CN109410089A - Low-voltage tripping and customer complaint prediction technique, device and storage medium - Google Patents
Low-voltage tripping and customer complaint prediction technique, device and storage medium Download PDFInfo
- Publication number
- CN109410089A CN109410089A CN201811633676.9A CN201811633676A CN109410089A CN 109410089 A CN109410089 A CN 109410089A CN 201811633676 A CN201811633676 A CN 201811633676A CN 109410089 A CN109410089 A CN 109410089A
- Authority
- CN
- China
- Prior art keywords
- sample
- prediction
- data
- prediction model
- result
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 49
- 238000003860 storage Methods 0.000 title claims abstract description 12
- 238000012549 training Methods 0.000 claims abstract description 89
- 230000007613 environmental effect Effects 0.000 claims abstract description 24
- 238000005070 sampling Methods 0.000 claims description 39
- 238000012545 processing Methods 0.000 claims description 34
- 241001269238 Data Species 0.000 claims description 21
- 238000012795 verification Methods 0.000 claims description 19
- 238000004590 computer program Methods 0.000 claims description 18
- 239000011159 matrix material Substances 0.000 claims description 9
- 235000013399 edible fruits Nutrition 0.000 claims description 7
- 238000005201 scrubbing Methods 0.000 claims description 5
- 238000004140 cleaning Methods 0.000 claims description 4
- 238000009826 distribution Methods 0.000 abstract description 18
- 238000010586 diagram Methods 0.000 description 7
- 230000005611 electricity Effects 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 4
- 239000012141 concentrate Substances 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000001131 transforming effect Effects 0.000 description 3
- 230000015572 biosynthetic process Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 238000003786 synthesis reaction Methods 0.000 description 2
- 230000032683 aging Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000035515 penetration Effects 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000010076 replication Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 230000009466 transformation Effects 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
-
- 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"
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
This application involves a kind of low-voltage trippings and customer complaint prediction technique, device, computer equipment and storage medium, by obtaining equipment account data, power supply environmental attribute data and platform area user characteristic data, obtain various data, the accuracy of prediction can be increased, by the equipment account data, the power supply environmental attribute data and described area's user characteristic data are cleaned and are arranged, obtain predicted characteristics data, it can make predicted characteristics data after input prediction chain of model, avoid the identification error of prediction model chain, it is made whether prediction model chain of the predicted characteristics data input based on classifier chains model training that low-voltage tripping occurs and customer complaint prediction respectively whether occurs, export prediction result, low-voltage tripping and customer complaint are predicted using prediction model chain, it can use the area the Gong Biantai low pressure of power distribution network Correlation between tripping and customer complaint effectively improves prediction result accuracy rate and prediction coverage rate.
Description
Technical field
This application involves field of computer technology, more particularly to a kind of low-voltage tripping and customer complaint prediction technique, dress
It sets, computer equipment and storage medium.
Background technique
With the increasingly raising of economic development and people's living standard, industry and commerce and the horizontal constantly growth of residential electricity consumption.Especially
It causes power load unprecedented soaring in summer high temperature weather period, the use of high-power electric appliance, and capacity of distribution transform occur can not expire
The confession that the problems such as the phenomenon that sufficient customer electricity increased requirement, platform area voltage instability, fault trip accordingly occurs, and thus causes
Electric class customer complaint quantity remains high.Currently, the processing for the area Gong Biantai low-voltage tripping and power supply class customer complaint problem
Method has replacement switch, adjustment subscribers' line to shunt load, capacity-increasing transformation, public change the brief urgencies settlement mechanism such as dedicated and length
Remote solution.Treating method lacks the work of anticipation property based on post-processing.
As intelligent distribution network is information-based, automation, interactive horizontal continuous improvement, electric power enterprise has accumulated a large amount of electricity consumptions
Data, domestic and foreign scholars propose many distribution transforming operating status prediction models based on power distribution network big data.Current research is more
Number with distribution transforming heavy-overload is predicted as point of penetration, and marks off be likely to occur fault trip and customer complaint problem on this basis
High risk platform area.But in addition to distribution transforming heavy-overload problem, still have a large amount of low-voltage trippings be due between branch, load between platform area
Distribute what the apparatus factors such as operational managements factor and switchgear, aging circuit such as unevenness, three-phase imbalance caused, it is therefore, right
The accuracy rate that the area the Gong Biantai low-voltage tripping of power distribution network and customer complaint are predicted is low.
Summary of the invention
Based on this, it is necessary in view of the above technical problems, provide a kind of area Gong Biantai low pressure that can be improved to power distribution network
The low-voltage tripping and customer complaint prediction technique, device, computer of the problem of accuracy rate that tripping and customer complaint are predicted
Equipment and storage medium.
A kind of low-voltage tripping and customer complaint prediction technique, which comprises
Obtain equipment account data, power supply environmental attribute data and platform area user characteristic data；
The equipment account data, the power supply environmental attribute data and described area's user characteristic data are carried out clear
It washes and arranges, obtain predicted characteristics data；
The predicted characteristics data are inputted into the prediction model chain based on classifier chains model training, pass through the prediction mould
Type chain predicts whether that low-voltage tripping occurs and whether customer complaint occurs, and obtains low-voltage tripping prediction result and customer complaint prediction
As a result.
It is described in one of the embodiments, to input the predicted characteristics data based on the pre- of classifier chains model training
Chain of model is surveyed, by the prediction model chain predict whether that low-voltage tripping occurs and whether customer complaint occurs, obtains low pressure jump
The step of lock prediction result and customer complaint prediction result, comprising:
The predicted characteristics data are inputted into each sub- prediction model chain in the prediction model chain respectively, it is pre- to export each son
Survey the preliminary low-voltage tripping prediction result and preliminary customer complaint prediction result of chain of model；
Each preliminary low-voltage tripping prediction result is voted, poll is highest to be determined as low-voltage tripping prediction result；
Each preliminary customer complaint prediction result is voted, poll is highest to be determined as customer complaint prediction result.
In one of the embodiments, it is described the predicted characteristics data are inputted respectively it is each in the prediction model chain
Sub- prediction model chain exports the preliminary low-voltage tripping prediction result and preliminary customer complaint prediction result of each sub- prediction model chain
Step, comprising:
The predicted characteristics data are inputted to the first prediction model of the sub- prediction model chain, output the first prediction knot
Fruit；
The predicted characteristics data and first prediction result are inputted to the second prediction mould of the sub- prediction model chain
Type exports the second prediction result；
Second prediction result is determined as the preliminary low-voltage tripping prediction result of sub- prediction model chain output and preliminary
Customer complaint prediction result.
