CN109034466A - A kind of laying rate of laying hen prediction technique based on Support vector regression - Google Patents
A kind of laying rate of laying hen prediction technique based on Support vector regression Download PDFInfo
- Publication number
- CN109034466A CN109034466A CN201810778061.9A CN201810778061A CN109034466A CN 109034466 A CN109034466 A CN 109034466A CN 201810778061 A CN201810778061 A CN 201810778061A CN 109034466 A CN109034466 A CN 109034466A
- Authority
- CN
- China
- Prior art keywords
- laying
- laying hen
- rate
- support vector
- laying rate
- 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 45
- 235000013601 eggs Nutrition 0.000 claims description 26
- 238000012549 training Methods 0.000 claims description 22
- 241000287828 Gallus gallus Species 0.000 claims description 19
- 238000012360 testing method Methods 0.000 claims description 13
- 235000021050 feed intake Nutrition 0.000 claims description 9
- 238000002790 cross-validation Methods 0.000 claims description 8
- 238000012706 support-vector machine Methods 0.000 claims description 6
- 238000005286 illumination Methods 0.000 claims description 5
- 241001269238 Data Species 0.000 claims description 4
- 238000004519 manufacturing process Methods 0.000 claims description 4
- 235000015097 nutrients Nutrition 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 3
- 230000003442 weekly effect Effects 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 2
- 230000001737 promoting effect Effects 0.000 abstract description 4
- 230000006870 function Effects 0.000 description 20
- 238000013528 artificial neural network Methods 0.000 description 9
- 238000009395 breeding Methods 0.000 description 6
- 230000001488 breeding effect Effects 0.000 description 6
- 238000010801 machine learning Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 2
- 230000036528 appetite Effects 0.000 description 2
- 235000019789 appetite Nutrition 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000037396 body weight Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000001351 cycling effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000001976 improved effect Effects 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000001373 regressive effect Effects 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 230000008093 supporting effect Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Animal Husbandry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Agronomy & Crop Science (AREA)
- Mining & Mineral Resources (AREA)
- General Health & Medical Sciences (AREA)
- Marine Sciences & Fisheries (AREA)
- Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The laying rate of laying hen prediction technique based on Support vector regression that present invention mainly discloses a kind of, comprising the following steps: 1), acquisition laying hen history laying rate and history influence factor data；2), establish prediction model: the prediction model based on time series will affect factor as input sample and establish the laying rate of laying hen prediction model based on Support vector regression using laying rate as output sample；3) it, using the laying rate of laying hen prediction model based on Support vector regression of foundation, is predicted according to following laying rate of the influence factor to laying hen, obtains the change in future of the laying rate index of laying hen.Laying rate of laying hen prediction model based on Support vector regression of the invention has good estimated performance, and high stability, has confidence level and promotional value.
Description
Technical field
The present invention relates to breeding layer chicken technical field, especially a kind of laying rate of laying hen based on Support vector regression is pre-
Survey method.
Background technique
Egg is indispensable using article in daily life, and our daily eaten eggs are mainly from each egg
Chicken breeding enterprise.It is studied according to domestic and foreign literature, the daily laying rate of chicken is not fixation, and laying rate of laying hen is one and is given birth to
Object, chemistry, physics and artificial many-sided complication system influenced, laying hen feed intake, the variation of external environment and chicken itself
The change of weight etc. factor can all cause very big influence to laying rate of laying hen, and laying rate can be with laying hen feed intake, laying hen
Living environment, own body weight etc. influences laying rate factor and generates nonlinear variation.Since the influence factor of laying rate is many
It is more, and laying rate is nonlinear change, is difficult to find out an accurate prediction model in a conventional manner to predict laying hen production
Egg rate.In order to maintain and improve the benefit of enterprise, comprehensive many factors, establishing a kind of laying rate of laying hen prediction model is very
It is necessary to.First, laying hen prediction model can understand the suitable environment of breeding layer chicken, there is positive side to breeding layer chicken
It helps.Second, establishing laying hen prediction model can be improved the benefit of enterprise, this is also asking of being concerned about the most of each breeding layer chicken enterprise
Topic.And there is presently no an accurate prediction models to realize the prediction to laying rate of laying hen.
