CN115358157B - Prediction analysis method and device for litter size of individual litters and electronic equipment - Google Patents
Prediction analysis method and device for litter size of individual litters and electronic equipment Download PDFInfo
- Publication number
- CN115358157B CN115358157B CN202211284150.0A CN202211284150A CN115358157B CN 115358157 B CN115358157 B CN 115358157B CN 202211284150 A CN202211284150 A CN 202211284150A CN 115358157 B CN115358157 B CN 115358157B
- Authority
- CN
- China
- Prior art keywords
- litter
- influence factor
- characteristic value
- short term
- term memory
- 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.)
- Active
Links
- 238000004458 analytical method Methods 0.000 title claims description 32
- 230000004083 survival effect Effects 0.000 claims abstract description 76
- 238000000034 method Methods 0.000 claims abstract description 30
- 238000012417 linear regression Methods 0.000 claims abstract description 25
- 238000005070 sampling Methods 0.000 claims abstract description 11
- 230000015654 memory Effects 0.000 claims description 104
- 238000012545 processing Methods 0.000 claims description 28
- 238000012549 training Methods 0.000 claims description 13
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 11
- 238000010586 diagram Methods 0.000 description 8
- 230000013011 mating Effects 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 238000009395 breeding Methods 0.000 description 4
- 230000001488 breeding effect Effects 0.000 description 4
- 238000004891 communication Methods 0.000 description 4
- 210000003754 fetus Anatomy 0.000 description 4
- 230000001605 fetal effect Effects 0.000 description 3
- 230000006403 short-term memory Effects 0.000 description 3
- 230000007787 long-term memory Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000009396 hybridization Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000010355 oscillation Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Hardware Design (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Geometry (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a method, a device and electronic equipment for predicting and analyzing the number of alive litter, which relate to the technical field of biological information and comprise the following steps: acquiring an influence factor characteristic value data set corresponding to the litter birth survival number influence factor; inputting the characteristic value data set of the influence factors into a litter survival number prediction model to predict litter survival number, so as to obtain a litter survival number prediction value; sampling the normal distribution sample set to obtain an influence factor characteristic value sample, and replacing the characteristic value sample in the influence factor characteristic value sample set to obtain an influence factor characteristic value new data set; inputting the new data set of the characteristic values of the influence factors into a litter birth liveness prediction model to obtain a new predicted value of the litter birth liveness; and constructing a linear regression model based on the influence factor characteristic value sample and the new predicted value of litter birth survival number. The invention can realize the prediction of litter size, determine the relationship between the characteristic value of the influence factor and the predicted value of litter size, and pointedly improve the litter size of individual litters.
Description
Technical Field
The invention relates to the technical field of biological information, in particular to a method and a device for predicting and analyzing the number of litters born alive and an electronic device.
Background
The prediction of litter size is a very important task. The traditional prediction method for litter size and litter size is usually that business personnel estimate the average litter size and litter size of a batch of sows through the experience accumulated in past breeding, and the method is very limited. Due to the influence of various factors, the litter birth live number of some sows is larger than the estimated average value, and the litter birth live number of some sows is smaller than the estimated average value. Therefore, the individual differences of the sows are usually covered only by observing the average litter size of a batch of sows, but the individual differences of the sows are very important, and a corresponding method for analyzing the influence of the individual differences of the sows on the litter size is lacked at present.
Disclosure of Invention
The invention provides a prediction analysis method and device for the number of live litter of an individual litter and electronic equipment, which can be used for predicting the number of live litter of the individual litter, determining the influence of individual differences of sows on the number of live litter of the individual litter, and further pointedly improving the number of live litter of the individual litter.
The invention provides a prediction analysis method for litter birth survival number, which comprises the following steps:
acquiring an influence factor characteristic value data set corresponding to the litter birth liveness influence factor;
inputting the influence factor characteristic value data set into a trained litter survival number prediction model to predict litter survival number, so as to obtain a predicted value of litter survival number; the litter birth survival doll prediction model is obtained by training based on an influence factor characteristic value sample set and a multi-sequence long-short term memory model;
sampling a normal distribution sample set to obtain an influence factor characteristic value sample, replacing the influence factor characteristic value sample set based on the influence factor characteristic value sample, and obtaining an influence factor characteristic value new data set, wherein the influence factor characteristic value sample and the influence factor characteristic value sample belong to characteristic value samples of the same litter birth survival number influence factor; the normal distribution sample set is obtained by fitting a plurality of influence factor characteristic value samples corresponding to the same litter parity number influence factor in the influence factor characteristic value sample set into normal distribution;
inputting the new influence factor characteristic value data set into the litter birth liveness number prediction model to obtain a new predicted value of the litter birth liveness number;
and constructing a linear regression model based on the influence factor characteristic value sample and the new litter birth survival number predicted value, so as to determine the relation between the influence factor characteristic value and the litter birth survival number predicted value based on the linear regression model.
According to the prediction analysis method for the litter size of the individual, which is provided by the invention, the method for inputting the characteristic value data set of the influence factor into the trained litter size prediction model to predict the litter size and obtain the predicted value of the litter size and comprises the following steps:
based on a plurality of long-short term memory sequence models of the litter birth litter offspring prediction model, carrying out weight setting on the influence factor characteristic values corresponding to the long-short term memory sequence models in the influence factor characteristic value data set to obtain a plurality of weighted characteristic values;
connecting the weighted characteristic values based on a connecting layer of the litter birth survival doll prediction model to obtain a target characteristic value;
and carrying out compact connection on the target characteristic values based on a compact layer of the litter liveness prediction model to obtain the litter liveness prediction value.