The training method of first prediction model includes: in one of the embodiments,
Each sample data is obtained, the sample data includes: the first of characteristic sample and the characteristic sample
As a result label；
Each sample data is based on mixing double sampling and is sampled processing, obtains training sample；
The training sample is inputted to the first prediction model to be trained, the first prediction model after being trained；
Verifying sample is obtained, the verifying sample includes: characteristic verifying sample；
By the first prediction model after the verifying sample input training, verification result is exported；
When the verification result is met the requirements, the first prediction model is obtained.
The training method of second prediction model includes: in one of the embodiments,
Each sample data is obtained, the sample data includes: the first of characteristic sample and the characteristic sample
As a result label and the second result label；
Each sample data is based on mixing double sampling and is sampled processing, obtains training sample；
The training sample is inputted to the second prediction model to be trained, the second prediction model after being trained；
Verifying sample is obtained, the verifying sample includes: characteristic verifying sample；
By the second prediction model after the verifying sample input training, verification result is exported；
When the verification result is met the requirements, the second prediction model is obtained.
It is described in one of the embodiments, that each sample data is sampled processing based on mixing double sampling, it obtains
The step of obtaining training sample, comprising:
Most class sample datas in each sample data are sampled processing using NCL sub- sampling, are trained
First training sample of sample；
Minority class sample data in each sample data is sampled processing using SMOTE oversampling, is instructed
Practice the second training sample of sample.
Most class sample datas by each sample data owe to take out using NCL in one of the embodiments,
The step of sample is sampled processing, obtains the first training sample of training sample, comprising:
It traverses each sample data and carries out data scrubbing, obtain most class sample datas；
Each most class sample datas are normalized, each treated sample data is obtained；
The Euclidean distance between each treated sample data is calculated, each treated sample data is obtained
Distance matrix；
Element in the upper triangle of the distance matrix is ranked up based on the height of similarity, obtains each sample number
According to similarity arrangement；
A sample data in sample data two-by-two is successively randomly selected according to the sequence that the similarity arranges, is obtained
First training sample of training sample.
A kind of low-voltage tripping and customer complaint prediction meanss, described device include:
Data acquisition module, for obtaining equipment account data, power supply environmental attribute data and platform area user characteristics number
According to；
Data processing module is used for the equipment account data, the power supply environmental attribute data and described area
User characteristic data is cleaned and is arranged, and predicted characteristics data are obtained；
Prediction module, for the predicted characteristics data to be inputted the prediction model chain based on classifier chains model training,
It by the prediction model chain predicts whether that low-voltage tripping occurs and whether customer complaint occurs, obtain low-voltage tripping prediction result
With customer complaint prediction result.
A kind of computer equipment, including memory and processor, the memory are stored with computer program, the processing
The step of device realizes the method when executing the computer program.
A kind of computer readable storage medium, is stored thereon with computer program, and the computer program is held by processor
The step of method is realized when row.
Above-mentioned low-voltage tripping and customer complaint prediction technique, device, computer equipment and storage medium, by obtaining equipment
Account data, power supply environmental attribute data and platform area user characteristic data, obtain various data, can increase prediction
Accuracy carries out the equipment account data, the power supply environmental attribute data and described area's user characteristic data clear
It washes and arranges, obtain predicted characteristics data, can make predicted characteristics data after input prediction chain of model, avoid prediction model chain
Prediction model chain of the predicted characteristics data input based on classifier chains model training is made whether to occur low by identification error
Whether pressure tripping occurs customer complaint prediction with each, exports prediction result, is thrown using prediction model chain low-voltage tripping and client
It tells prediction, can use the correlation between the area the Gong Biantai low-voltage tripping of power distribution network and customer complaint, effectively improve prediction knot
Fruit accuracy rate and prediction coverage rate.
Detailed description of the invention
Fig. 1 is the application scenario diagram of one embodiment mesolow tripping and customer complaint prediction technique；
Fig. 2 is the flow diagram of one embodiment mesolow tripping and customer complaint prediction technique；
Fig. 3 is the flow diagram of one embodiment mesolow tripping and customer complaint prediction technique；
Fig. 4 is the flow diagram of another embodiment mesolow tripping and customer complaint prediction technique；
Fig. 5 is that the training process of the first prediction model of the tripping of one embodiment mesolow and customer complaint prediction technique is shown
It is intended to；
Fig. 6 is the training process of the second prediction model of the tripping of another embodiment mesolow and customer complaint prediction technique
Schematic diagram；
Fig. 7 is the structural block diagram of one embodiment mesolow tripping and customer complaint prediction meanss；
Fig. 8 is the internal structure chart of computer equipment in one embodiment.
Specific embodiment
It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, not
For limiting the application.
Low-voltage tripping provided by the present application and customer complaint prediction technique, can be applied to application environment as shown in Figure 1
In.Wherein, terminal 102 is communicated with server 104 by network by network.Server 104 is set by the acquisition of terminal 102
Standby account data, power supply environmental attribute data and platform area user characteristic data；Server 104 by equipment account data, power supply
Environmental attribute data and platform area user characteristic data are cleaned and are arranged, and predicted characteristics data are obtained；Server 104 will be pre-
It surveys characteristic and inputs the prediction model chain based on classifier chains model training, predict whether that low pressure occurs by prediction model chain
It trips and whether customer complaint occurs, obtain low-voltage tripping prediction result and customer complaint prediction result.Wherein, terminal 102 can
With but be not limited to various personal computers, laptop, smart phone, tablet computer and portable wearable device, take
Business device 104 can be realized with the server cluster of the either multiple server compositions of independent server.
In one embodiment, as shown in Fig. 2, a kind of low-voltage tripping and customer complaint prediction technique are provided, with the party
Method is applied to be illustrated for the server in Fig. 1, including step S220 to step S260:
Step S220 obtains equipment account data, power supply environmental attribute data and platform area user characteristic data.
Wherein, equipment account data include: the data such as transformer capacity, low-voltage customer number, the duration that puts into operation, switchtype,
Transformer capacity can be assignment and become rated capacity (kVA), and low-voltage customer number can refer to low-voltage customer quantity in platform area, put into operation
Duration can be assignment change and put into operation duration (year), and switchtype can refer to trip switch device model.Power supply environment attribute number
According to including: the data such as temperature, distribution transformer load rate, festivals or holidays label, temperature can be in a few days mean temperature (DEG C), distribution transformer load rate
It can be the day highest load factor that assignment becomes, festivals or holidays label can refer to whether be the legal festivals and holidays.Platform area user characteristics number
According to including: the data such as load nature of electricity consumed, regionalism, load nature of electricity consumed can refer to resident, industry, business and synthesis, regionalism
It can refer to city, cities and towns, villages within the city.Equipment account data refer to the equipment account data in the area Gong Biantai of power distribution network, supply
Electrical environment attribute data refers to the power supply environmental attribute data in the area Gong Biantai of power distribution network, and platform area user characteristic data refers to
The platform area user characteristic data in the area Gong Biantai of power distribution network.