Machine learning (Machine Learning, ML) is one of the research field of modern intelligence computation radix, it is main
It is that the mankind sum up the study inducing ability come to the simulation of the Nature biobehavioral, its objective is by excavating observation
Regular nature in data, and the prediction using this regular characteristic spread to unknown data.The essence of machine learning is exactly logical
The model that computer establishes a hypothesis to some problem is crossed, true model is approached using the model of hypothesis.It supports
A kind of method of the vector machine (Support vector machine, SVM) as machine learning, regression model are built upon system
Count the supervised learning method on the basis of model theory and structure minimization principle.This method, will be original by introducing kernel function
The features such as input is mapped to the feature space that High-dimensional Linear can divide, and has bustling ability strong, is less prone to over-fitting, can be well
The problems such as solving non-linear small sample, various dimensions and local minimum point.
Summary of the invention
In view of the deficienciess of the prior art, to provide a kind of laying rate of laying hen based on Support vector regression pre- by the present invention
Survey method has been put forward for the first time the laying rate of laying hen prediction model based on Support vector regression, has good to the laying rate of laying hen
Good estimated performance, and high stability have confidence level and promotional value.
In order to achieve the above object, the present invention is achieved through the following technical solutions: one kind being based on Support vector regression
Laying rate of laying hen prediction technique, comprising the following steps:
1) laying rate of the history of laying hen and the influence factor data of history, are acquired；
2), establish prediction model: the prediction model based on time series will affect factor as input sample, will lay eggs
Rate establishes the laying rate of laying hen prediction model based on Support vector regression as output sample；
3), using the laying rate of laying hen prediction model based on Support vector regression of foundation, according to influence factor to egg
Following laying rate of chicken is predicted, the change in future of the laying rate index of laying hen is obtained.
The present invention is further arranged to: in the step 1), influence factor include laying hen feed intake, laying hen chicken age, weight,
Temperature, light irradiation time, intensity of illumination and whether take nutrient.
The present invention is further arranged to: in the step 1), the acquisition method of laying rate is always to lay eggs so that laying hen is daily
Amount obtains the daily laying rate of every laying hen divided by the sum of laying hen.
The present invention is further arranged to: the acquisition method of the laying hen feed intake is to be removed with total collection capacity that laying hen is daily
With the sum of laying hen, the daily laying hen feed intake of every laying hen is obtained.
The present invention is further arranged to: the acquisition method of the weight is to acquire the weight of a laying hen weekly, is used
Lagrange interpolation method calculates the daily weight of laying hen.
The present invention is further arranged to: the acquisition method of the temperature is, the maximum temperature of acquisition same day henhouse and minimum
Temperature.
The present invention is further arranged to: the step 2) is specifically, establish prediction model: choosing two parts of historical datas, often
Part historical data includes influence factor and laying rate, and the prediction model based on time series chooses gaussian kernel function, to influence
Factor is as input sample, and using laying rate as output sample, using a copy of it historical data as training sample, training is based on
The laying rate of laying hen prediction model of Support vector regression；Using another historical data as test sample, test is based on support
The laying rate of laying hen prediction model that vector machine returns.
