CN109754107A - Prediction technique and prediction meanss - Google Patents
Prediction technique and prediction meanss Download PDFInfo
- Publication number
- CN109754107A CN109754107A CN201711089197.0A CN201711089197A CN109754107A CN 109754107 A CN109754107 A CN 109754107A CN 201711089197 A CN201711089197 A CN 201711089197A CN 109754107 A CN109754107 A CN 109754107A
- Authority
- CN
- China
- Prior art keywords
- prediction
- prediction error
- error sequence
- value
- sequence
- 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
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000012417 linear regression Methods 0.000 claims description 28
- 238000012545 processing Methods 0.000 claims description 24
- 238000003860 storage Methods 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 7
- 238000003066 decision tree Methods 0.000 claims description 6
- 238000003062 neural network model Methods 0.000 claims description 3
- 230000008901 benefit Effects 0.000 claims description 2
- 238000007637 random forest analysis Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 18
- 230000006870 function Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present disclosure discloses a kind of prediction technique and prediction meanss, are related to computer field.Method therein includes: to determine prediction error sequence according to the predicted value and true value of prediction model；Exceptional value in removal prediction error sequence；Calculate the mean value of remaining prediction error sequence；The value to be predicted of prediction model is compensated using the mean value of remaining prediction error sequence.To using historical forecast error information, treat predicted value and compensate, improve the accuracy of prediction.
Description
Technical field
This disclosure relates to computer field, in particular to a kind of prediction technique and prediction meanss.
Background technique
With the development of science and technology and the increase of forecast demand, more and more prediction models are suggested and apply.
The target of prediction model is that predicted value is equal to actual value as far as possible.However, the situation of reality is that predicted value is frequent with actual value
It is unequal.The difference of predicted value and actual value is exactly the prediction error of these prediction models.
Summary of the invention
Inventors have found that due to the randomness of reality, alternatively, since the feature considered inside different prediction models is each
It is different, cause always to have some features not found the reasons such as not to be considered into, causes prediction error occur, influence the standard of prediction
True property.
An embodiment of the present disclosure technical problem to be solved is: improving the accuracy of prediction.
According to one aspect of the disclosure, a kind of prediction technique is proposed, comprising: according to the predicted value of prediction model and really
Value determines prediction error sequence；Exceptional value in removal prediction error sequence；Calculate the mean value of remaining prediction error sequence；Benefit
The value to be predicted of prediction model is compensated with the mean value of remaining prediction error sequence.
Optionally, the exceptional value in the removal prediction error sequence includes: to calculate the mean value and mark of prediction error sequence
It is quasi- poor；The zone of reasonableness of prediction error is determined according to the mean value of prediction error sequence and standard deviation；In removal prediction error sequence
Prediction error except zone of reasonableness.
Optionally, the zone of reasonableness for predicting error is to predict the mean value of error sequence plus or minus according to prediction error sequence
Floating range determined by the standard deviation of column.
Optionally, the exceptional value in the removal prediction error sequence includes: to determine linear return using prediction error sequence
Return model；Deviate the prediction error that linear regression model (LRM) is more than preset range in removal prediction error sequence.
Optionally, the mean value that the value to be predicted of prediction model subtracts remaining prediction error sequence is compensated as prediction model
Value to be predicted afterwards.
Optionally, prediction model for example, neural network model, moving average model(MA model), gradient promote decision tree, random
Forest decision tree etc., but it is not limited to examples cited.
According to another aspect of the disclosure, a kind of prediction meanss are proposed, comprising: error sequence module, for according to pre-
The predicted value and true value for surveying model determine prediction error sequence；Series processing module, for removing in prediction error sequence
Exceptional value；Compensating module, for calculating the mean value of remaining prediction error sequence；Utilize the mean value of remaining prediction error sequence
The value to be predicted of prediction model is compensated.
Optionally, the series processing module includes: First ray processing unit or the second series processing unit；
The First ray processing unit, for calculating the mean value and standard deviation of prediction error sequence；According to prediction error
The mean value and standard deviation of sequence determine the zone of reasonableness of prediction error；It is pre- except zone of reasonableness in removal prediction error sequence
Survey error；
The second series processing unit, for determining linear regression model (LRM) using prediction error sequence；Removal prediction misses
Deviate the prediction error that linear regression model (LRM) is more than preset range in difference sequence.