Step S240 cleans equipment account data, power supply environmental attribute data and platform area user characteristic data
And arrangement, obtain predicted characteristics data.
Wherein, the data of equipment account data, power supply environmental attribute data and platform area user characteristic data are carried out clear
Reason, such as: equipment account data, power supply environmental attribute data and platform area user characteristic data are examined and are verified again,
Duplicate data information is deleted, and corrects existing wrong data.By equipment account data, the power supply environment attribute number after cleaning
Accordingly and platform area user characteristic data carry out data preparation, such as: transformer capacity, low-voltage customer number, the duration that puts into operation data, can
To be handled by the processing mode of branch mailbox, acquisition predicted characteristics data, temperature, distribution transformer load rate, the data of load nature of electricity consumed,
It can be handled by standardized processing mode, obtain predicted characteristics data, regionalism, switchtype, festivals or holidays mark
The data of label can be handled by the processing mode of coding, obtain predicted characteristics data.
Predicted characteristics data are inputted the prediction model chain based on classifier chains model training, pass through prediction by step S260
Chain of model predicts whether that low-voltage tripping occurs and whether customer complaint occurs, and obtains low-voltage tripping prediction result and customer complaint is pre-
Survey result.
Wherein, classifier chains (Classifier chains, CC) are a kind of multi-tag classification proposed by READ J et al.
Algorithm, in order to make full use of the correlation between label, classifier chains obtain each fundamental classifier during prediction
Prediction result is added to the characteristic variable space of all fundamental classifiers thereafter, provides predictive information for other labels, forms
The classifier of one chain form.Low-voltage tripping prediction result can have: low-voltage tripping occurs or there is no low-voltage trippings；Visitor
It complains prediction result that can have in family: customer complaint occurs or there is no customer complaints.Low-voltage tripping prediction result refers to institute
The low-voltage tripping prediction result in the area Gong Biantai of the power distribution network of prediction, customer complaint prediction result refer to predicted power distribution network
The area Gong Biantai power supply class client customer complaint prediction result.Such as: prediction target is the low-voltage tripping and public affairs in the area Gong Biantai
Two respective labels of customer complaint in the area Bian Tai, it is now assumed that characteristic variable be X, prediction label whether occur low-voltage tripping be y1 and
Customer service whether occurs to complain to be y2, can be predicted by a prediction model in characteristic variable X input prediction chain of model
Another prediction model in prediction result y1 and characteristic variable X input prediction chain of model is obtained prediction result by as a result y1
Y2 is also possible to obtain prediction result y2 by a prediction model in characteristic variable X input prediction chain of model, will predict
As a result another prediction model in y2 and characteristic variable X input prediction chain of model obtains prediction result y1, and exports y1 and y2
Prediction result.
In above-mentioned low-voltage tripping and customer complaint prediction technique, by obtaining equipment account data, power supply environment attribute number
Accordingly and platform area user characteristic data, various data are obtained, the accuracy of prediction can be increased, by the equipment account number
It is cleaned and is arranged according to, the power supply environmental attribute data and described area's user characteristic data, obtain predicted characteristics number
According to, can make predicted characteristics data after input prediction chain of model, avoid prediction model chain identification error, by the predicted characteristics
Prediction model chain of the data input based on classifier chains model training is made whether that low-voltage tripping occurs and respectively whether client occurs
Prediction is complained, prediction result is exported, low-voltage tripping and customer complaint are predicted using prediction model chain, can use low-voltage tripping
Correlation between customer complaint effectively improves prediction result accuracy rate and prediction coverage rate.
In one embodiment, predicted characteristics data are inputted into the prediction model chain based on classifier chains model training, led to
It crosses prediction model chain to predict whether that low-voltage tripping occurs and whether customer complaint occurs, obtains low-voltage tripping prediction result and client
The step of complaining prediction result, comprising:
Predicted characteristics data are distinguished into each sub- prediction model chain in input prediction chain of model, export each sub- prediction model chain
Preliminary low-voltage tripping prediction result and preliminary customer complaint prediction result；Each preliminary low-voltage tripping prediction result is thrown
Ticket, poll is highest to be determined as low-voltage tripping prediction result；Each preliminary customer complaint prediction result is voted, poll highest
Be determined as customer complaint prediction result.
Wherein, as shown in figure 3, there can be multiple sub- prediction model chains in prediction model chain, predicted characteristics data are inputted
Each sub- prediction model chain, each sub- prediction model chain can all export preliminary low-voltage tripping prediction result and preliminary customer complaint prediction knot
Fruit votes to preliminary low-voltage tripping prediction result and preliminary customer complaint prediction result, obtains prediction result.Such as: assuming that
Prediction model chain has sub- prediction model chain 1, sub- prediction model chain 2, sub- prediction model chain 3, sub- prediction model chain 4, and prediction is special
Sign data input sub- prediction model chain 1, sub- prediction model chain 2, sub- prediction model chain 3, sub- prediction model chain 4, sub- prediction respectively
Chain of model 1 exports preliminary low-voltage tripping prediction result as low-voltage tripping occurs, and preliminary customer complaint prediction result is that client occurs
It complains；It is that there is no low-voltage tripping, preliminary customer complaint predictions that sub- prediction model chain 2, which exports preliminary low-voltage tripping prediction result,
It as a result is generation customer complaint；Sub- prediction model chain 3 exports preliminary low-voltage tripping prediction result as low-voltage tripping, preliminary visitor occurs
It is that customer complaint occurs that prediction result is complained at family；Sub- prediction model chain 4 exports preliminary low-voltage tripping prediction result as low pressure occurs
Tripping, preliminary customer complaint prediction result are that there is no customer complaints；Prediction result based on each sub- prediction model chain carries out
Ballot, it is 1 ticket there is no the poll of low-voltage tripping, the poll that customer complaint occurs is that the poll that low-voltage tripping occurs, which is 3 tickets,
3 tickets are 1 ticket there is no the poll of customer complaint, then the low-voltage tripping prediction result of prediction model chain output is that low pressure occurs
Tripping, the customer complaint prediction result of prediction model chain output are that customer complaint occurs.Using prediction model chain to low-voltage tripping
It is predicted with customer complaint, can use the correlation between low-voltage tripping and customer complaint, effectively improve prediction result accuracy rate
With prediction coverage rate.