According to the prediction analysis method for the litter size of the individual litters provided by the invention, the method for performing weight setting on the influence factor characteristic value data set and the influence factor characteristic value corresponding to the long-short term memory sequence model to obtain a plurality of weighted characteristic values based on a plurality of long-short term memory sequence models of the litter size prediction model comprises the following steps:
respectively carrying out weight setting on the characteristic values of the influence factors received by the long-short term memory sequence model based on a plurality of series-connected long-short term memory sequence units in the long-short term memory sequence model, wherein the weighting characteristic values are obtained;
the last long-short term memory sequence unit in the plurality of series-connected long-short term memory sequence units inputs the weighting processing result of the characteristic value of the influence factor to the next long-short term memory sequence unit so as to be used for the next long-short term memory sequence unit to carry out weight setting on the characteristic value of the influence factor received by the next long-short term memory sequence unit;
the weighting characteristic value is the weighting processing result of the last long-short term memory sequence unit in the plurality of series-connected long-short term memory sequence units on the characteristic value of the influence factor.
According to the prediction analysis method for the litter size of the individual, provided by the invention, the characteristic value of the influence factor comprises a continuity characteristic and a discrete characteristic;
the long-short term memory sequence unit is used for conducting vectorization processing on the discrete type features on the basis of the nested layer to obtain vectorized features, then combining the vectorized features and the continuity features on the basis of the connecting layer to obtain combined features, and conducting weight setting on the combined features on the basis of the long-short term memory sequence layer.
According to the prediction analysis method for the litter size of the individual, which is provided by the invention, the litter size influence factor comprises a sow influence factor, a boar influence factor, a breeder influence factor and a pig farm influence factor.
The prediction analysis method for the litter size of the individual, provided by the invention, further comprises the following steps:
after the litter birth liveness number predicted value is obtained, the litter birth liveness number predicted value is compared with a litter birth liveness number actual value based on an Adam optimizer to obtain a comparison result, and parameters of the litter birth liveness prediction model are updated based on the comparison result.
The invention also provides a prediction analysis device for the number of born alive litters of an individual, which comprises:
the acquisition module is used for acquiring an influence factor characteristic value data set corresponding to the litter birth liveness influence factor;
the first processing module is used for inputting the influence factor characteristic value data set into a trained litter survival number prediction model to predict litter survival number so as to obtain a predicted value of litter survival number; the litter birth survival doll prediction model is obtained by training based on an influence factor characteristic value sample set and a multi-sequence long-short term memory model;
the second processing module is used for sampling the normal distribution sample set to obtain an influence factor characteristic value sample, replacing the influence factor characteristic value sample set based on the influence factor characteristic value sample, and obtaining an influence factor characteristic value new data set, wherein the influence factor characteristic value sample and the influence factor characteristic value sample belong to the characteristic value sample of the influence factor with the same litter birth liveness; the normal distribution sample set is obtained by fitting a plurality of influence factor characteristic value samples corresponding to the same litter of birth liveness influence factors in the influence factor characteristic value sample set into normal distribution;
the third processing module is used for inputting the new influence factor characteristic value data set to the litter birth liveness number prediction model to obtain a new predicted value of the litter birth liveness number;
and the fourth processing module is used for constructing a linear regression model based on the influence factor characteristic value sample and the new litter birth vitality predicted value so as to determine the relation between the influence factor characteristic value and the litter birth vitality predicted value based on the linear regression model.
According to the prediction analysis device for the litter size of the individual, the litter size influence factor comprises a sow influence factor, a boar influence factor, a breeder influence factor and a pig farm influence factor.
According to the present invention, there is provided an apparatus for predicting and analyzing litter size of individual litters, comprising:
and the updating module is used for comparing the predicted litter size value with the actual litter size value based on an Adam optimizer after the predicted litter size value is obtained, obtaining a comparison result, and updating the parameters of the litter size prediction model based on the comparison result.
The present invention also provides an electronic device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for predicting and analyzing litter size of individuals as described in any one of the above.
According to the prediction analysis method and device for the litter size of the individual litters and the electronic equipment, the litter size and the litter size are predicted by processing the characteristic value of the influence factor through the litter size and the litter size prediction model, the litter size and the litter size are predicted by training the litter size and the litter size prediction model based on the first data set and the multi-sequence long-short term memory model, and compared with the situation that people predict the litter size and the litter size through breeding experience, the prediction result of the litter size and the litter size prediction model is more objective and accurate. And the influence factor characteristic value and the corresponding litter size alive number predicted value are subjected to relational analysis through a relational analysis model, so that the relation between the influence factor characteristic value and the litter size alive number predicted value can be determined, the litter size alive of an individual can be pertinently improved, and the economic benefit is improved.