Optionally, the zone of reasonableness of the prediction error in the First ray processing unit is the mean value for predicting error sequence
Plus or minus the floating range according to determined by the standard deviation of prediction error sequence.
Optionally, the compensating module, for the value to be predicted of prediction model to be subtracted remaining prediction error sequence
Mean value is as the compensated value to be predicted of prediction model.
According to another aspect of the present disclosure, a kind of prediction meanss are proposed, comprising: memory；And it is coupled to the storage
The processor of device, the processor is configured to the instruction based on storage in the memory, executes prediction technique above-mentioned.
According to the another aspect of the disclosure, proposes a kind of computer readable storage medium, is stored thereon with computer program,
The step of program realizes prediction technique above-mentioned when being executed by processor.
The disclosure utilizes historical forecast error information, treats predicted value and compensates, improves the accuracy of prediction.
Detailed description of the invention
Attached drawing needed in embodiment or description of Related Art will be briefly described below.According to following ginseng
According to the detailed description of attached drawing, the disclosure can be more clearly understood,
It should be evident that the accompanying drawings in the following description is only some embodiments of the present disclosure, skill common for this field
For art personnel, without any creative labor, it is also possible to obtain other drawings based on these drawings.
Fig. 1 is the flow diagram of disclosure prediction technique one embodiment.
Fig. 2 is the flow diagram of one embodiment of the exceptional value in disclosure removal prediction error sequence.
Fig. 3 is the flow diagram of the further embodiment of the exceptional value in disclosure removal prediction error sequence.
Fig. 4 is the structural schematic diagram of disclosure prediction meanss one embodiment.
Fig. 5 is the structural schematic diagram of another embodiment of disclosure prediction meanss.
Fig. 6 is the structural schematic diagram of disclosure prediction meanss further embodiment.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present disclosure, the technical solution in the embodiment of the present disclosure is carried out clear, complete
Site preparation description.
In order to improve the accuracy of prediction, the disclosure is proposed.
Fig. 1 is the flow diagram of disclosure prediction technique one embodiment.
As shown in Figure 1, the prediction technique 10 of the present embodiment includes:
Step 110, prediction error sequence is determined according to the predicted value of prediction model and true value.
Wherein, prediction model for example, neural network model, moving average model(MA model), gradient promote decision tree, random gloomy
Woods decision tree etc., but it is not limited to examples cited.Under normal conditions, prediction model first passes through training, is then used to predict again.In advance
The training and prediction for surveying model can refer to the prior art.
Wherein, the difference of predicted value and actual value is exactly the prediction error of these prediction models.Multiple groups predicted value and reality
The sequence of differences of value is exactly the prediction error sequence of these prediction models.Prediction error sequence usually can be historical forecast error
Data determine historical forecast error sequence according to the historical forecast value of prediction model and history true value.
Step 120, the exceptional value in removal prediction error sequence.
In prediction error sequence, different a small number of prediction errors are showed from most prediction error, can be considered as pre-
Survey the exceptional value in error sequence.Therefore, a zone of reasonableness of prediction error is determined according to most normal prediction error,
Prediction error except zone of reasonableness is determined as exceptional value, and is removed.
Step 130, the mean value of remaining prediction error sequence is calculated.
Assuming that the sequence ε of n prediction error1,ε2,...,εnRemove remaining m prediction error ε after exceptional value '1,ε
'2,...,ε'm.According to central-limit theorem, when enough with the sample of remaining prediction error sequence, remaining prediction is missed
Difference sequence is Normal Distribution, can refer to subsequent formula 1 and formula 2, calculates the equal of remaining prediction error sequence
Value (being set as μ ') and standard deviation (being set as σ ').
Step 140, the value to be predicted of prediction model is compensated using the mean value of remaining prediction error sequence.
In one embodiment, the value to be predicted of prediction model subtracts the mean value of remaining prediction error sequence as prediction
Value to be predicted after model compensation.
The present embodiment utilizes historical forecast error information, treats predicted value and compensates, improves the accuracy of prediction.
The method that the disclosure also proposes the exceptional value in two kinds of illustrative removal prediction error sequences.
Fig. 2 is the flow diagram of one embodiment of the exceptional value in disclosure removal prediction error sequence.