In one embodiment, predicted characteristics data are distinguished into each sub- prediction model chain in input prediction chain of model, it is defeated
The step of preliminary low-voltage tripping prediction result and preliminary customer complaint prediction result of each sub- prediction model chain out, comprising:
Predicted characteristics data are inputted to the first prediction model of sub- prediction model chain, export the first prediction result；It will prediction
Characteristic and the first prediction result input the second prediction model of sub- prediction model chain, export the second prediction result；By second
Prediction result is determined as the preliminary low-voltage tripping prediction result and preliminary customer complaint prediction result of sub- prediction model chain output.
Wherein, the first prediction model in each sub- prediction model chain can be for predicting preliminary low-voltage tripping prediction result
Model, be also possible to the model for predicting preliminary customer complaint prediction result, in group prediction model chain first prediction
When model is the model for predicting preliminary low-voltage tripping prediction result, the second prediction model is then for predicting that preliminary client throws
Predicted characteristics data are inputted the first prediction model by the model for telling prediction result, and the first prediction model predicts preliminary low pressure and jumps
Lock prediction result exports preliminary low-voltage tripping prediction result (i.e. the first prediction result), by preliminary low-voltage tripping prediction result and
Predicted characteristics data input the second prediction model, and the second prediction model is based on preliminary low-voltage tripping prediction result and predicted characteristics number
It is predicted that preliminary customer complaint prediction result out, and export preliminary low-voltage tripping prediction result and preliminary customer complaint prediction result
(i.e. the second prediction result).
When the first prediction model in group prediction model chain is the model for predicting preliminary customer complaint prediction result,
Second prediction model is then the model for predicting preliminary low-voltage tripping prediction result, by the first prediction of predicted characteristics data input
Model, the first prediction model predict preliminary customer complaint prediction result, and exporting preliminary customer complaint prediction result, (i.e. first is pre-
Survey result), preliminary customer complaint prediction result and predicted characteristics data are inputted into the second prediction model, the second prediction model is based on
Preliminary customer complaint prediction result and predicted characteristics data predict preliminary low-voltage tripping prediction result, and export preliminary low pressure and jump
Lock prediction result and preliminary customer complaint prediction result (i.e. the second prediction result).Using prediction model chain to low-voltage tripping and visitor
Prediction is complained at family, can use the correlation between low-voltage tripping and customer complaint, effectively improves prediction result accuracy rate and pre-
Survey coverage rate.
In one embodiment, as shown in figure 4, predicted characteristics data to be inputted to the first prediction mould of each sub- prediction model chain
Type exports each first prediction result；It is corresponding that predicted characteristics data and each first prediction result are inputted into each sub- prediction model chain
Second prediction model exports each second prediction result；Each second prediction result is corresponded to and is determined as each sub- prediction model chain output
Preliminary low-voltage tripping prediction result and preliminary customer complaint prediction result, to each preliminary low-voltage tripping prediction result and each preliminary
Customer complaint prediction result is voted, and prediction result is obtained.Low-voltage tripping and customer complaint are predicted using prediction model chain,
It can use the correlation between low-voltage tripping and customer complaint, effectively improve prediction result accuracy rate and prediction coverage rate.
Assuming that prediction model chain has sub- prediction model chain 1, sub- prediction model chain 2, the first prediction in sub- prediction model chain 1
Model is the model for predicting preliminary low-voltage tripping prediction result, and the first prediction model in sub- prediction model chain 2 is to be used for
Predicted characteristics data are inputted the first prediction mould in sub- prediction model chain 1 by the model for predicting preliminary customer complaint prediction result
Type, the preliminary low-voltage tripping prediction result of output are that low-voltage tripping occurs, and low-voltage tripping will occur and predicted characteristics data input
The second prediction model in sub- prediction model chain 1, the preliminary low-voltage tripping prediction result of output and preliminary customer complaint prediction knot
Fruit are as follows: low-voltage tripping occurs and customer complaint occurs；Predicted characteristics data are inputted to the first prediction mould in sub- prediction model chain 2
Type, the preliminary customer complaint prediction result of output are that customer complaint occurs, and customer complaint will occur and predicted characteristics data input
The second prediction model in sub- prediction model chain 2, the preliminary low-voltage tripping prediction result of output and preliminary customer complaint prediction knot
Fruit are as follows: occur low-voltage tripping and occur customer complaint, the prediction result based on sub- prediction model chain 1 and sub- prediction model chain 2 into
Row ballot, the poll that low-voltage tripping occurs is 2 tickets, is 0 ticket there is no the poll of low-voltage tripping, the poll of customer complaint occurs
It is 0 ticket there is no the poll of customer complaint for 2 tickets, then the low-voltage tripping prediction result of prediction model chain output and client throw
Tell prediction result are as follows: low-voltage tripping occurs and customer complaint occurs.
In one embodiment, the training method of the first prediction model includes: to obtain each sample data, sample data packet
It includes: the first result label of characteristic sample and characteristic sample；Each sample data is based on mixing double sampling to take out
Sample processing, obtains training sample；Training sample is inputted to the first prediction model to be trained, the first prediction mould after being trained
Type；Verifying sample is obtained, verifying sample includes: characteristic verifying sample；The first prediction after sample input is trained will be verified
Model exports verification result；When verification result is met the requirements, the first prediction model is obtained.
Wherein, when the first prediction model is trained, using the characteristic sample in sample data as characteristic variable, sample
First result label of the characteristic sample in notebook data is as prediction label.Characteristic sample includes a large amount of table of equipment
Account data, a large amount of power supply environmental attribute data and a large amount of platform area user characteristic data, the first knot of characteristic sample
Fruit label refers to the result label according to the corresponding first result setting of each characteristic sample, as shown in figure 5, by each sample
Data are based on mixing double sampling and are sampled processing, obtain training sample, and training sample is inputted to the first prediction mould to be trained
Type, the first result of the first prediction model to be trained based on characteristic sample and characteristic sample in each training sample
Label is trained, the first prediction model after being trained；Verifying sample is obtained, verifying sample includes: characteristic verifying
Sample；The first prediction model after sample input is trained will be verified, exports verification result；When verification result is met the requirements, obtain
Obtain the first prediction model.
When the first result label of characteristic sample be according to the corresponding low-voltage tripping result of each characteristic sample (i.e.
First result) setting result label, obtain the first prediction model be the model for predicting preliminary low-voltage tripping prediction result, work as spy
The the first result label for levying data sample is set according to the corresponding customer complaint result of each characteristic sample (i.e. the first result)
The result label set, obtaining the first prediction model is the model for predicting preliminary customer complaint prediction result.By mixing double sampling
It is sampled processing, obtains training sample, avoids the problem that important information loss and information redundancy.