Drawings
In order to more clearly illustrate the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a schematic flow chart of a method for predictive analysis of litter size of individuals provided by the invention;
FIG. 2 is a schematic diagram of a linear regression model provided by the present invention;
FIG. 3 is a schematic diagram of a first part of a litter birth survival doll prediction model provided by the present invention;
FIG. 4 is a schematic diagram of a second part of a litter birth survival doll prediction model provided by the invention;
FIG. 5 is a schematic diagram of a third part of a litter birth survival doll prediction model provided by the invention;
FIG. 6 is a fourth structural diagram of a litter box prediction model provided by the present invention;
FIG. 7 is a schematic diagram of the structure of an apparatus for predicting and analyzing litter size of individuals according to the present invention;
fig. 8 is a schematic structural diagram of an electronic device provided in the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The method, the device and the electronic device for predicting and analyzing the litter size of individuals will be described with reference to fig. 1 to 8.
As shown in fig. 1, the present invention provides a method for analyzing the prediction of litter size, comprising:
and 110, acquiring an influence factor characteristic value data set corresponding to the litter birth survival number influence factor.
It is understood that the litter size influence factor is a factor for imaging the litter size of the individual sow, and the litter size influence factor may be divided into different categories. And specifically quantifying the litter birth survival number influence factor to obtain a corresponding influence factor characteristic value. The impact factor characteristic value data set includes a plurality of impact factor characteristic values.
It can be understood that the influence factor characteristic value data sets formed by the influence factor characteristic values corresponding to a plurality of different types and different births are all input into the trained litter vitality prediction model to predict litter vitality, so as to obtain the litter vitality prediction value.
Wherein, the characteristic values of the influence factors of different fetal times can be the same or different.
It can be understood that a plurality of influence factor characteristic values corresponding to a certain type of litter birth alive litter size influence factor in the influence factor characteristic value sample set are fitted into a normal distribution sample set, a characteristic value sample of the influence factor is randomly extracted from the influence factor characteristic value sample set to serve as a first sample, a characteristic value sample of the influence factor is randomly extracted from the normal distribution sample set to serve as a second sample, the first sample and the second sample belong to the same litter birth alive litter size influence factor, for example, the same litter size belongs to a sow, and the first sample is replaced by the second sample to obtain an influence factor characteristic value new data set.
And step 140, inputting the new influence factor characteristic value data set to the litter liveness prediction model to obtain a new predicted value of the litter liveness.
It can be understood that after obtaining a new predicted value of litter parity number, step 130 is repeated to obtain new sets of impact factor characteristic values corresponding to different samples of impact factor characteristic values obtained by sampling, and further step 140 is performed to obtain new predicted values of litter parity number.
And 150, constructing a linear regression model based on the influence factor characteristic value sample and the new litter birth survival number predicted value, and determining the relation between the influence factor characteristic value and the litter birth survival number predicted value based on the linear regression model.
It is understood that a linear regression model is constructed based on the new predicted litter size values obtained in step 140 and the corresponding samples of the characteristic values of the impact factors, and the linear regression model characterizes the relationship between the characteristic values of the impact factors and the predicted litter size values. That is, the characteristic value of the influence factor is linearly related to the predicted litter birth survival number.
The linear regression model in this example is shown in fig. 2, and fig. 2 shows the correspondence between the number of births of a sow and the number of litter survivors predicted.
In some embodiments, the inputting the characteristic value data set of the influence factor into a trained litter liveness prediction model for litter liveness prediction to obtain a predicted litter liveness value includes:
based on a plurality of long-short term memory sequence models of the litter birth survival doll prediction model, carrying out weight setting on the influence factor characteristic values corresponding to the long-short term memory sequence models in the influence factor characteristic value data set to obtain a plurality of weighted characteristic values;
connecting the weighted characteristic values based on a connecting Layer (connected Layer) of the litter birth survival prediction model to obtain a target characteristic value;
and carrying out compact connection on the target characteristic values based on a compact Layer (Dense Layer) of the litter size prediction model to obtain the litter size number prediction value.
It is understood that the litter size prediction model is shown in fig. 3, 4, 5 and 6, and includes a plurality of long and short term memory sequence models, such as a sow mating birth sequence model, a boar mating birth sequence model and a breeder mating record sequence model. Each long-short term memory sequence model correspondingly processes different kinds of characteristic values of the influence factors. And the connection layer is used for connecting the weighted characteristic values respectively output by the long and short term memory sequence models based on a concatenate function to obtain a target characteristic value.
And the compact layer obtains scalar characteristics (scalar) based on the target characteristic value, namely a littermate survival number predicted value.
In some embodiments, the plurality of long-short term memory sequence models based on the litter size prediction model, for the data set of impact factor characteristic values, and carrying out weight setting on the characteristic values of the influence factors corresponding to the long-short term memory sequence model to obtain a plurality of weighted characteristic values, wherein the weight setting comprises the following steps:
respectively carrying out weight setting on the characteristic values of the influence factors received by the long-short term memory sequence model based on a plurality of series-connected long-short term memory sequence units (LSTM Layer) in the long-short term memory sequence model, wherein the weighting characteristic values are obtained;
the last long-short term memory sequence unit in the plurality of series-connected long-short term memory sequence units inputs the weighting processing result of the characteristic value of the influence factor to the next long-short term memory sequence unit so as to be used for the next long-short term memory sequence unit to carry out weight setting on the characteristic value of the influence factor received by the next long-short term memory sequence unit;
the weighting characteristic value is the weighting processing result of the last long-short term memory sequence unit in the plurality of series-connected long-short term memory sequence units on the characteristic value of the influence factor.
It is understood that each long-short term memory sequence model includes a plurality of long-short term memory sequence units connected in series, and the number of the long-short term memory sequence units included in different long-short term memory sequence models may be the same or different.