The present invention is further arranged to: the step 2) is specifically, establish prediction model: set the historical data of laying rate asWherein N is data length, YnThe laying rate laid eggs in other words n-th day for n-th group；If the historical data of influence factor
ForxnFor the vector of the influence factor data composition of seven groups of laying rate,It is produced for i-th kind
The influence factor n-th group of data of egg rate, building time series are Xn=[xn,xn+1,...,xn+d], XnFor xnTo xn+dThe square of composition
Battle array, d are the width of data set, determine mapping relations:
Yn=F (Xn), n=1,2 ..., N (1)；
Training sample is constructed,Wherein XnFor input sample, YnTo export sample, gaussian kernel function, choosing are chosen
Two parts of historical datas are taken, it is sample by the historical data of a copy of it that every part of historical data, which includes influence factor and laying rate,
Collect training SVM, and obtain the parameter combination of optimal penalty factor and kernel function width cs, substitutes into support vector machines and instructed
Practice, using the approximation capability of SVM, so that the input and output of SVM are approximant (1), to establish the laying hen of Support vector regression
Laying rate prediction model；The influence factor in another historical data is substituted into established SVM later, as test sample,
The result predicted.
The present invention is further arranged to: the method for obtaining the parameter combination of optimal penalty factor and kernel function width cs,
Be using grid data service and cross-validation method to penalty factor and kernel function width cs parameter combination performance carry out evaluation to
Select optimal (C, σ) parameter combination, the specific steps are as follows:
Step 1: initialization (C, σ) range of choice, subset number, optimizing termination condition and grid search step-length；
Step 2: reference SVM learns sample in all (C, σ) is combined, using cross validation mean square deviation (MSE)
It indicates:
In formula (2), yiFor training set output valve, i.e., i-th point of actual temperature error or error rate；y′iIt is i-th
The predicted value of a point；
Step 3: comparing MSE, choose the smallest parameter combination of MSE as optimized parameter, for identical MSE, due to mistake
The big state that will lead to overfitting then chooses the lesser parameter combination of C；
Step 4: constantly reducing search range and step-size in search according to step 3, repeat step 2, until MSE or search
Step-length reaches setting value, and last (C, σ) parameter combination is best optimizing (C, σ) parameter combination.
The present invention is further arranged to: further including that assessment is returned based on support vector machines between the step 2) and step 3)
The performance for the laying rate of laying hen prediction model returned.
The present invention has the beneficial effect that
The present invention is based on Support vector regression principle, will affect laying hen by the machine learning quickly grown in recent years
Input sample of the influence factor of laying rate as SVM, laying rate of laying hen data are established for the first time and are proposed as output sample
Laying rate of laying hen prediction model based on Support vector regression, and laying rate is predicted by the model.SVM is in laying hen
Performance is better than BP neural network on laying rate prediction model, and SVM prediction model result is relatively stable, and divides from the angle of popularization
Analysis, the prediction model time-consuming based on SVM will be much smaller than the time based on BP neural network.To sum up, it is based on Support vector regression
Laying rate of laying hen prediction there is good estimated performance, and high stability has confidence level and promotional value.
Detailed description of the invention
Fig. 1 is a kind of laying rate of laying hen prediction model based on Support vector regression of the invention；
Fig. 2 is the prediction result of the invention based on Support vector regression.
Specific embodiment
In conjunction with attached drawing, present pre-ferred embodiments are described in further details.
A kind of laying rate of laying hen prediction technique based on Support vector regression, comprising the following steps:
1) laying rate of the history of laying hen and the influence factor data of history, are acquired；The influence factor packet of laying rate of laying hen
It includes laying hen feed intake, laying hen chicken age (day), weight, temperature, light irradiation time, intensity of illumination and whether takes nutrient.Application
People cooperates with Bozhou City Hai Lanhe breeding layer chicken company of Anhui Province, lays eggs since laying hen, during laying hen is sold, by egg
The laying rate and external influence factor of chicken are recorded, are analyzed.