In one embodiment, the training method of the second prediction model includes: to obtain each sample data, sample data packet
It includes: the first result label and the second result label of characteristic sample and characteristic sample；Each sample data is based on mixed
It closes double sampling and is sampled processing, obtain training sample；Training sample is inputted to the second prediction model to be trained, is trained
The second prediction model afterwards；Verifying sample is obtained, verifying sample includes: characteristic verifying sample；It will verifying sample input instruction
The second prediction model after white silk exports verification result；When verification result is met the requirements, the second prediction model is obtained.
Wherein, when the second prediction model is trained, by the characteristic sample and characteristic sample in sample data
The first result label as characteristic variable, the second result label of the characteristic sample in sample data is as pre- mark
Label.Characteristic sample includes that a large amount of equipment account data, a large amount of power supply environmental attribute data and a large amount of platform area use
Family characteristic, the first result label of characteristic sample refer to being set according to corresponding first result of each characteristic sample
Second result label of the result label set, characteristic sample is referred to according to each characteristic sample and characteristic sample
The corresponding second result setting of the first result result label, the first result can refer to customer complaint as a result, being also possible to
Refer to low-voltage tripping as a result, the second result is then low-voltage tripping as a result, when the first result when the first result is customer complaint result
When for low-voltage tripping result, the second result be then customer complaint as a result,
As shown in fig. 6, each sample data, which is based on mixing double sampling, is sampled processing, training sample is obtained, will be trained
Sample inputs the second prediction model to be trained, and the second prediction model to be trained is based on the characteristic sample in each training sample
Originally, the second result label of the first result label of characteristic sample and characteristic sample is trained, after being trained
The second prediction model；Verifying sample is obtained, verifying sample includes: characteristic verifying sample；It will verifying sample input training
The second prediction model afterwards exports verification result；When verification result is met the requirements, the second prediction model is obtained.
When the first result label of characteristic sample is the result label of low-voltage tripping result, the second prediction mould is obtained
Type is to predict the model of preliminary customer complaint prediction result, when the first result label of characteristic sample is customer complaint result
Result label, obtain the second prediction model be the model for predicting preliminary low-voltage tripping prediction result.Based on NCL sub- sampling and
The mixing repeat replication of SMOTE oversampling combination can be effectively reduced oversampling algorithm and largely add synthesis sample, may make
At minority class sample information redundancy issue and sub- sampling algorithm be easily lost in most class samples part important information the two
Influence of the problem to predictablity rate and prediction coverage rate.
Wherein, before carrying out mixing double sampling, it is necessary first to which the training sample that confirmation is formed through double sampling is concentrated a small number of
The target proportion of class and most class training samples, and thus calculate the number of oversampling and sub- sampling addition and removal.Assuming that one
It is n1 that a unbalanced data, which concentrates the quantity of minority class training sample, and the quantity of most class training samples is n0, through double sampling shape
At training sample to concentrate minority class and the target accounting of most class training samples be k1:k0, such as sub- sampling and SMOTE oversampling
Combined mixing double sampling, oversampling and sub- sampling need the training sample number N for adding and removing that can be calculated by formula 1:
N=round [(k1n0)-(k0n1)] (1)
In formula 1, round indicates to be rounded calculated result by rounding up；Wherein k1+k0=1, therefore only need to set
Determine the i.e. controllable mixing double sampling target proportion of parameter k1；To ensure that training sample concentrates most class training samples numbers to be not less than
Minority class and N value are not negative, and the upper limit value of k1 is set as 0.5, and lower limit value is the minority class accounting that sample data is concentrated: n1/
(n1+n0)。
In one embodiment, each sample data is based on mixing double sampling and is sampled processing, obtain training sample
Step, comprising: most class sample datas in each sample data are sampled processing using NCL sub- sampling, obtain training sample
This first training sample；Minority class sample data in each sample data is sampled processing using SMOTE oversampling, is obtained
Obtain the second training sample of training sample.
Wherein, NCL sub- sampling refers to retaining all minority class sample datas, and to class samples most present on its neighborhood
Notebook data is cleared up.The NCL sub- sampling on basis is divided into two steps: step 1, traverses each sample that sample data is concentrated,
Three of them closest sample is found out, if it is a small number of that arbitrary sample x, which belongs in most classes and three of them closest sample at least two,
Class, then x is identified as noise data, is cleared up；Step 2, if sample x belongs in minority class and three of them closest sample
At least two be most classes, then clears up most class samples in neighbouring sample.
SMOTE oversampling is a kind of oversampling method for solving the problems, such as unbalanced data that Chawla et al. is proposed, master
Wanting thought is to synthesize new samples between two neighbouring samples by stochastic linear interpolation method, thus what acquisition specified number
Sample.In this application, oversampling is carried out to the minority class sample in sample data using SMOTE algorithm.Synthesize sample XnewMeter
Calculation mode such as formula 2:
In formula 2: rand (0,1) indicates a random number of section (0,1), and x is any minority class sample,For the k of x
Random one in a closest sample.In the application, k default setting is 5, and oversampling number is N.
In one embodiment, most class sample datas in each sample data are sampled place using NCL sub- sampling
The step of managing, obtaining the first training sample of training sample, comprising: each sample data of traversal carries out data scrubbing, obtains
Most class sample datas；Each most class sample datas are normalized, each treated sample data is obtained；Meter
The Euclidean distance between each treated sample data is calculated, the distance matrix of each treated sample data is obtained；It adjusts the distance square
Element in the upper triangle of battle array is ranked up based on the height of similarity, obtains the similarity arrangement of each sample data；According to
The sequence of similarity arrangement successively randomly selects a sample data in sample data two-by-two, obtains the first instruction of training sample
Practice sample.
Wherein, traversing the step of each sample data carries out data scrubbing, obtains most class sample datas includes: step
One, traversal sample data concentrate each sample, find out three of them closest sample, if arbitrary sample x belong to most classes and its
At least two be minority class in three closest samples, then x is identified as noise data, is cleared up；Step 2, if sample x
Belonging in minority class and three of them closest sample at least two is most classes, then clears up most class samples in neighbouring sample,
Remaining most class samples are determined as most class sample datas.Element in upper triangle can have it is multiple, in upper triangle
Each element represents is the distance between sample calculated value two-by-two, be expressed as similarity.