As shown in Fig. 2, the method 20 of the exceptional value in embodiment removal prediction error sequence includes:
Step 210, it is assumed that obtain the sequence ε of n prediction error1,ε2,...,εn, according to central-limit theorem, every time
Prediction error be independent of each other stochastic variable, with prediction error sequence sample it is enough when, predict error sequence
It is Normal Distribution, the mean value (being set as μ) and standard deviation (being set as σ) of prediction error sequence can be calculated accordingly.
In addition, according to law of great number, when enough with the sample of prediction error sequence, the close totality of the mean value of sample
Mean value can represent the mean value of prediction error with the mean value of prediction error sequence accordingly.
Step 220, the zone of reasonableness of prediction error is determined according to the mean value of prediction error sequence and standard deviation.
Wherein, the zone of reasonableness for predicting error is to predict the mean value of error sequence plus or minus according to prediction error sequence
Standard deviation determined by floating range.Formula is expressed as follows:
μ-p × σ < ε < μ+p × σ (formula 3)
Wherein, p × σ indicates the floating range according to determined by the standard deviation of prediction error sequence, and p is float factor,
Numerical value is, for example, 0.5,1,1.5,2,3 etc., but is first not limited to examples cited, and p is bigger, predicts that the zone of reasonableness of error is bigger, goes
The prediction error removed is fewer, can need to be arranged the value of p according to business.For example, about removing 35% prediction error, about when p=1
65% prediction error is retained.
Step 230, the prediction error in removal prediction error sequence except zone of reasonableness.
For example, ε1,ε2,...,εnIn prediction error or not the zone of reasonableness shown in formula 3 will be removed.
To eliminate the exceptional value in prediction error sequence according to central-limit theorem and law of great number.
Fig. 3 is the flow diagram of the further embodiment of the exceptional value in disclosure removal prediction error sequence.
As shown in figure 3, the method 30 of the exceptional value in embodiment removal prediction error sequence includes:
Step 310, linear regression model (LRM) is determined using prediction error sequence.
For example, using prediction error sequence, using prediction error sequence, can attempt to use preceding n prediction error as
Variable x1,x2,...,xn, the latter prediction error is y=w as target value y, i.e. equation of linear regression1×x1+w2×x2+
...wn×xn+ b, wherein w1,w2,...,wn, b is known variables, removes linear regression equation using least square method, is obtained
The value of known variables, to obtain equation of linear regression (i.e. linear regression model (LRM)).
Step 320, deviate the prediction error that linear regression model (LRM) is more than preset range in removal prediction error sequence.
To eliminate the exceptional value in prediction error sequence using linear regression method.
Fig. 4 is the structural schematic diagram of disclosure prediction meanss one embodiment.
As shown in figure 4, the prediction meanss 40 of the present embodiment include:
Error sequence module 410, for determining prediction error sequence according to the predicted value and true value of prediction model.
Series processing module 420, for removing the exceptional value in prediction error sequence.
Compensating module 430 utilizes remaining prediction error sequence for calculating the mean value of remaining prediction error sequence
Mean value compensates the value to be predicted of prediction model.
Wherein, compensating module 430, the mean value for the predicted value of prediction model to be subtracted remaining prediction error sequence are made
For the compensated predicted value of prediction model.
The present embodiment utilizes historical forecast error information, treats predicted value and compensates, improves the accuracy of prediction.
Fig. 5 is the structural schematic diagram of another embodiment of disclosure prediction meanss.
As shown in figure 5, the series processing module 420 in the prediction meanss 50 of the present embodiment includes: that First ray processing is single
Member 421 or the second series processing unit 422.
First ray processing unit 421, for calculating the mean value and standard deviation of prediction error sequence.According to prediction error sequence
The mean value and standard deviation of column determine the zone of reasonableness of prediction error, and the prediction in error sequence except zone of reasonableness is predicted in removal
Error.
Wherein, the zone of reasonableness for predicting error is to predict the mean value of error sequence plus or minus according to prediction error sequence
Standard deviation determined by floating range.
Second series processing unit 422, for determining linear regression model (LRM) using prediction error sequence.Removal prediction error
Deviate the prediction error that linear regression model (LRM) is more than preset range in sequence.
The method that the present embodiment proposes the exceptional value in two kinds of illustrative removal prediction error sequences.
Fig. 6 is the structural schematic diagram of disclosure prediction meanss further embodiment.
As shown in fig. 6, the prediction meanss 60 of the present embodiment include: memory 610 and the place for being coupled to the memory 610
Device 620 is managed, processor 620 is configured as executing in any one aforementioned embodiment based on the instruction being stored in memory 610
Prediction technique.