Since the number of chicken is different, and during investigation, laying hen is there are natural death, therefore the acquisition of laying rate
Method is, with the daily total egg production of laying hen divided by the sum of laying hen, to obtain the daily laying rate of every laying hen.The laying hen is adopted
The acquisition method of appetite is, with the daily total collection capacity of laying hen divided by the sum of laying hen, to obtain the daily laying hen of every laying hen and adopt
Appetite.Since condition limits, the acquisition method of the weight is to acquire the weight of a laying hen weekly, is inserted using Lagrange
Value method calculates the daily weight of laying hen.Lagrange interpolation method is a kind of polynomial interpolation, can provide one and wear just
The polynomial function for crossing known point on two-dimensional surface, according to this multinomial, it is estimated that the weight that laying hen is daily.Due to chicken
House illumination condition remains unchanged, and the application ignores the factor of this influence laying rate of intensity of illumination.When using nutrient solution, by drug
It is mixed in water, can guarantee every laying hen all with taking.Due to environment and equipment limit, the more difficult reality of mean temperature on the same day is measured
It is existing, therefore the acquisition method of temperature is, the maximum temperature and minimum temperature of acquisition same day henhouse replace mean temperature, can satisfy
The relevance of temperature and egg production.It then can accurately monitor, record as light irradiation time.So far, institute's number in need has been handled
According to.
2), establish prediction model: the prediction model based on time series will affect factor as input sample, will lay eggs
Rate establishes the laying rate of laying hen prediction model based on Support vector regression as output sample；
It establishes prediction model: choosing two parts of historical datas, every part of historical data includes influence factor and laying rate, is based on
The prediction model of time series chooses gaussian kernel function, using influence factor as input sample, using laying rate as output sample
This, using a copy of it historical data as training sample, laying rate of laying hen prediction model of the training based on Support vector regression；
Using another historical data as test sample, the laying rate of laying hen prediction model based on Support vector regression is tested.
Specifically, the blue brown laying hen in selection two batches sea of the same race, record sell period from laying eggs first day to the 378th day, lay eggs
The influence factor data and laying rate data of rate.Prediction model based on time series, by one group of laying rate of first chicken
Influence factor and laying rate data input respectively as the training of SVM, export sample, and the laying hen that Training Support Vector Machines return produces
Egg rate prediction model；Then SVM is inputted using the influence factor data of one group of laying rate of second batch laying hen as test sample, obtained
The rate score of laying eggs predicted out is finally compared with the practical laying rate of second batch laying hen.
SVM recurrence is related to the selection of kernel function type, and different kernel functions can generate different influences to prediction model.
Gaussian kernel function (RBF kernel function) is chosen in the application to predict laying rate model.RBF kernel function is relative to other core letters
For number, have number of parameters few, the advantages such as numerical value restrictive condition is few and performance is good.
The parameter of SVM determines its learning ability and generalization ability, and the laying hen based on Support vector regression of foundation produces
Egg rate prediction model, parameter include penalty factor and kernel function width cs, wherein the size of penalty factor and kernel function width cs
There is vital influence to SVM.C is used to adjust the balance between model complexity and experience error, and σ influences feature space
The complexity of middle sample data distribution.The present invention is using grid data service and cross-validation method to penalty factor and kernel function
Width cs parameter combination performance carries out evaluation to select optimal (C, σ) parameter combination, obtains optimal Support vector regression
Laying rate of laying hen prediction model.
Grid data service (grid search) study precision with higher, algorithm be simple, easy to accomplish and can search
Rope to delimit grid in optimal solution.Grid data service is that penalty factor and kernel function width cs are taken M, N number of value respectively, to M
The combination of × N number of (C, σ), is respectively trained different models, then estimates that it learns precision, to be learnt in these combinations
The highest combination of precision is used as optimized parameter.Cross validation (cross validation) be it is a kind of for eliminate sample with
The statistical method for the training deviation that machine generates.Training data is divided into K subset by it, using wherein any one subset as
Other K-1 subsets are obtained decision function as training set by test set.By not repetitive cycling until each subset
It is all predicted once as test set, is finally rounded the average value of body mean square error as final prediction error, to evade
Overfitting problem.Therefore, the application evaluates the parameter combination performance of selection using cross validation in grid data service
To select best parameter group, the accuracy rate of comprehensive SVM simulation velocity and prediction model chooses K=20 as number of subsets.