NCL sub- sampling based on basis is sampled most class sample datas of acquisition, most class numbers of discovery NCL cleaning
Mesh is on the low side, and minority class classification performance cannot be significantly improved after sub- sampling.Therefore, clear in each sample data progress data of traversal
Reason is normalized most class sample datas on the basis of obtaining most class sample datas, obtains each treated sample
Notebook data, and calculate the Euclidean distance between each pair of sample in most class sample datas obtains each treated sample data
Distance matrix.Wherein, the distance between p-th and q-th sample is expressed as DIP, q, DIP, qIt is smaller to illustrate sample p and sample q
Similarity is higher.Then, the upper triangular element according to DI (distance matrix) carries out the similarity of sample two-by-two from high to low
Sequence, obtains the similarity arrangement of each sample data, and successively randomly selects the corresponding each pair of sample of similarity according to putting in order
One of them in this is cleared up, until the most class quantity summations cleared up with two steps in front reach N, NCL sub- sampling
Stop, remaining sample is determined as to the first training sample of training sample.The most class numbers for solving NCL cleaning are on the low side,
The problem of minority class classification performance cannot be significantly improved after sub- sampling.
It should be understood that although each step in the flow chart of Fig. 2 is successively shown according to the instruction of arrow, this
A little steps are not that the inevitable sequence according to arrow instruction successively executes.Unless expressly state otherwise herein, these steps
It executes there is no the limitation of stringent sequence, these steps can execute in other order.Moreover, at least part in Fig. 2
Step may include that perhaps these sub-steps of multiple stages or stage are executed in synchronization to multiple sub-steps
It completes, but can execute at different times, the execution sequence in these sub-steps or stage, which is also not necessarily, successively to be carried out,
But it can be executed in turn or alternately at least part of the sub-step or stage of other steps or other steps.
In one embodiment, as shown in fig. 7, providing a kind of low-voltage tripping and customer complaint prediction meanss, comprising: number
According to acquisition module 310, data processing module 320 and prediction module 330, in which:
Data acquisition module 310, for obtaining equipment account data, power supply environmental attribute data and platform area user characteristics
Data；
Data processing module 320 is used for the equipment account data, the power supply environmental attribute data and described
Area's user characteristic data is cleaned and is arranged, and predicted characteristics data are obtained；
Prediction module 330, for the predicted characteristics data to be inputted the prediction model based on classifier chains model training
Chain by the prediction model chain predicts whether that low-voltage tripping occurs and whether customer complaint occurs, and obtains low-voltage tripping prediction
As a result with customer complaint prediction result.
In one embodiment, prediction module 330 further include: tentative prediction unit, for predicted characteristics data to be distinguished
Each sub- prediction model chain in input prediction chain of model exports the preliminary low-voltage tripping prediction result and just of each sub- prediction model chain
Walk customer complaint prediction result；Low-voltage tripping ballot unit, for each preliminary low-voltage tripping prediction result to be voted, poll
It is highest to be determined as low-voltage tripping prediction result；Customer complaint vote unit, for by each preliminary customer complaint prediction result into
Row ballot, poll is highest to be determined as customer complaint prediction result.
In one embodiment, tentative prediction unit is also used to: predicted characteristics data are inputted the of sub- prediction model chain
One prediction model exports the first prediction result；Predicted characteristics data and the first prediction result are inputted the of sub- prediction model chain
Two prediction models export the second prediction result；The preliminary low pressure that second prediction result is determined as sub- prediction model chain output is jumped
Lock prediction result and preliminary customer complaint prediction result.
In one embodiment, low-voltage tripping and customer complaint prediction meanss further include: sample data obtains module, is used for
Each sample data is obtained, sample data includes: the first result label of characteristic sample and characteristic sample；Sample sampling
Module is sampled processing for each sample data to be based on mixing double sampling, obtains training sample；Model training module is used
In training sample to be inputted to the first prediction model to be trained, the first prediction model after being trained；Authentication module, for obtaining
Verifying sample is taken, verifying sample includes: characteristic verifying sample, the first prediction model after sample input is trained will be verified,
Verification result is exported, when verification result is met the requirements, obtains the first prediction model.
In one embodiment, sample data obtains module and is also used to: obtaining each sample data, sample data includes: spy
Levy the first result label and the second result label of data sample and characteristic sample；Sample decimation blocks are also used to: will be each
Sample data is based on mixing double sampling and is sampled processing, obtains training sample；Model training module is also used to: by training sample
Input the second prediction model to be trained, the second prediction model after being trained；Authentication module is also used to: obtaining verifying sample
This, verifying sample includes: characteristic verifying sample；The second prediction model after sample input is trained, output verifying will be verified
As a result；When verification result is met the requirements, the second prediction model is obtained.
In one embodiment, sample decimation blocks include: first sample sampling unit, for will be in each sample data
Most class sample datas are sampled processing using NCL sub- sampling, obtain the first training sample of training sample；Second sample is taken out
Sample unit is trained for the minority class sample data in each sample data to be sampled processing using SMOTE oversampling
Second training sample of sample.
In one embodiment, first sample sampling unit is also used to: each sample data of traversal carries out data scrubbing,
Obtain most class sample datas；Each most class sample datas are normalized, each treated sample number is obtained
According to；The Euclidean distance between each treated sample data is calculated, the distance matrix of each treated sample data is obtained；To away from
It is ranked up from the element in the upper triangle of matrix based on the height of similarity, obtains the similarity arrangement of each sample data；
A sample data in sample data two-by-two is successively randomly selected according to the sequence that similarity arranges, obtains the of training sample
One training sample.
Specific restriction about low-voltage tripping and customer complaint prediction meanss may refer to above for low-voltage tripping and
The restriction of customer complaint prediction technique, details are not described herein.Each mould in above-mentioned low-voltage tripping and customer complaint prediction meanss
Block can be realized fully or partially through software, hardware and combinations thereof.Above-mentioned each module can be embedded in the form of hardware or independence
In processor in computer equipment, it can also be stored in a software form in the memory in computer equipment, in order to
Processor, which calls, executes the corresponding operation of the above modules.
In one embodiment, a kind of computer equipment is provided, which can be server, internal junction
Composition can be as shown in Figure 8.The computer equipment include by system bus connect processor, memory, network interface and
Database.Wherein, the processor of the computer equipment is for providing calculating and control ability.The memory packet of the computer equipment
Include non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program and data
Library.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The calculating
The database of machine equipment is for storing sample data.The network interface of the computer equipment is used to pass through network with external terminal
Connection communication.To realize a kind of low-voltage tripping and customer complaint prediction technique when the computer program is executed by processor.
It will be understood by those skilled in the art that structure shown in Fig. 8, only part relevant to application scheme is tied
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme, specific computer equipment
It may include perhaps combining certain components or with different component layouts than more or fewer components as shown in the figure.