Wherein, memory 610 is such as may include system storage, fixed non-volatile memory medium.System storage
Device is for example stored with operating system, application program, Boot loader (Boot Loader) and other programs etc..
Device 60 can also include input/output interface 630, network interface 640, memory interface 650 etc..These interfaces
It can for example be connected by bus 660 between 630,640,650 and memory 610 and processor 620.Wherein, input and output
The input-output equipment such as interface 630 is display, mouse, keyboard, touch screen provide connecting interface.Network interface 640 is various
Networked devices provide connecting interface.The external storages such as memory interface 650 is SD card, USB flash disk provide connecting interface.
According to the another aspect of the disclosure, proposes a kind of computer readable storage medium, is stored thereon with computer program,
The step of program realizes prediction technique above-mentioned when being executed by processor.
Those skilled in the art should be understood that embodiment of the disclosure can provide as method, system or computer journey
Sequence product.Therefore, complete hardware embodiment, complete software embodiment or combining software and hardware aspects can be used in the disclosure
The form of embodiment.Moreover, it wherein includes the calculating of computer usable program code that the disclosure, which can be used in one or more,
Machine can use the meter implemented in non-transient storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.)
The form of calculation machine program product.
The disclosure is referring to method, the process of equipment (system) and computer program product according to the embodiment of the present disclosure
Figure and/or block diagram describe.It is interpreted as to be realized by computer program instructions each in flowchart and/or the block diagram
The combination of process and/or box in process and/or box and flowchart and/or the block diagram.It can provide these computer journeys
Sequence instruct to general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices processor with
A machine is generated, so that the instruction generation executed by computer or the processor of other programmable data processing devices is used for
Realize the dress for the function of specifying in one or more flows of the flowchart and/or one or more blocks of the block diagram
It sets.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
The foregoing is merely the preferred embodiments of the disclosure, not to limit the disclosure, all spirit in the disclosure and
Within principle, any modification, equivalent replacement, improvement and so on be should be included within the protection scope of the disclosure.
Claims (12)
1. a kind of prediction technique, comprising:
Prediction error sequence is determined according to the predicted value of prediction model and true value；
Exceptional value in removal prediction error sequence；
Calculate the mean value of remaining prediction error sequence；
The value to be predicted of prediction model is compensated using the mean value of remaining prediction error sequence.
2. the method as described in claim 1, the removal predicts that the exceptional value in error sequence includes:
Calculate the mean value and standard deviation of prediction error sequence；
The zone of reasonableness of prediction error is determined according to the mean value of prediction error sequence and standard deviation；
Prediction error in removal prediction error sequence except zone of reasonableness.
3. method according to claim 2, wherein predict error zone of reasonableness be predict error sequence mean value add or
Subtract the floating range according to determined by the standard deviation of prediction error sequence.
4. the method as described in claim 1, the removal predicts that the exceptional value in error sequence includes:
Linear regression model (LRM) is determined using prediction error sequence；
Deviate the prediction error that linear regression model (LRM) is more than preset range in removal prediction error sequence.
5. the method for claim 1, wherein the value to be predicted of prediction model subtracts the equal of remaining prediction error sequence
Value is used as the compensated value to be predicted of prediction model.
6. the method for claim 1, wherein prediction model includes: neural network model, moving average model(MA model), gradient
Promote decision tree, random forest decision tree.
7. a kind of prediction meanss, comprising:
Error sequence module, for determining prediction error sequence according to the predicted value and true value of prediction model；
Series processing module, for removing the exceptional value in prediction error sequence；
Compensating module, for calculating the mean value of remaining prediction error sequence；Utilize the mean value pair of remaining prediction error sequence
The value to be predicted of prediction model compensates.
8. device as claimed in claim 7, the series processing module includes: at First ray processing unit or the second sequence
Manage unit；
The First ray processing unit, for calculating the mean value and standard deviation of prediction error sequence；According to prediction error sequence
Mean value and standard deviation determine prediction error zone of reasonableness；Prediction in removal prediction error sequence except zone of reasonableness misses
Difference；
The second series processing unit, for determining linear regression model (LRM) using prediction error sequence；Removal prediction error sequence
Deviate the prediction error that linear regression model (LRM) is more than preset range in column.
