CN113570156A - Thermal environment prediction model and thermal environment prediction method based on agricultural facilities - Google Patents
Thermal environment prediction model and thermal environment prediction method based on agricultural facilities Download PDFInfo
- Publication number
- CN113570156A CN113570156A CN202110950819.4A CN202110950819A CN113570156A CN 113570156 A CN113570156 A CN 113570156A CN 202110950819 A CN202110950819 A CN 202110950819A CN 113570156 A CN113570156 A CN 113570156A
- Authority
- CN
- China
- Prior art keywords
- thermal environment
- term
- long
- short
- prediction result
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000007774 longterm Effects 0.000 claims abstract description 91
- 230000015654 memory Effects 0.000 claims abstract description 66
- 230000007613 environmental effect Effects 0.000 claims abstract description 13
- 238000000605 extraction Methods 0.000 claims description 9
- 230000007787 long-term memory Effects 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 230000000306 recurrent effect Effects 0.000 claims description 3
- 230000006403 short-term memory Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 4
- 230000007547 defect Effects 0.000 abstract description 2
- 239000010410 layer Substances 0.000 description 8
- CURLTUGMZLYLDI-UHFFFAOYSA-N Carbon dioxide Chemical compound O=C=O CURLTUGMZLYLDI-UHFFFAOYSA-N 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 238000012549 training Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 230000007246 mechanism Effects 0.000 description 5
- 238000004590 computer program Methods 0.000 description 4
- 239000000284 extract Substances 0.000 description 4
- 239000002689 soil Substances 0.000 description 4
- 230000003068 static effect Effects 0.000 description 4
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 4
- 230000001364 causal effect Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 241000251468 Actinopterygii Species 0.000 description 2
- 230000004913 activation Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 229910002092 carbon dioxide Inorganic materials 0.000 description 2
- 239000001569 carbon dioxide Substances 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 235000013311 vegetables Nutrition 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000009360 aquaculture Methods 0.000 description 1
- 244000144974 aquaculture Species 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000010339 dilation Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000013209 evaluation strategy Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000010946 mechanistic model Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 230000003134 recirculating effect Effects 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 239000002356 single layer Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000031068 symbiosis, encompassing mutualism through parasitism Effects 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
- G06N3/061—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using biological neurons, e.g. biological neurons connected to an integrated circuit
-
- 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
-
- 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
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Business, Economics & Management (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Neurology (AREA)
- Marketing (AREA)
- Animal Husbandry (AREA)
- Agronomy & Crop Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Marine Sciences & Fisheries (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Primary Health Care (AREA)
- Mining & Mineral Resources (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Evolutionary Biology (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
Abstract
The invention provides a thermal environment prediction model and a thermal environment prediction method based on agricultural facilities, wherein the model comprises the following steps: the first memory module is used for extracting the characteristics of the time series information corresponding to the thermal environment to obtain the long-term space-time characteristics of the thermal environment, and the time series information comprises: an environmental parameter and a target parameter of the thermal environment; the second memory module is used for extracting short-term space-time characteristics of the thermal environment based on the long-term space-time characteristics; the autoregressive module is used for determining an initial prediction result based on the historical information of the target parameter; and the output module is used for outputting a final prediction result of the thermal environment based on the long-term space-time characteristic, the short-term space-time characteristic and the initial prediction result. The method is used for solving the defect that the effect is poor in the long-term prediction task of the thermal environment due to the fact that the characteristics of the thermal environment of the agricultural facility are not combined in the prior art.
Description
Technical Field
The invention relates to the technical field of agricultural facilities, in particular to a thermal environment prediction model and a thermal environment prediction method based on agricultural facilities.
Background
The prediction method for the thermal environment in facility agriculture (greenhouse, industrial fish and vegetable symbiosis, soilless culture and recirculating aquaculture) can be divided into mechanism model prediction, machine learning model prediction and time series model prediction. The latter two methods are called "black box models" because they do not care for the laws and principles of physics in the system. The application of these methods in mechanism predictive modeling will be described in detail below.