A kind of computer equipment, including memory and processor, the memory are stored with computer program, the processing
The step of device realizes low-voltage tripping and customer complaint prediction technique when executing the computer program.
A kind of computer readable storage medium, is stored thereon with computer program, and the computer program is held by processor
The step of low-voltage tripping and customer complaint prediction technique are realized when row.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
Each technical characteristic of above embodiments can be combined arbitrarily, for simplicity of description, not to above-described embodiment
In each technical characteristic it is all possible combination be all described, as long as however, the combination of these technical characteristics be not present lance
Shield all should be considered as described in this specification.
The several embodiments of the application above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the concept of this application, various modifications and improvements can be made, these belong to the protection of the application
Range.Therefore, the scope of protection shall be subject to the appended claims for the application patent.
Claims (10)
1. a kind of low-voltage tripping and customer complaint prediction technique, which comprises
Obtain equipment account data, power supply environmental attribute data and platform area user characteristic data；
By the equipment account data, the power supply environmental attribute data and described area's user characteristic data carry out cleaning and
It arranges, obtains predicted characteristics data；
The predicted characteristics data are inputted into the prediction model chain based on classifier chains model training, pass through the prediction model chain
It predicts whether that low-voltage tripping occurs and whether customer complaint occurs, obtain low-voltage tripping prediction result and customer complaint prediction knot
Fruit.
2. the method according to claim 1, wherein described be based on classifier for predicted characteristics data input
The prediction model chain of chain model training by the prediction model chain predicts whether that low-voltage tripping occurs and client whether occurs to throw
The step of telling, obtaining low-voltage tripping prediction result and customer complaint prediction result, comprising:
The predicted characteristics data are inputted into each sub- prediction model chain in the prediction model chain respectively, export each sub- prediction mould
The preliminary low-voltage tripping prediction result and preliminary customer complaint prediction result of type chain；
Each preliminary low-voltage tripping prediction result is voted, poll is highest to be determined as low-voltage tripping prediction result；
Each preliminary customer complaint prediction result is voted, poll is highest to be determined as customer complaint prediction result.
3. according to the method described in claim 2, it is characterized in that, it is described the predicted characteristics data are inputted respectively it is described pre-
Survey each sub- prediction model chain in chain of model, the preliminary low-voltage tripping prediction result of each sub- prediction model chain of output and preliminary client
The step of complaining prediction result, comprising:
The predicted characteristics data are inputted to the first prediction model of the sub- prediction model chain, export the first prediction result；
The predicted characteristics data and first prediction result are inputted to the second prediction model of the sub- prediction model chain, it is defeated
Second prediction result out；
Second prediction result is determined as the preliminary low-voltage tripping prediction result and preliminary client that sub- prediction model chain exports
Complain prediction result.
4. according to the method described in claim 3, it is characterized in that, the training method of first prediction model includes:
Each sample data is obtained, the sample data includes: the first result of characteristic sample and the characteristic sample
Label；
Each sample data is based on mixing double sampling and is sampled processing, obtains training sample；
The training sample is inputted to the first prediction model to be trained, the first prediction model after being trained；
Verifying sample is obtained, the verifying sample includes: characteristic verifying sample；
By the first prediction model after the verifying sample input training, verification result is exported；
When the verification result is met the requirements, the first prediction model is obtained.
5. according to the method described in claim 3, it is characterized in that, the training method of second prediction model includes:
Each sample data is obtained, the sample data includes: the first result of characteristic sample and the characteristic sample
Label and the second result label；
Each sample data is based on mixing double sampling and is sampled processing, obtains training sample；
The training sample is inputted to the second prediction model to be trained, the second prediction model after being trained；
Verifying sample is obtained, the verifying sample includes: characteristic verifying sample；
By the second prediction model after the verifying sample input training, verification result is exported；
When the verification result is met the requirements, the second prediction model is obtained.
6. according to the described in any item methods of claim 4 or 5, which is characterized in that described to be based on mixing by each sample data
Close the step of double sampling is sampled processing, obtains training sample, comprising:
Most class sample datas in each sample data are sampled processing using NCL sub- sampling, obtain training sample
The first training sample；
Minority class sample data in each sample data is sampled processing using SMOTE oversampling, obtains training sample
This second training sample.
7. according to the described in any item methods of claim 6, which is characterized in that most classes by each sample data
The step of sample data is sampled processing, is obtained the first training sample of training sample using NCL sub- sampling, comprising:
It traverses each sample data and carries out data scrubbing, obtain most class sample datas；
Each most class sample datas are normalized, each treated sample data is obtained；
The Euclidean distance between each treated sample data is calculated, the distance of each treated sample data is obtained
Matrix；
Element in the upper triangle of the distance matrix is ranked up based on the height of similarity, obtains each sample data
Similarity arrangement；
A sample data in sample data two-by-two is successively randomly selected according to the sequence that the similarity arranges, is trained
First training sample of sample.