9. device as claimed in claim 8, wherein the zone of reasonableness of the prediction error in the First ray processing unit is
The mean value plus or minus the floating range according to determined by the standard deviation of prediction error sequence for predicting error sequence.
10. device as claimed in claim 7, wherein the compensating module, it is surplus for subtracting the value to be predicted of prediction model
The mean value of remaining prediction error sequence is as the compensated value to be predicted of prediction model.
11. a kind of prediction meanss, comprising:
Memory；And
It is coupled to the processor of the memory, the processor is configured to the instruction based on storage in the memory,
Execute such as prediction technique of any of claims 1-6.
12. a kind of computer readable storage medium, is stored thereon with computer program, power is realized when which is executed by processor
Benefit requires the step of prediction technique described in any one of 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711089197.0A CN109754107A (en) | 2017-11-08 | 2017-11-08 | Prediction technique and prediction meanss |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711089197.0A CN109754107A (en) | 2017-11-08 | 2017-11-08 | Prediction technique and prediction meanss |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109754107A true CN109754107A (en) | 2019-05-14 |
Family
ID=66400250
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711089197.0A Pending CN109754107A (en) | 2017-11-08 | 2017-11-08 | Prediction technique and prediction meanss |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109754107A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110533182A (en) * | 2019-08-01 | 2019-12-03 | 北京三快在线科技有限公司 | A kind of data processing method and device |
| CN111966226A (en) * | 2020-09-03 | 2020-11-20 | 福州大学 | A fault-tolerant method and system for tactile communication based on compensatory long short-term memory network |
-
2017
- 2017-11-08 CN CN201711089197.0A patent/CN109754107A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110533182A (en) * | 2019-08-01 | 2019-12-03 | 北京三快在线科技有限公司 | A kind of data processing method and device |
| CN111966226A (en) * | 2020-09-03 | 2020-11-20 | 福州大学 | A fault-tolerant method and system for tactile communication based on compensatory long short-term memory network |
| CN111966226B (en) * | 2020-09-03 | 2022-05-10 | 福州大学 | Touch communication fault-tolerant method and system based on compensation type long-term and short-term memory network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chang et al. | Integrating berth allocation and quay crane assignments | |
| CN104462755B (en) | Electronic equipment spare parts configuration computational methods based on reliability model | |
| US20190039243A1 (en) | Simulation device and simulation method for robot system | |
| CN103914433B (en) | For the system and method for factorization square matrix again on parallel processor | |
| CN109754107A (en) | Prediction technique and prediction meanss | |
| CN110134598A (en) | A kind of batch processing method, apparatus and system | |
| US20160154874A1 (en) | Method for determining condition of category division of key performance indicator, and computer and computer program therefor | |
| Annighöfer et al. | Automated selection, sizing, and mapping of integrated modular avionics modules | |
| KR101698335B1 (en) | Program and method for movement flow simulation about nuclear fuel | |
| Anand et al. | Resource allocation problem for multi versions of software system | |
| US20180046959A1 (en) | Similar project identification | |
| US20170024708A1 (en) | Product disassembling method with disassembling sequence optimization and non-transitory computer readable medium thereof | |
| CA2962139A1 (en) | Parallel solution generation | |
| US10929177B2 (en) | Managing resources for multiple trial distributed processing tasks | |
| JP2016045692A (en) | Apparatus and program for estimating the number of bugs | |
| CN104572134B (en) | A kind of optimization method and device | |
| JP7328126B2 (en) | Production simulation device and production simulation method | |
| JP7103908B2 (en) | Elevator fit judgment device, fit judgment method, and fit judgment program | |
| Rodovalho et al. | Estimate of hourly productivity applied to elaboration and implementation of mining plans | |
| Natu et al. | EasyDist: An End-to-End distributed deep learning tool for cloud | |
| Underwood et al. | Comparing lifecycle sustainment strategies in an electronic component obsolescence environment | |
| CN114816758B (en) | Resource allocation method and device | |
| CN116954652A (en) | Method and device for predicting update time of open source component and electronic equipment | |
| Velampudi et al. | K OUT OF N INTEGRATED RELIABILITY MODEL WITH MULTIPLE STAGE OPTIMIZATION USING FOUR COMPONENTS OF HEURISTIC PROGRAMMING APPROACH. | |
| Lorenzo et al. | Study of performance issues on a SMP-NUMA system using the roofline model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190514 |
|
| RJ01 | Rejection of invention patent application after publication |