Specific step is as follows:
Step 1: initialization (C, σ) range of choice, subset number, optimizing termination condition and grid search step-length；
Step 2: reference SVM learns sample in all (C, σ) is combined, using cross validation mean square deviation (MSE)
It indicates:
In formula (1), yiFor training set output valve, i.e., i-th point of actual temperature error or error rate；y′iIt is i-th
The predicted value of a point.
Step 3: comparing MSE, choose the smallest parameter combination of MSE as optimized parameter.For identical MSE, due to mistake
The big state that will lead to overfitting then chooses the lesser parameter combination of C.
Step 4: constantly reducing search range and step-size in search according to step 3, repeat step 2, until MSE or search
Step-length reaches setting value, and last parameter combination is best optimizing result.
As shown in Figure 1, set the historical data of laying hen day laying rate asWherein N is data length, chooses laying hen the
Once start to lay eggs until the 378th day sells as data length, YnThe laying rate laid eggs in other words n-th day for n-th group.If shadow
Ring factor historical data bexnFor the influence factor data composition of seven groups of laying rate
Vector,For the influence factor n-th group of data of i-th kind of laying rate.Building time series is X hereinn=[xn,xn+1,...,
xn+d], XnFor xnTo xn+dThe matrix of composition, d are the width of data set, choose d=5 herein.
The key for then establishing laying rate prediction model is to determine mapping relations:
Yn=F (Xn), n=1,2 ..., N (2)
Building training sample is needed thus,Wherein XnFor input sample, YnTo export sample, and with first chicken
Laying rate influence factor and laying rate data be sample set training SVM.According to the above-mentioned grid data service used and friendship
Fork proof method looks for optimal (C, σ) parameter combination, and substitutes into support vector machines and be trained.Using the approximation capability of SVM,
So that the input and output of SVM are approximant (2), to establish laying rate prediction model.Later by seven laying rate of second batch chicken
Influence factor, 378 groups of data of same each single item substitute into trained SVM, as test sample, the result predicted.
Low-dimensional linearly inseparable is transformed into the process that High-dimensional Linear can divide by structure sheaf, that is, kernel function of SVM, and used herein is RBF core
Function, furthermore existing conventional treatment method, therefore excessive introduction is not done herein.
As shown in Fig. 2, being emulated using MATLAB, used tool box is least square method supporting vector machine
(LSSVM).Resulting predicted value and laying rate are relatively coincide, and the tendency of prediction model is also coincide with the tendency of practical laying rate.
And emulated by applicant's many experiments, prediction effect is more stable, and the laying rate of laying hen of the invention returned based on SVM predicts mould
The available good prediction result of type.
Due to before this without related laying rate prediction model, it is also necessary to laying rate of laying hen of the assessment based on Support vector regression
The performance of prediction model；For the performance for further assessing the laying rate of laying hen prediction model that SVM is returned, applicant is similarly to count
According to laying rate prediction model being established based on BP neural network, for comparing with the performance of SVM regressive prediction model.This hair
It is bright to use duration used in network training, test, the index of mean square error MSE and correlation coefficient r as prediction model performance.Duration
More short then network advantage is bigger, and the value of MSE is smaller to illustrate that model prediction performance is better, and the value of correlation coefficient r illustrates reality closer to 1
The degree of correlation is higher between actual value and predicted value, its calculation formula is:
In formula, ytiAnd ypiThe actual value and predicted value of respectively i-th sample,WithRespectively represent the reality of n sample
Actual value and predicted value mean value.
The test result of SVM and BP neural network is shown in Table 1:
| Prediction model | SVM | BP neural network |
| MSE | 0.0010 | 0.0017 |
| r | 0.9923 | 0.9884 |
| Training time (s) | 0.0183 | 11.6577 |
| Testing time (s) | 0.0338 | 0.0154 |
The laying rate prediction result of 1 SVM of table and BP neural network compares
The prediction result of consolidated statement 1 is it is found that SVM performance on laying rate of laying hen prediction model is better than BP neural network.And
SVM prediction model result is relatively stable, and the fluctuation based on BP neural network is larger, and stability is poor.And from the angle of popularization
Analysis, the prediction model time-consuming based on SVM will be much smaller than the time based on BP neural network.To sum up, the laying hen returned based on SVM
Laying rate prediction has good estimated performance, and high stability, has confidence level and promotional value.
3), using the laying rate of laying hen prediction model based on Support vector regression of foundation, according to influence factor to egg
Following laying rate of chicken is predicted, the change in future of the laying rate index of laying hen is obtained.
Above-described embodiment is only used for illustrating inventive concept of the invention, rather than the restriction to rights protection of the present invention,
It is all to be made a non-material change to the present invention using this design, protection scope of the present invention should all be fallen into.
Claims (10)
1. a kind of laying rate of laying hen prediction technique based on Support vector regression, it is characterised in that: the following steps are included:
1) laying rate of the history of laying hen and the influence factor data of history, are acquired；
2), establish prediction model: the prediction model based on time series will affect factor as input sample, laying rate made
To export sample, the laying rate of laying hen prediction model based on Support vector regression is established；
3), using the laying rate of laying hen prediction model based on Support vector regression of foundation, according to influence factor to laying hen
Following laying rate is predicted, the change in future of the laying rate index of laying hen is obtained.
2. a kind of laying rate of laying hen prediction technique based on Support vector regression according to claim 1, feature exist
In: in the step 1), influence factor include laying hen feed intake, laying hen chicken age, weight, temperature, light irradiation time, intensity of illumination with
And whether take nutrient.
3. a kind of laying rate of laying hen prediction technique based on Support vector regression according to claim 1, feature exist
In: in the step 1), the acquisition method of laying rate is, with the daily total egg production of laying hen divided by the sum of laying hen, to obtain every
The daily laying rate of laying hen.
4. a kind of laying rate of laying hen prediction technique based on Support vector regression according to claim 2, feature exist
In: the acquisition method of the laying hen feed intake is, with the daily total collection capacity of laying hen divided by the sum of laying hen, to obtain every laying hen
Daily laying hen feed intake.
5. a kind of laying rate of laying hen prediction technique based on Support vector regression according to claim 2, feature exist
In: the acquisition method of the weight is that the weight for acquiring a laying hen weekly calculates laying hen using Lagrange interpolation method
Daily weight.
6. a kind of laying rate of laying hen prediction technique based on Support vector regression according to claim 2, feature exist
In: the acquisition method of the temperature is the maximum temperature and minimum temperature of acquisition same day henhouse.
7. a kind of laying rate of laying hen prediction technique based on Support vector regression according to claim 1, feature exist
In: the step 2) is specifically, establish prediction model: choosing two parts of historical datas, every part of historical data includes influence factor
And laying rate, the prediction model based on time series choose gaussian kernel function, using influence factor as input sample, to lay eggs
Rate is as output sample, and using a copy of it historical data as training sample, laying hen of the training based on Support vector regression is produced
Egg rate prediction model；Using another historical data as test sample, the laying rate of laying hen based on Support vector regression is tested
Prediction model.
8. a kind of laying rate of laying hen prediction technique based on Support vector regression according to claim 7, feature exist
In: the step 2) is specifically, establish prediction model: set the historical data of laying rate asWherein N is data length,
YnThe laying rate laid eggs in other words n-th day for n-th group；If the historical data of influence factor is
I=7, xnFor the vector of the influence factor data composition of seven groups of laying rate,For the influence factor n-th group number of i-th kind of laying rate
According to building time series is Xn=[xn,xn+1,...,xn+d], XnFor xnTo xn+dThe matrix of composition, d are the width of data set, really
Determine mapping relations:
Yn=F (Xn), n=1,2 ..., N (1)；
Training sample is constructed,Wherein XnFor input sample, YnTo export sample, gaussian kernel function is chosen, chooses two parts
Historical data, every part of historical data include influence factor and laying rate, are sample set training by the historical data of a copy of it
SVM, and the parameter combination of optimal penalty factor and kernel function width cs is obtained, it substitutes into support vector machines and is trained, utilize
The approximation capability of SVM, so that the input and output of SVM are approximant (1), so that the laying rate of laying hen for establishing Support vector regression is pre-
Survey model；The influence factor in another historical data established SVM is substituted into as test sample later to be predicted
Result.
9. a kind of laying rate of laying hen prediction technique based on Support vector regression according to claim 8, feature exist
In: the method for obtaining the parameter combination of optimal penalty factor and kernel function width cs is tested using grid data service and intersection
Demonstration carries out evaluating to select optimal (C, σ) parameter combination to penalty factor and kernel function width cs parameter combination performance,
Specific step is as follows:
Step 1: initialization (C, σ) range of choice, subset number, optimizing termination condition and grid search step-length；
Step 2: reference SVM learns sample in all (C, σ) is combined, and is indicated using cross validation mean square deviation (MSE):
In formula (2), yiFor training set output valve, i.e., i-th point of actual temperature error or error rate；y′iIt is i-th point
Predicted value；
Step 3: comparing MSE, choose the smallest parameter combination of MSE as optimized parameter, for identical MSE, due to excessive
It will lead to the state of overfitting, then choose the lesser parameter combination of C；
Step 4: constantly reducing search range and step-size in search according to step 3, repeat step 2, until MSE or step-size in search
Reach setting value, last (C, σ) parameter combination is best optimizing (C, σ) parameter combination.
10. a kind of laying rate of laying hen prediction technique based on Support vector regression according to claim 1, feature exist
In: it further include assessing the laying rate of laying hen prediction model based on Support vector regression between the step 2) and step 3)
Performance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810778061.9A CN109034466A (en) | 2018-07-16 | 2018-07-16 | A kind of laying rate of laying hen prediction technique based on Support vector regression |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810778061.9A CN109034466A (en) | 2018-07-16 | 2018-07-16 | A kind of laying rate of laying hen prediction technique based on Support vector regression |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109034466A true CN109034466A (en) | 2018-12-18 |
Family
ID=64642559
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810778061.9A Pending CN109034466A (en) | 2018-07-16 | 2018-07-16 | A kind of laying rate of laying hen prediction technique based on Support vector regression |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109034466A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109496987A (en) * | 2018-12-30 | 2019-03-22 | 江苏省家禽科学研究所 | A kind of selection improving local egg articles system egg number |
| CN112348238A (en) * | 2020-10-27 | 2021-02-09 | 浙江师范大学 | Egg yield prediction PSO-SVM regression model based on principal component analysis |
| CN113033890A (en) * | 2021-03-20 | 2021-06-25 | 南通天成现代农业科技有限公司 | Method for analyzing laying performance of laying hens based on vector autoregressive model |
| CN115481808A (en) * | 2022-09-23 | 2022-12-16 | 江苏天成科技集团有限公司 | Laying hen laying rate prediction method based on MDT-LSSVM model |
| CN116579508A (en) * | 2023-07-13 | 2023-08-11 | 海煜(福州)生物科技有限公司 | Fish prediction method, device, equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103412543A (en) * | 2013-08-15 | 2013-11-27 | 马从国 | Layer chicken coop laminated cage culturing environment illumination control system |
| CN107289998A (en) * | 2017-05-12 | 2017-10-24 | 淮阴工学院 | Pig house environment temperature intelligent monitoring system based on CAN |
-
2018
- 2018-07-16 CN CN201810778061.9A patent/CN109034466A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103412543A (en) * | 2013-08-15 | 2013-11-27 | 马从国 | Layer chicken coop laminated cage culturing environment illumination control system |
| CN107289998A (en) * | 2017-05-12 | 2017-10-24 | 淮阴工学院 | Pig house environment temperature intelligent monitoring system based on CAN |
Non-Patent Citations (2)
| Title |
|---|
| 向昌盛: ""基于支持向量机的时间序列组合预测模型"", 《中国优秀博硕士学位论文全文数据库(博士)农业科技辑》 * |