8. a kind of low-voltage tripping and customer complaint prediction meanss, which is characterized in that described device includes:
Data acquisition module, for obtaining equipment account data, power supply environmental attribute data and platform area user characteristic data；
Data processing module is used for the equipment account data, the power supply environmental attribute data and described area user
Characteristic is cleaned and is arranged, and predicted characteristics data are obtained；
Prediction module passes through for the predicted characteristics data to be inputted the prediction model chain based on classifier chains model training
The prediction model chain predicts whether that low-voltage tripping occurs and whether customer complaint occurs, and obtains low-voltage tripping prediction result and visitor
Complain prediction result in family.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the step of processor realizes any one of claims 1 to 7 the method when executing the computer program.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any one of claims 1 to 7 is realized when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811633676.9A CN109410089B (en) | 2018-12-29 | 2018-12-29 | Low-voltage trip and customer complaint prediction method, device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811633676.9A CN109410089B (en) | 2018-12-29 | 2018-12-29 | Low-voltage trip and customer complaint prediction method, device and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109410089A true CN109410089A (en) | 2019-03-01 |
| CN109410089B CN109410089B (en) | 2020-11-03 |
Family
ID=65462757
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811633676.9A Active CN109410089B (en) | 2018-12-29 | 2018-12-29 | Low-voltage trip and customer complaint prediction method, device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109410089B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110033140A (en) * | 2019-04-22 | 2019-07-19 | 广东工业大学 | A kind of distribute-electricity transformer district tripping prediction technique, system and device |
| CN110135614A (en) * | 2019-03-26 | 2019-08-16 | 广东工业大学 | It is a kind of to be tripped prediction technique based on rejecting outliers and the 10kV distribution low-voltage of sampling techniques |
| CN111178957A (en) * | 2019-12-23 | 2020-05-19 | 广西电网有限责任公司 | Method for early warning sudden increase of electric quantity of electricity consumption customer |
| CN111382897A (en) * | 2019-10-25 | 2020-07-07 | 广州供电局有限公司 | Transformer area low-voltage trip prediction method and device, computer equipment and storage medium |
| CN115456210A (en) * | 2022-08-22 | 2022-12-09 | 国网浙江省电力有限公司杭州市临安区供电公司 | Power utilization complaint early warning method based on cascade logistic regression Bayesian algorithm |
| CN115879586A (en) * | 2022-01-11 | 2023-03-31 | 北京中关村科金技术有限公司 | Complaint prediction optimization method and device based on ablation experiment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104376510A (en) * | 2014-12-05 | 2015-02-25 | 国家电网公司 | Method of predicting and accessing level of wildfire-caused trip risk in power transmission lines |
| CN106980929A (en) * | 2017-01-05 | 2017-07-25 | 国网福建省电力有限公司 | A kind of power failure complaint risk Forecasting Methodology based on random forest |
| CN107506849A (en) * | 2017-07-24 | 2017-12-22 | 国网江西省电力公司电力科学研究院 | A kind of intelligent optimization distribution transforming, which has a power failure, studies and judges system |
-
2018
- 2018-12-29 CN CN201811633676.9A patent/CN109410089B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104376510A (en) * | 2014-12-05 | 2015-02-25 | 国家电网公司 | Method of predicting and accessing level of wildfire-caused trip risk in power transmission lines |
| CN106980929A (en) * | 2017-01-05 | 2017-07-25 | 国网福建省电力有限公司 | A kind of power failure complaint risk Forecasting Methodology based on random forest |
| CN107506849A (en) * | 2017-07-24 | 2017-12-22 | 国网江西省电力公司电力科学研究院 | A kind of intelligent optimization distribution transforming, which has a power failure, studies and judges system |
Non-Patent Citations (1)
| Title |
|---|
| 王乐: "厦门配电网可靠性提升措施的研究与应用", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110135614A (en) * | 2019-03-26 | 2019-08-16 | 广东工业大学 | It is a kind of to be tripped prediction technique based on rejecting outliers and the 10kV distribution low-voltage of sampling techniques |
| CN110033140A (en) * | 2019-04-22 | 2019-07-19 | 广东工业大学 | A kind of distribute-electricity transformer district tripping prediction technique, system and device |
| CN111382897A (en) * | 2019-10-25 | 2020-07-07 | 广州供电局有限公司 | Transformer area low-voltage trip prediction method and device, computer equipment and storage medium |
| CN111178957A (en) * | 2019-12-23 | 2020-05-19 | 广西电网有限责任公司 | Method for early warning sudden increase of electric quantity of electricity consumption customer |
| CN111178957B (en) * | 2019-12-23 | 2023-04-14 | 广西电网有限责任公司 | Method for early warning sudden increase of electric quantity of electricity consumption customer |
| CN115879586A (en) * | 2022-01-11 | 2023-03-31 | 北京中关村科金技术有限公司 | Complaint prediction optimization method and device based on ablation experiment and storage medium |
| CN115879586B (en) * | 2022-01-11 | 2024-01-02 | 北京中关村科金技术有限公司 | Complaint prediction optimization method and device based on ablation experiment and storage medium |
| CN115456210A (en) * | 2022-08-22 | 2022-12-09 | 国网浙江省电力有限公司杭州市临安区供电公司 | Power utilization complaint early warning method based on cascade logistic regression Bayesian algorithm |
| CN115456210B (en) * | 2022-08-22 | 2024-04-12 | 国网浙江省电力有限公司杭州市临安区供电公司 | Power consumption complaint early warning method based on cascading logistic regression Bayesian algorithm |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109410089B (en) | 2020-11-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109410089A (en) | Low-voltage tripping and customer complaint prediction technique, device and storage medium | |
| Gamarra et al. | Computational optimization techniques applied to microgrids planning: A review | |
| Huang et al. | Evaluation of AMI and SCADA data synergy for distribution feeder modeling | |
| Xiang et al. | Economic planning of electric vehicle charging stations considering traffic constraints and load profile templates | |
| Tascikaraoglu et al. | Short-term residential electric load forecasting: A compressive spatio-temporal approach | |
| Dong et al. | Emerging techniques in power system analysis | |
| Billinton et al. | Unit commitment risk analysis of wind integrated power systems | |
| ES2451368T3 (en) | Improvement of distribution grid for plug-in electric vehicles | |
| US20220260619A1 (en) | Abnormal electricity use recognition method and device, and computer-readable medium | |
| Taylor et al. | Distribution modeling requirements for integration of PV, PEV, and storage in a smart grid environment | |
| Nguyen et al. | Optimal number, location, and size of distributed generators in distribution systems by symbiotic organism search based method | |
| CN102722764A (en) | Integrated power grid optimization auxiliary decision analysis system | |
| Alharbi et al. | Electric vehicle charging facility as a smart energy microhub | |
| Wang et al. | Long‐Term Maintenance Scheduling of Smart Distribution System through a PSO‐TS Algorithm | |
| Li et al. | Resilient outage recovery of a distribution system: Co-optimizing mobile power sources with network structure | |
| Liu et al. | Does environmental heterogeneity affect the productive efficiency of grid utilities in China? | |
| Rylander et al. | Application of new method for distribution-wide assessment of Distributed Energy Resources | |
| Govindarajan et al. | Renewable energy for electricity use in India: evidence from India’s smart cities mission | |
| Mazza et al. | Determination of the relevant periods for intraday distribution system minimum loss reconfiguration | |
| De Filippo et al. | Robust optimization for virtual power plants | |
| Li et al. | Artificial Intelligence Enabled Computational Methods for Smart Grid Forecast and Dispatch | |
| Yan et al. | Active distribution system synthesis via unbalanced graph generative adversarial network | |
| Pappas et al. | Adaptive load forecasting of the Hellenic electric grid | |
| Safdar et al. | Electricity Cost Prediction using Autoregressive Integrated Moving Average (ARIMA) in Korea | |
| Li et al. | Learning to bundle proactively for on-demand meal delivery |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20200924 Address after: 510620 Tianhe District, Guangzhou, Tianhe South Road, No. two, No. 2, No. Applicant after: Guangzhou Power Supply Bureau of Guangdong Power Grid Co.,Ltd. Address before: 510620 Tianhe District, Guangzhou, Tianhe South Road, No. two, No. 2, No. Applicant before: GUANGZHOU POWER SUPPLY Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |