing data in K group data sets is filled up using Bootstrap methods, in group m, is usedRepresent this group of m
In containing n0 without missing XiObserved value number,It represents and contains ni X in this group of miMissing values number, then from
In randomly select n numerical value, the n value randomly selected as Filling power be used for pairIn missing values be filled;
Above step is computed repeatedly, until xmIn each value obtain one fill up value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810163770.6A CN108519989A (en) | 2018-02-27 | 2018-02-27 | The reduction retroactive method and device of a kind of day electricity missing data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810163770.6A CN108519989A (en) | 2018-02-27 | 2018-02-27 | The reduction retroactive method and device of a kind of day electricity missing data |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108519989A true CN108519989A (en) | 2018-09-11 |
Family
ID=63433330
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810163770.6A Pending CN108519989A (en) | 2018-02-27 | 2018-02-27 | The reduction retroactive method and device of a kind of day electricity missing data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108519989A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109445972A (en) * | 2018-09-21 | 2019-03-08 | 深圳供电局有限公司 | Data recovery method, device, equipment and storage medium |
CN111125635A (en) * | 2019-11-29 | 2020-05-08 | 广东电网有限责任公司 | User table data filling method, device, terminal and storage medium |
CN111127247A (en) * | 2019-11-21 | 2020-05-08 | 深圳供电局有限公司 | Electric quantity data acquisition method and device, computer equipment and storage medium |
CN111178788A (en) * | 2020-01-09 | 2020-05-19 | 国网福建省电力有限公司 | Power data auxiliary management method and system for power selling side |
CN111667123A (en) * | 2020-06-17 | 2020-09-15 | 上海积成能源科技有限公司 | Method for supplementing missing value by applying multiple interpolation in power load prediction |
CN116415123A (en) * | 2023-03-07 | 2023-07-11 | 清华大学 | Method and system for analyzing total water flow data of community |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102521080A (en) * | 2011-12-07 | 2012-06-27 | 福建省电力有限公司 | Computer data recovery method for electricity-consumption information collecting system for power consumers |
US20130036082A1 (en) * | 2011-08-05 | 2013-02-07 | International Business Machines Corporation | Multiple imputation of missing data in multi-dimensional retail sales data sets via tensor factorization |
CN103353895A (en) * | 2013-07-22 | 2013-10-16 | 广西电网公司电力科学研究院 | Pre-processing method of power distribution network line loss data |
US20160133193A1 (en) * | 2014-11-10 | 2016-05-12 | Samsung Display Co., Ltd. | Display apparatus, method and apparatus for controlling the same |
US20170249434A1 (en) * | 2016-02-26 | 2017-08-31 | Daniela Brunner | Multi-format, multi-domain and multi-algorithm metalearner system and method for monitoring human health, and deriving health status and trajectory |
CN107368543A (en) * | 2017-06-28 | 2017-11-21 | 国网辽宁省电力有限公司 | A kind of electric power acquisition data restorative procedure based on MCMC algorithms |
CN107480298A (en) * | 2017-08-30 | 2017-12-15 | 北京恒华龙信数据科技有限公司 | A kind of electric quantity data restorative procedure and device |
CN107577649A (en) * | 2017-09-26 | 2018-01-12 | 广州供电局有限公司 | The interpolation processing method and device of missing data |
-
2018
- 2018-02-27 CN CN201810163770.6A patent/CN108519989A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130036082A1 (en) * | 2011-08-05 | 2013-02-07 | International Business Machines Corporation | Multiple imputation of missing data in multi-dimensional retail sales data sets via tensor factorization |
CN102521080A (en) * | 2011-12-07 | 2012-06-27 | 福建省电力有限公司 | Computer data recovery method for electricity-consumption information collecting system for power consumers |
CN103353895A (en) * | 2013-07-22 | 2013-10-16 | 广西电网公司电力科学研究院 | Pre-processing method of power distribution network line loss data |
US20160133193A1 (en) * | 2014-11-10 | 2016-05-12 | Samsung Display Co., Ltd. | Display apparatus, method and apparatus for controlling the same |
US20170249434A1 (en) * | 2016-02-26 | 2017-08-31 | Daniela Brunner | Multi-format, multi-domain and multi-algorithm metalearner system and method for monitoring human health, and deriving health status and trajectory |
CN107368543A (en) * | 2017-06-28 | 2017-11-21 | 国网辽宁省电力有限公司 | A kind of electric power acquisition data restorative procedure based on MCMC algorithms |
CN107480298A (en) * | 2017-08-30 | 2017-12-15 | 北京恒华龙信数据科技有限公司 | A kind of electric quantity data restorative procedure and device |
CN107577649A (en) * | 2017-09-26 | 2018-01-12 | 广州供电局有限公司 | The interpolation processing method and device of missing data |
Non-Patent Citations (1)
Title |
---|
谷海彤 等: ""DA多重插补法在电网电能量数据缺失处理中的应用"", 《广西科技大学学报》 * |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109445972A (en) * | 2018-09-21 | 2019-03-08 | 深圳供电局有限公司 | Data recovery method, device, equipment and storage medium |
CN109445972B (en) * | 2018-09-21 | 2022-11-04 | 深圳供电局有限公司 | Data recovery method, device, equipment and storage medium |
CN111127247A (en) * | 2019-11-21 | 2020-05-08 | 深圳供电局有限公司 | Electric quantity data acquisition method and device, computer equipment and storage medium |
CN111127247B (en) * | 2019-11-21 | 2024-01-19 | 深圳供电局有限公司 | Electric quantity data acquisition method, device, computer equipment and storage medium |
CN111125635A (en) * | 2019-11-29 | 2020-05-08 | 广东电网有限责任公司 | User table data filling method, device, terminal and storage medium |
CN111125635B (en) * | 2019-11-29 | 2023-04-18 | 广东电网有限责任公司 | User table data filling method, device, terminal and storage medium |
CN111178788A (en) * | 2020-01-09 | 2020-05-19 | 国网福建省电力有限公司 | Power data auxiliary management method and system for power selling side |
CN111667123A (en) * | 2020-06-17 | 2020-09-15 | 上海积成能源科技有限公司 | Method for supplementing missing value by applying multiple interpolation in power load prediction |
CN116415123A (en) * | 2023-03-07 | 2023-07-11 | 清华大学 | Method and system for analyzing total water flow data of community |
CN116415123B (en) * | 2023-03-07 | 2023-09-19 | 清华大学 | Method and system for analyzing total water flow data of community |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108519989A (en) | The reduction retroactive method and device of a kind of day electricity missing data | |
CN106570778B (en) | A kind of method that data integration based on big data is calculated with line loss analyzing | |
Farmer | Information dimension and the probabilistic structure of chaos | |
US8175830B2 (en) | Frequency estimation of rare events by adaptive thresholding | |
CN102426674B (en) | Power system load prediction method based on Markov chain | |
CN102509173B (en) | A kind of based on markovian power system load Accurate Prediction method | |
CN106682763B (en) | Power load optimization prediction method for large amount of sample data | |
Cheng et al. | Enhanced state estimation and bad data identification in active power distribution networks using photovoltaic power forecasting | |
CN111210093A (en) | Daily water consumption prediction method based on big data | |
CN111368259B (en) | Distribution network voltage data missing filling method and related device | |
CN108132423B (en) | A kind of method for rapidly positioning based on state transition probability power system monitoring data distortion | |
Liu et al. | An efficient simulation procedure for point estimation of expected shortfall | |
Hodge et al. | Examining information entropy approaches as wind power forecasting performance metrics | |
CN109583773A (en) | A kind of method, system and relevant apparatus that taxpaying credit integral is determining | |
CN104102804A (en) | Method and device for predicting service life of component of equipment | |
CN112906762A (en) | Method and system for defining and identifying critical state of water resource system | |
Lev | An information theory analysis of budget variances | |
CN115130924A (en) | Microgrid power equipment asset evaluation method and system under source grid storage background | |
Chen et al. | Regime-switching processes and mean-reverting volatility models in value-at-risk estimation: Evidence from the Taiwan stock index | |
Mitic | Credible value-at-risk | |
Kley-Holsteg et al. | Probabilistic multi-step-ahead short-term water demand forecasting with Lasso | |
CN114784795A (en) | Wind power prediction method and device, electronic equipment and storage medium | |
CN114595952A (en) | Electricity stealing behavior detection method based on attention network improved convolutional neural network | |
Van Montfort | Statistical remarks on laboratory‐evaluating programs for comparing laboratories and methods | |
CN111164633B (en) | Method and device for adjusting scoring card model, server and storage medium |
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 | ||
AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20221223 |
|
AD01 | Patent right deemed abandoned |