| 顾凯成: ""改进网格搜索的支持向量机参数优化研究及应用"", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109496987A (en) * | 2018-12-30 | 2019-03-22 | 江苏省家禽科学研究所 | A kind of selection improving local egg articles system egg number |
| CN112348238A (en) * | 2020-10-27 | 2021-02-09 | 浙江师范大学 | Egg yield prediction PSO-SVM regression model based on principal component analysis |
| CN113033890A (en) * | 2021-03-20 | 2021-06-25 | 南通天成现代农业科技有限公司 | Method for analyzing laying performance of laying hens based on vector autoregressive model |
| CN115481808A (en) * | 2022-09-23 | 2022-12-16 | 江苏天成科技集团有限公司 | Laying hen laying rate prediction method based on MDT-LSSVM model |
| CN116579508A (en) * | 2023-07-13 | 2023-08-11 | 海煜(福州)生物科技有限公司 | Fish prediction method, device, equipment and storage medium |
| CN116579508B (en) * | 2023-07-13 | 2023-10-13 | 海煜(福州)生物科技有限公司 | Fish prediction method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109034466A (en) | A kind of laying rate of laying hen prediction technique based on Support vector regression | |
| CN112508275B (en) | Power distribution network line load prediction method and equipment based on clustering and trend indexes | |
| CN106971240A (en) | The short-term load forecasting method that a kind of variables choice is returned with Gaussian process | |
| CN110728401A (en) | Short-term power load prediction method of neural network based on squirrel and weed hybrid algorithm | |
| CN112906298B (en) | Blueberry yield prediction method based on machine learning | |
| CN110111848A (en) | A kind of human cyclin expressing gene recognition methods based on RNN-CNN neural network fusion algorithm | |
| Liu et al. | Prediction of dissolved oxygen content in aquaculture of Hyriopsis cumingii using Elman neural network | |
| CN114492191A (en) | Heat station equipment residual life evaluation method based on DBN-SVR | |
| CN107145941A (en) | The real-time dynamic acquisition method of light requirement based on optimal light quality and photon flux density | |
| CN114242156A (en) | Real-time prediction method and system for relative abundance of pathogenic vibrios on marine micro-plastic | |
| CN105930531A (en) | Method for optimizing cloud dimensions of agricultural domain ontological knowledge on basis of hybrid models | |
| Wu et al. | Predicting daily global solar radiation in various climatic regions of China based on hybrid support vector machines with meta-heuristic algorithms | |
| CN110516835A (en) | A kind of Multi-variable Grey Model optimization method based on artificial fish-swarm algorithm | |
| CN106502093A (en) | Water island dosing On-Line Control Method and device based on GA SVR | |
| CN113221447A (en) | Soil humidity prediction method for optimizing BP neural network based on improved genetic algorithm | |
| CN111126827A (en) | Input-output accounting model construction method based on BP artificial neural network | |
| CN113591399B (en) | Short-term wind power prediction method | |
| CN113222288B (en) | Classified evolution and prediction method of village and town community space development map | |
| CN116316542A (en) | Regional distributed photovoltaic power prediction method, regional distributed photovoltaic power prediction device, computer equipment and storage medium | |
| CN103778327A (en) | Curve fitting method | |
| CN114139783A (en) | Wind power short-term power prediction method and device based on nonlinear weighted combination | |
| CN109858127B (en) | Blue algae bloom prediction method based on recursive time sequence deep confidence network | |
| Yan et al. | Water demand forecast model of least squares support vector machine based on particle swarm optimization | |
| CN112581311A (en) | Method and system for predicting long-term output fluctuation characteristics of aggregated multiple wind power plants | |
| Geng et al. | Study on prediction of dissolved oxygen content in aquaculture water |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20181218 |
|
| WD01 | Invention patent application deemed withdrawn after publication |