In the same long-short term memory sequence model, different long-short term memory sequence units correspondingly receive the characteristic values of the influence factors of the same type and different fetus times for weight setting.
In the same long-short term memory sequence model, the first long-short term memory sequence unit carries out weight setting according to the input influence factor characteristic value, and the next long-short term memory sequence unit needs to carry out weight setting on the influence factor characteristic value by combining the output result of the last long-short term memory sequence unit.
In some embodiments, the impact factor characteristic values include a continuity characteristic and a discrete characteristic;
the long-short term memory sequence unit is used for vectorizing the discrete type features based on an Embedding Layer (Embedding Layer) to obtain vectorized features, combining the vectorized features and the continuity features based on a connecting Layer to obtain combined features, and performing weight setting on the combined features based on the long-short term memory sequence Layer.
It is understood that the long-short term memory sequence unit comprises a nesting layer, a connection layer and a long-short term memory sequence layer.
The characteristic value of the influence factor includes a continuous type and a discrete type. The discrete characteristic values need to be vectorized through a nested layer built in the long-short term memory sequence unit, the continuous characteristic values are directly used as input, and then all the characteristic values are connected through a connecting layer built in the long-short term memory sequence unit and are used as input of each long-short term memory sequence unit.
In some embodiments, the litter size impact factor comprises a sow impact factor, a boar impact factor, a breeder impact factor, and a pig farm impact factor.
It is understood that, according to the knowledge of the experts in the field, the impact factors on litter size are mainly divided into four main categories: sow related, boar related, breeder related and farm related, comprising a total of 25 characteristic values of the influencing factors. Some of the eigenvalues are discrete and some are continuous, and the specific impact factor eigenvalues are shown in the following table:
in some embodiments, the method for predictive analysis of litter size in an individual further comprises:
after the litter liveness prediction value is obtained, the litter liveness prediction value is compared with the actual litter liveness value based on an Adam optimizer to obtain a comparison result, and parameters of the litter liveness prediction model are updated based on the comparison result.
It can be understood that the Adam optimizer is also an adaptive moment estimation optimizer, and after the Adam optimizer combines the litter liveness prediction value and the litter liveness actual value, the Adam optimizer performs back propagation to update the parameters of the litter liveness prediction model.
The Adam optimizer absorbs the advantages of a gradient descent algorithm and a momentum gradient descent algorithm of a self-adaptive learning rate, can adapt to sparse gradients (namely natural language and computer vision problems), and can relieve the problem of gradient oscillation.
In other embodiments, a method for predictive analysis of litter size in an individual, comprising:
three long-short term memory sequence models are first built from three different timing dimensions, for example, based on the fetal sequence of a single sow, the date sequence of a single boar mating, and based on the time sequence of the hybridization operations of a breeder. Then, the output results of the last long-short term memory sequence unit of the three long-short term memory sequence models are combined, and the litter size is predicted.
Selecting data from the nth year to the (n + m) th year, taking the data from the nth year to the (n + k) th year as a training set, taking the data from the (n + k + 1) th year as a verification set, and taking the data from the (n + m) th year as a test set, wherein n can be 2016, k can be 3, and m can be 5.
Sow mating fetal sequence looking 2 sequence units ahead, for example, predicting litter size of third fetus of single sow, looking at first fetus and second fetus of the sow; similarly, boar breeding records and breeder record sequences look forward at 4 sequence units.
The characteristic value of the influence factor includes both a continuous type and a discrete type. The discrete eigenvalues need to be vectorized through the nested layer, the continuous eigenvalues are directly used as input, and then all eigenvalues are connected through the connecting layer to be used as input of each long-short term memory sequence unit.
And the three long-short term memory sequence models are calculated in parallel, the output of the last long-short term memory sequence unit of each long-short term memory sequence model is connected through a connecting layer, and a scalar characteristic is output through a compact layer and is used as a prediction result of the litter size.
The model updates the parameters using Adam optimizer backpropagation.
In summary, the method for analyzing the prediction of the litter size of an individual includes: acquiring an influence factor characteristic value data set corresponding to the litter birth survival number influence factor; inputting the characteristic value data set of the influence factors into a trained litter liveness prediction model to predict litter liveness, so as to obtain a predicted value of litter liveness; the litter birth survival doll prediction model is obtained by training based on an influence factor characteristic value sample set and a multi-sequence long-short term memory model; sampling a normal distribution sample set to obtain an influence factor characteristic value sample, replacing the influence factor characteristic value sample set based on the influence factor characteristic value sample, and obtaining an influence factor characteristic value new data set, wherein the influence factor characteristic value sample and the influence factor characteristic value sample belong to characteristic value samples of the same litter birth survival number influence factor; the normal distribution sample set is obtained by fitting a plurality of influence factor characteristic value samples corresponding to the same litter of birth liveness influence factors in the influence factor characteristic value sample set into normal distribution; inputting the new influence factor characteristic value data set into the litter birth liveness number prediction model to obtain a new predicted value of the litter birth liveness number; and constructing a linear regression model based on the influence factor characteristic value sample and the new litter birth survival number predicted value, so as to determine the relation between the influence factor characteristic value and the litter birth survival number predicted value based on the linear regression model.
According to the prediction analysis method for the litter size of the individual litters, the litter size and the number of the litters is predicted by processing the characteristic value of the influence factor through a litter size and the number of the litters, the litter size and the number of the litters are predicted by a litter size and number prediction model which is obtained based on the first data set and training of a multi-sequence long-short term memory model, and compared with the situation that people predict the litter size and the number of the litters through breeding experience, the prediction result of the litter size and the number prediction model is more objective and accurate. And the influence factor characteristic value and the corresponding litter size and liveness prediction value are subjected to relational analysis through a relational analysis model, so that the relation between the influence factor characteristic value and the litter size and liveness prediction value can be determined, the litter size of an individual can be improved in a targeted manner, and the economic benefit is improved.
The prediction analysis device for litter size of individual litter provided by the present invention will be described below, and the prediction analysis device for litter size of individual litter described below and the prediction analysis method for litter size of individual litter described above can be referred to each other.
As shown in fig. 7, the present invention also provides a device 700 for predicting and analyzing the number of litter litters, comprising:
an obtaining module 710, configured to obtain an influence factor characteristic value data set corresponding to a litter birth liveness influence factor;
the first processing module 720 is configured to input the data set of the characteristic values of the impact factors into a trained litter liveness prediction model to perform litter liveness prediction, so as to obtain a predicted value of the litter liveness; the litter birth survival offspring prediction model is obtained by training based on an influence factor characteristic value sample set and a multi-sequence long-short term memory model;
the second processing module 730 is configured to sample a normal distribution sample set to obtain an influence factor characteristic value sample, replace the influence factor characteristic value sample set based on the influence factor characteristic value sample, and obtain an influence factor characteristic value new data set, where the influence factor characteristic value sample and the influence factor characteristic value sample belong to a characteristic value sample of an influence factor with the same litter birth liveness; the normal distribution sample set is obtained by fitting a plurality of influence factor characteristic value samples corresponding to the same litter of birth liveness influence factors in the influence factor characteristic value sample set into normal distribution;
a third processing module 740, configured to input the new data set of the characteristic values of the impact factors into the litter birth liveness prediction model to obtain a new predicted value of the litter birth liveness;
and a fourth processing module 750, configured to construct a linear regression model based on the sample of the characteristic value of the impact factor and the new predicted litter size, so as to determine a relationship between the characteristic value of the impact factor and the predicted litter size based on the linear regression model.
In some embodiments, the litter size impact factor comprises a sow impact factor, a boar impact factor, a breeder impact factor, and a pig farm impact factor.
In some embodiments, the prediction analysis apparatus 700 for litter size of individuals further comprises:
and the updating module is used for comparing the predicted litter birth liveness number with the actual litter birth liveness number value based on an Adam optimizer after the predicted litter birth liveness number value is obtained, obtaining a comparison result, and updating the parameters of the litter birth liveness prediction model based on the comparison result.
The electronic device, the computer program product, and the storage medium provided by the present invention are described below, and the electronic device, the computer program product, and the storage medium described below may be referred to in correspondence with the above-described method for predicting and analyzing the litter size of an individual litter.
Fig. 8 illustrates a physical structure diagram of an electronic device, and as shown in fig. 8, the electronic device may include: a processor (processor) 810, a communication Interface 820, a memory 830 and a communication bus 840, wherein the processor 810, the communication Interface 820 and the memory 830 communicate with each other via the communication bus 840. Processor 810 may invoke logic instructions in memory 830 to perform a method of predictive analysis of litter size, the method comprising:
acquiring an influence factor characteristic value data set corresponding to the litter birth liveness influence factor;
inputting the influence factor characteristic value data set into a trained litter survival number prediction model to predict litter survival number, so as to obtain a predicted value of litter survival number; the litter birth survival doll prediction model is obtained by training based on an influence factor characteristic value sample set and a multi-sequence long-short term memory model;
sampling a normal distribution sample set to obtain an influence factor characteristic value sample, replacing the influence factor characteristic value sample set based on the influence factor characteristic value sample, and obtaining an influence factor characteristic value new data set, wherein the influence factor characteristic value sample and the influence factor characteristic value sample belong to characteristic value samples of the same litter birth survival number influence factors; the normal distribution sample set is obtained by fitting a plurality of influence factor characteristic value samples corresponding to the same litter of birth liveness influence factors in the influence factor characteristic value sample set into normal distribution;
inputting the new data set of the characteristic values of the influence factors into the litter birth liveness number prediction model to obtain a new predicted value of the litter birth liveness number;
and constructing a linear regression model based on the sample of the characteristic value of the influence factor and the new predicted litter birth liveness value, so as to determine the relationship between the characteristic value of the influence factor and the predicted litter birth liveness value based on the linear regression model.
In addition, the logic instructions in the memory 830 may be implemented in software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solution of the present invention or a part thereof which substantially contributes to the prior art may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, the present invention also provides a computer program product comprising a computer program, the computer program being stored on a non-transitory computer-readable storage medium, the computer program, when executed by a processor, being capable of executing the method for predictive analysis of litter size in individuals provided by the above methods, the method comprising:
acquiring an influence factor characteristic value data set corresponding to the litter birth survival number influence factor;
inputting the influence factor characteristic value data set into a trained litter survival number prediction model to predict litter survival number, so as to obtain a predicted value of litter survival number; the litter birth survival doll prediction model is obtained by training based on an influence factor characteristic value sample set and a multi-sequence long-short term memory model;
sampling a normal distribution sample set to obtain an influence factor characteristic value sample, replacing the influence factor characteristic value sample set based on the influence factor characteristic value sample, and obtaining an influence factor characteristic value new data set, wherein the influence factor characteristic value sample and the influence factor characteristic value sample belong to characteristic value samples of the same litter birth survival number influence factor; the normal distribution sample set is obtained by fitting a plurality of influence factor characteristic value samples corresponding to the same litter parity number influence factor in the influence factor characteristic value sample set into normal distribution;
inputting the new data set of the characteristic values of the influence factors into the litter birth liveness number prediction model to obtain a new predicted value of the litter birth liveness number;
and constructing a linear regression model based on the sample of the characteristic value of the influence factor and the new predicted litter birth liveness value, so as to determine the relationship between the characteristic value of the influence factor and the predicted litter birth liveness value based on the linear regression model.
In yet another aspect, the present invention also provides a non-transitory computer readable storage medium, on which a computer program is stored, the computer program, when executed by a processor, implementing a method for predictive analysis of litter size for individuals provided by the methods above, the method comprising:
acquiring an influence factor characteristic value data set corresponding to the litter birth survival number influence factor;
inputting the influence factor characteristic value data set into a trained litter survival number prediction model to predict litter survival number, so as to obtain a predicted value of litter survival number; the litter birth survival offspring prediction model is obtained by training based on an influence factor characteristic value sample set and a multi-sequence long-short term memory model;
sampling a normal distribution sample set to obtain an influence factor characteristic value sample, replacing the influence factor characteristic value sample set based on the influence factor characteristic value sample, and obtaining an influence factor characteristic value new data set, wherein the influence factor characteristic value sample and the influence factor characteristic value sample belong to characteristic value samples of the same litter birth survival number influence factor; the normal distribution sample set is obtained by fitting a plurality of influence factor characteristic value samples corresponding to the same litter of birth liveness influence factors in the influence factor characteristic value sample set into normal distribution;
inputting the new influence factor characteristic value data set into the litter birth liveness number prediction model to obtain a new predicted value of the litter birth liveness number;
and constructing a linear regression model based on the influence factor characteristic value sample and the new litter birth survival number predicted value, so as to determine the relation between the influence factor characteristic value and the litter birth survival number predicted value based on the linear regression model.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. Based on the understanding, the above technical solutions substantially or otherwise contributing to the prior art may be embodied in the form of a software product, which may be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method according to the various embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (5)
1. A method for predictive analysis of litter size, comprising:
acquiring an influence factor characteristic value data set corresponding to the litter birth survival number influence factor;
inputting the influence factor characteristic value data set into a trained litter survival number prediction model to predict litter survival number, so as to obtain a predicted value of litter survival number; the litter birth survival doll prediction model is obtained by training based on an influence factor characteristic value sample set and a multi-sequence long-short term memory model;
sampling a normal distribution sample set to obtain an influence factor characteristic value sample, replacing the influence factor characteristic value sample set based on the influence factor characteristic value sample, and obtaining an influence factor characteristic value new data set, wherein the influence factor characteristic value sample and the influence factor characteristic value sample belong to characteristic value samples of the same litter birth survival number influence factor; the normal distribution sample set is obtained by fitting a plurality of influence factor characteristic value samples corresponding to the same litter of birth liveness influence factors in the influence factor characteristic value sample set into normal distribution;
inputting the new influence factor characteristic value data set into the litter birth liveness number prediction model to obtain a new predicted value of the litter birth liveness number;
constructing a linear regression model based on the influence factor characteristic value sample and the new litter birth survival number predicted value, and determining the relation between the influence factor characteristic value and the litter birth survival number predicted value based on the linear regression model;
inputting the influence factor characteristic value data set into a trained litter survival number prediction model to predict litter survival number, and obtaining a litter survival number prediction value, wherein the method comprises the following steps:
based on a plurality of long-short term memory sequence models of the litter birth litter offspring prediction model, carrying out weight setting on the influence factor characteristic values corresponding to the long-short term memory sequence models in the influence factor characteristic value data set to obtain a plurality of weighted characteristic values;
connecting the weighted characteristic values based on a connecting layer of the litter birth survival doll prediction model to obtain a target characteristic value;
based on a compact layer of the litter liveness prediction model, carrying out compact connection on the target characteristic values to obtain a litter liveness prediction value;
the weight setting is carried out on the influence factor characteristic values corresponding to the long-short term memory sequence models in the influence factor characteristic value data set by the plurality of long-short term memory sequence models based on the litter birth survival doll prediction model to obtain a plurality of weighted characteristic values, and the weight setting comprises the following steps:
respectively carrying out weight setting on the influence factor characteristic values received by the long-short term memory sequence model on the basis of a plurality of long-short term memory sequence units connected in series in the long-short term memory sequence model, wherein the weighting characteristic values are obtained;
the last long-short term memory sequence unit in the plurality of series-connected long-short term memory sequence units inputs the weighting processing result of the characteristic value of the influence factor to the next long-short term memory sequence unit so as to be used for the next long-short term memory sequence unit to carry out weight setting on the characteristic value of the influence factor received by the next long-short term memory sequence unit;
the weighted characteristic value is a weighted processing result of the last long-short term memory sequence unit in the plurality of series-connected long-short term memory sequence units on the characteristic value of the influence factor;
the influence factor characteristic value comprises a continuity characteristic and a discrete characteristic;
the long-short term memory sequence unit is used for conducting vectorization processing on the discrete type features on the basis of the nested layer to obtain vectorized features, then combining the vectorized features and the continuity features on the basis of the connecting layer to obtain combined features, and conducting weight setting on the combined features on the basis of the long-short term memory sequence layer;
the litter survival number influence factor comprises a sow influence factor, a boar influence factor, a breeder influence factor and a pig farm influence factor.
2. The method for predictive analysis of litter size in an individual of claim 1, further comprising:
after the litter birth liveness number predicted value is obtained, the litter birth liveness number predicted value is compared with a litter birth liveness number actual value based on an Adam optimizer to obtain a comparison result, and parameters of the litter birth liveness prediction model are updated based on the comparison result.
3. An apparatus for predictive analysis of litter size, comprising:
the acquisition module is used for acquiring an influence factor characteristic value data set corresponding to the litter birth liveness influence factor;
the first processing module is used for inputting the characteristic value data set of the influence factor into a trained litter liveness prediction model to predict litter liveness so as to obtain a predicted value of litter liveness; the litter birth survival doll prediction model is obtained by training based on an influence factor characteristic value sample set and a multi-sequence long-short term memory model;
the second processing module is used for sampling the normal distribution sample set to obtain an influence factor characteristic value sample, replacing the influence factor characteristic value sample set based on the influence factor characteristic value sample, and obtaining an influence factor characteristic value new data set, wherein the influence factor characteristic value sample and the influence factor characteristic value sample belong to the characteristic value sample of the influence factor with the same litter birth liveness; the normal distribution sample set is obtained by fitting a plurality of influence factor characteristic value samples corresponding to the same litter parity number influence factor in the influence factor characteristic value sample set into normal distribution; the second processing module is further configured to perform weight setting on the influence factor characteristic value data set and the influence factor characteristic values corresponding to the long-short term memory sequence models based on a plurality of long-short term memory sequence models of the litter birth litter model prediction model to obtain a plurality of weighted characteristic values; based on a connection layer of the litter birth litter prediction model, connecting the weighted characteristic values to obtain a target characteristic value; based on a compact layer of the litter liveness prediction model, carrying out compact connection on the target characteristic values to obtain a litter liveness prediction value; the second processing module is further configured to perform weight setting on the influence factor characteristic values received by the second processing module based on a plurality of long-short term memory sequence units connected in series in the long-short term memory sequence model, respectively, where the weighted characteristic values obtain the weighted characteristic values; the last long-short term memory sequence unit in the plurality of series-connected long-short term memory sequence units inputs the weighting processing result of the characteristic value of the influence factor to the next long-short term memory sequence unit so as to be used for the next long-short term memory sequence unit to carry out weight setting on the characteristic value of the influence factor received by the next long-short term memory sequence unit; the weighted characteristic value is a weighted processing result of the last long-short term memory sequence unit in the plurality of series-connected long-short term memory sequence units on the characteristic value of the influence factor; the influence factor characteristic value comprises a continuity characteristic and a discrete characteristic; the long-short term memory sequence unit is used for conducting vectorization processing on the discrete type features on the basis of the nested layer to obtain vectorized features, then combining the vectorized features and the continuity features on the basis of the connecting layer to obtain combined features, and conducting weight setting on the combined features on the basis of the long-short term memory sequence layer;
the third processing module is used for inputting the new data set of the characteristic values of the influence factors into the litter birth liveness number prediction model to obtain a new predicted value of the litter birth liveness number;
the fourth processing module is used for constructing a linear regression model based on the influence factor characteristic value sample and the new litter birth survival number predicted value so as to determine the relation between the influence factor characteristic value and the litter birth survival number predicted value based on the linear regression model;
the litter birth live litter number influence factor comprises a sow influence factor, a boar influence factor, a breeder influence factor and a pig farm influence factor.
4. The apparatus for predictive analysis of litter size of individuals as set forth in claim 3, further comprising:
and the updating module is used for comparing the predicted litter size value with the actual litter size value based on an Adam optimizer after the predicted litter size value is obtained, obtaining a comparison result, and updating the parameters of the litter size prediction model based on the comparison result.
5. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor, when executing the program, implements a method for predictive analysis of individual litter size as set forth in claim 1 or 2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211284150.0A CN115358157B (en) | 2022-10-20 | 2022-10-20 | Prediction analysis method and device for litter size of individual litters and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211284150.0A CN115358157B (en) | 2022-10-20 | 2022-10-20 | Prediction analysis method and device for litter size of individual litters and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115358157A CN115358157A (en) | 2022-11-18 |
| CN115358157B true CN115358157B (en) | 2023-02-28 |
Family
ID=84008829
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211284150.0A Active CN115358157B (en) | 2022-10-20 | 2022-10-20 | Prediction analysis method and device for litter size of individual litters and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115358157B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105095690A (en) * | 2015-07-21 | 2015-11-25 | 华南农业大学 | Mathematic model for evaluating fertilization capacities of duroc boars and establishment model thereof |
| CN108805176A (en) * | 2018-05-21 | 2018-11-13 | 青岛农业大学 | A kind of fish meal feeding volume prediction technique returned based on random forest |
| CN113642231A (en) * | 2021-07-09 | 2021-11-12 | 西北大学 | CNN-GRU landslide displacement prediction method based on compression excitation network and application |
| CN113868954A (en) * | 2021-09-30 | 2021-12-31 | 苏州科技大学 | Air conditioner load prediction method and system |
| CN113924579A (en) * | 2019-05-29 | 2022-01-11 | 豪夫迈·罗氏有限公司 | Integrated neural network for determining protocol configuration |
| CN114626626A (en) * | 2022-03-25 | 2022-06-14 | 云启智慧科技有限公司 | Model training method and device, storage medium and electronic equipment |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11914674B2 (en) * | 2011-09-24 | 2024-02-27 | Z Advanced Computing, Inc. | System and method for extremely efficient image and pattern recognition and artificial intelligence platform |
| CN108830417B (en) * | 2018-06-13 | 2021-01-08 | 山东师范大学 | ARMA (autoregressive moving average) and regression analysis based life energy consumption prediction method and system |
| US11705226B2 (en) * | 2019-09-19 | 2023-07-18 | Tempus Labs, Inc. | Data based cancer research and treatment systems and methods |
-
2022
- 2022-10-20 CN CN202211284150.0A patent/CN115358157B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105095690A (en) * | 2015-07-21 | 2015-11-25 | 华南农业大学 | Mathematic model for evaluating fertilization capacities of duroc boars and establishment model thereof |
| CN108805176A (en) * | 2018-05-21 | 2018-11-13 | 青岛农业大学 | A kind of fish meal feeding volume prediction technique returned based on random forest |
| CN113924579A (en) * | 2019-05-29 | 2022-01-11 | 豪夫迈·罗氏有限公司 | Integrated neural network for determining protocol configuration |
| CN113642231A (en) * | 2021-07-09 | 2021-11-12 | 西北大学 | CNN-GRU landslide displacement prediction method based on compression excitation network and application |
| CN113868954A (en) * | 2021-09-30 | 2021-12-31 | 苏州科技大学 | Air conditioner load prediction method and system |
| CN114626626A (en) * | 2022-03-25 | 2022-06-14 | 云启智慧科技有限公司 | Model training method and device, storage medium and electronic equipment |
Non-Patent Citations (2)
| Title |
|---|
| 我国南亚热带地区瘦肉型母猪窝产活仔数月份变化的逐步回归分析;陈景仁;《广东畜牧兽医科技》;19940615(第02期);全文 * |
| 窝产活仔数和初生窝重对21日龄仔猪断奶窝重及断奶成活率的相关性研究;金大春 等;《中国畜牧杂志》;20141010;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115358157A (en) | 2022-11-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111124840B (en) | Method and device for predicting alarm in business operation and maintenance and electronic equipment | |
| US11650968B2 (en) | Systems and methods for predictive early stopping in neural network training | |
| CN111243736A (en) | Survival risk assessment method and system | |
| Khorshidi-Jalali et al. | Comparison of artificial neural network and regression models for prediction of body weight in Raini Cashmere goat | |
| US11424034B2 (en) | Systems and methods for predicting animal health | |
| CN113011532B (en) | Classification model training method, device, computing equipment and storage medium | |
| CN115034315B (en) | Service processing method and device based on artificial intelligence, computer equipment and medium | |
| CN110956310A (en) | Fish feed feeding amount prediction method and system based on feature selection and support vector | |
| CN114219562A (en) | Model training method, enterprise credit evaluation method and device, equipment and medium | |
| CN114418182B (en) | Beef cattle breeding optimization method and device based on machine learning | |
| CN111062511A (en) | Aquaculture disease prediction method and system based on decision tree and neural network | |
| CN113743461A (en) | Unmanned aerial vehicle cluster health degree assessment method and device | |
| CN115358157B (en) | Prediction analysis method and device for litter size of individual litters and electronic equipment | |
| CN109934352B (en) | Automatic evolution method of intelligent model | |
| Silveira et al. | Regression trees in genomic selection for carcass traits in pigs | |
| CN114818463B (en) | Vulnerability assessment method and system of feature-based pre-training model selection algorithm | |
| CN115907775A (en) | Personal credit assessment rating method based on deep learning and application thereof | |
| CN112395810B (en) | Fish attack behavior quantification method and device based on artificial neural network and storage medium | |
| CN112529009B (en) | Image feature mining method and device, storage medium and electronic equipment | |
| Shertzer et al. | Identifying growth morphs from mixtures of size-at-age data | |
| CN113744892A (en) | Embryo euploidy prediction method, embryo euploidy prediction device, electronic equipment and storage medium | |
| CN118194163B (en) | Boar scoring method and device, electronic equipment and storage medium | |
| CN117373688B (en) | Chronic disease data processing method, device, electronic equipment and storage medium | |
| Njubi et al. | Milk yield prediction in Kenyan Holstein-Friesian cattle using computer neural networks system | |
| CN117725979B (en) | Model training method and device, electronic equipment and computer readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20231221 Address after: No. 17, Seventh Street, Binhai Economic and Technological Development Zone, Tianjin 300457 Patentee after: C.P. PREMIX (TIANJIN) CO.,LTD. Address before: No. 38-21045, Yuyang Road, Yukou New Energy Industrial Base, Pingguyuanyukou, Zhongguancun Science and Technology Park, Pinggu District, Beijing 101206 (cluster registration) Patentee before: Chia Tai Agricultural Science Research Co.,Ltd. |
|
| TR01 | Transfer of patent right |