The mechanism model uses biological and physical principles to quantitatively analyze relevant factors in a system and establishes an equilibrium equation through the law of conservation of energy and mass. The mechanism model comprises a static model and a dynamic model: static models are also referred to as steady-state models; the dynamic model is usually described by a differential equation, which describes the law of change of the system. The earliest static model of the greenhouse environment was established by the student Businger, which laid the foundation for the later study. Although static models are easy to implement, they are less accurate. Thus, the japanese scholar Takakura et al established a first relatively complete kinetic model of an unheated single-layer glass greenhouse that fully describes the heat and moisture transfer processes in the greenhouse.
With the continuous improvement of the mechanism model, the method also brings convenience for the prediction and regulation of the agricultural facility environment. However, due to the large number of parameters and physical variables used by the mechanistic model, agricultural systems tend to be time-varying. Consequently, tuning is difficult in development and practice, and more researchers are beginning to focus on data-based models. And with the development of artificial intelligence, the computational performance and the data productivity are further improved, so that the application of a machine learning model in agricultural environment modeling is promoted.
Although the time series prediction can be realized by the prior art through a machine learning mode, the characteristics of the thermal environment of agricultural facilities are not combined, and the satisfactory effect cannot be achieved in the task of predicting the thermal environment for a long time. Therefore, how to realize accurate prediction of the thermal environment of agricultural facilities is an important issue to be solved in the industry at present.
Disclosure of Invention
The invention provides a thermal environment prediction model and a thermal environment prediction method based on agricultural facilities, which are used for solving the defect of poor effect in a long-term thermal environment prediction task caused by the fact that the characteristics of the thermal environment of the agricultural facilities are not combined in the prior art and realizing accurate prediction of the thermal environment.
The invention provides a thermal environment prediction model based on agricultural facilities, which comprises the following components: the first memory module is used for extracting the characteristics of time series information corresponding to the thermal environment to obtain the long-term space-time characteristics of the thermal environment, wherein the time series information comprises: an environmental parameter and a target parameter of the thermal environment;
a second memory module for extracting short-term spatiotemporal features of the thermal environment based on the long-term spatiotemporal features;
the autoregressive module is used for determining an initial prediction result based on the historical information of the target parameter;
an output module to output a final prediction result of the thermal environment based on the long-term spatiotemporal features, the short-term spatiotemporal features, and the initial prediction result.
According to the invention, the second memory module comprises: an attention unit and at least two second memory units;
the attention unit is used for connecting the first memory module and the second memory module, adjusting the long-term space-time characteristics, and transmitting the adjusted long-term space-time characteristics to the second memory unit.
According to the thermal environment prediction model based on the agricultural facility, the first memory module comprises at least two first memory units;
the attention unit is specifically configured to, after obtaining the long-term spatiotemporal feature, calculate an attention weight corresponding to each of the first memory units based on the long-term spatiotemporal feature and the short-term spatiotemporal feature, adjust the long-term spatiotemporal feature based on the attention weight, and transmit the adjusted long-term spatiotemporal feature to the second memory unit.
According to the thermal environment prediction model based on the agricultural facility, the second memory unit is used for determining the short-term space-time characteristics based on the adjusted long-term space-time characteristics and the historical information.
According to the agricultural facility-based thermal environment prediction model provided by the invention, the second memory module is a Recurrent Neural Network (RNN), and the second memory unit comprises: long and short term memory networks LSTM.
According to the invention, the first memory unit comprises: time domain convolutional network TCN.
According to the thermal environment prediction model based on the agricultural facility, the output module is specifically used for summing the long-term space-time characteristics, the short-term space-time characteristics and the initial prediction result to obtain the final prediction result.
The invention also provides a thermal environment prediction method based on the thermal environment prediction model of the agricultural facility, which comprises the following steps:
performing feature extraction on time series information corresponding to the thermal environment to obtain long-term space-time features of the thermal environment, wherein the time series information comprises: an environmental parameter and a target parameter of the thermal environment;
extracting short-term spatiotemporal features of the thermal environment based on the long-term spatiotemporal features;
determining an initial prediction result based on the historical information of the target parameter;
outputting a final prediction result of the thermal environment based on the long-term spatiotemporal features, the short-term spatiotemporal features, and the initial prediction result.
The method for predicting the thermal environment, provided by the invention, is used for extracting the short-term space-time characteristics of the thermal environment based on the long-term space-time characteristics, and comprises the following steps:
and adjusting the long-term spatiotemporal features, and determining the short-term spatiotemporal features based on the adjusted long-term spatiotemporal features and the historical information.
The thermal environment prediction method provided by the invention is characterized in that the final prediction result of the thermal environment is output based on the long-term space-time characteristics, the short-term space-time characteristics and the initial prediction result, and comprises the following steps:
and summing the long-term space-time characteristics, the short-term space-time characteristics and the initial prediction result to obtain the final prediction result.
The invention provides a thermal environment prediction model and a thermal environment prediction method based on agricultural facilities, which can obtain a prediction result of a thermal environment through the thermal environment prediction model, wherein the thermal environment prediction model comprises the following steps: the device comprises a first memory module, a second memory module, an autoregressive module and an output module. The invention carries out feature extraction on time series information corresponding to a thermal environment through a first memory module to obtain long-term space-time features of the thermal environment, wherein the time series information comprises: the thermal environment prediction model is effectively based on the relevant parameters of the thermal environment, and obtains the long-term space-time characteristics of the thermal environment; the second memory module extracts the short-term space-time characteristics of the thermal environment based on the long-term space-time characteristics of the thermal environment, so that the short-term space-time characteristics of the thermal environment are obtained; the autoregressive module is used for determining an initial prediction result based on the historical information of the target parameter; the output module is used for outputting the final prediction result of the thermal environment based on the long-term space-time characteristics, the short-term space-time characteristics and the initial prediction result.
Drawings
In order to more clearly illustrate the technical solutions of the present invention or the prior art, the drawings needed for the description of the embodiments or 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 those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a schematic structural diagram of a thermal environment prediction model provided by the present invention;
FIG. 2 is a second schematic structural diagram of a thermal environment prediction model provided by the present invention;
FIG. 3 is a third schematic structural diagram of a thermal environment prediction model provided by the present invention;
FIG. 4 is a flow chart of a thermal environment prediction method provided by the present invention;
fig. 5 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 clearer, 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 agricultural facility-based thermal environment prediction model of the present invention is described below in conjunction with fig. 1-3.
The embodiment of the invention provides a thermal environment prediction model based on agricultural facilities, a structural schematic diagram of the model is shown in figure 1, and the model specifically comprises:
the first memory module 101 is configured to perform feature extraction on time series information corresponding to a thermal environment to obtain long-term spatiotemporal features of the thermal environment, where the time series information includes: an environmental parameter and a target parameter of the thermal environment.
The second memory module 102 is configured to extract short-term spatiotemporal features of the thermal environment based on the long-term spatiotemporal features.
And the autoregressive module 103 is used for determining an initial prediction result based on the historical information of the target parameter.
And the output module 104 is used for outputting a final prediction result of the thermal environment based on the long-term space-time characteristic, the short-term space-time characteristic and the initial prediction result.
Wherein the long-term spatiotemporal features include: seasonal information, change trend information, temporal and spatial correlation and the like of the thermal environment; short-term spatiotemporal features include: short-term fluctuations in thermal environment, etc.
Specifically, the method is mainly aimed at the task of predicting the time series information of the thermal environment of the agricultural facility. Wherein the environmental parameters include: outdoor temperature, outdoor humidity, soil temperature, soil moisture, atmospheric pressure, indoor lighting, indoor carbon dioxide (CO)2) Concentration, indoor temperature, indoor humidity, circulating water temperature, and the like. Wherein the environmental parameters include: exogenous variables and target variables. The target variable corresponds to the final prediction result.
Then, the outdoor temperature, outdoor humidity, soil temperature, soil moisture, atmospheric pressure, indoor illumination, and indoor carbon dioxide (CO)2) The following description will be given, taking as examples the exogenous variables such as concentration, indoor temperature, indoor humidity, and circulating water temperature, and the target variables such as circulating water temperature:
wherein u is used for exogenous variable and target variable respectivelyt∈RNAnd vtE.g. R, which together form the input parameter z of the modelt={ut;vt}. The invention aims to predict target variables of H time steps in the future by using environmental parameters of historical T time steps, namely time sequence information. That is, the thermal environment prediction model establishes a mapping function f (-) of exogenous variables and target variables, see equation (1)
Namely, it is
Y=f(X) (1)
Wherein the input is changedThe amount of the compound is X,the predicted target variable is Y, Y ═ vT+1;vT+2;…;vT+H}∈RHT is a constant and H is a constant
In one embodiment, the first memory module 101 includes at least two first memory units, and the first memory units include: time domain Convolutional Network (TCN for short).
Specifically, due to the characteristics of strong coupling, large inertia, nonlinearity and the like of the facility agriculture thermal environment, data multivariable and long-time sequence need to be considered comprehensively when the model is created, and therefore the TCN model is adopted by the first memory module. Wherein the TCN model ensures that future information is not revealed by causal convolution. Furthermore, the TCN model can efficiently obtain space-time features of arbitrary length by deeper and expanding convolutions.
Therefore, the invention can capture long-term time series information and extract the long-term space-time characteristics of the thermal environment based on the long-term time series information.
Specifically, as shown in fig. 2, each TCN model includes 2 causal convolution TCN sub-modules 201, and each TCN sub-module includes: an extension convolutional layer 2011, a weight normalization layer 2012, an activation function (ReLU) 2013, and a Dropout layer 2014. The causal convolution refers to the convolution of the neuron with the time step t by using the information before the t-th time step in the previous layer, so that the leakage of the time sequence information can be effectively ensured. For a convolution kernel f of size 1 × k, the receptive field of the feature map Z exhibits a multiple increase compared to the general convolution module after adding the dilation convolution factor d, see equation (2). Further, the model has good generalization performance through weight normalization 2012, the ReLU activation function 2013 and the Dropout layer 2014. Finally, residual concatenation is deployed for the model by 1 × 1 convolution 202.
Wherein k is a constant, t is a constant, i is a constant, d is an expansion convolution factor, f is a convolution kernel, and Z is a result of each layer of convolution calculation.
Based on the TCN model, the invention extracts the long-term space-time characteristics of the thermal environment, and the long-term space-time characteristics at the time t are miThe long-term space-time characteristic of the TCN model output is expressed as M ═ { M ═ M1,m2,…,mT}∈RT×p。
In one embodiment, the second memory module 102 includes: an attention unit and at least two second memory units; and the attention unit is used for connecting the first memory module 101 and the second memory module 102, adjusting the long-term space-time characteristics, and transmitting the adjusted long-term space-time characteristics to the second memory unit.
In one embodiment, the second memory module 102 is a Recurrent Neural Network (RNN), and the second memory unit includes: long and short term memory networks LSTM.
The invention adopts the RNN network to effectively capture the short-term fluctuation in the thermal environment.
In one embodiment, the attention unit is specifically configured to, after obtaining the long-term spatiotemporal features, calculate an attention weight corresponding to each first memory unit based on the long-term spatiotemporal features and the short-term spatiotemporal features, adjust the long-term spatiotemporal features based on the attention weights, and transmit the adjusted long-term spatiotemporal features to the second memory unit.
Specifically, the invention effectively fuses the long-term space-time characteristics and the short-term space-time characteristics, so that the RNN can adaptively select a valuable first memory unit for reasoning. For example, at time t, the output result of the LSTM network hidden layer unit is outputOutput m of TCN model as Query target (Query) of attention Uniti∈RpAs a query Key (Key) for attention units, and further, byAnd miAttention weights are calculated, see formula (3) and formula (4):
wherein the hidden unit of the LSTM network outputs dt-1And st-1By mixing d witht-1And st-1Splicing to obtainWattn∈Rm×2q、Uattn∈Rm×pAnd vattn∈RmIs a training parameter for the attention unit,as a query target for attention units, miAs a query key for attention units. Equation (4) for the intermediate energy value to be calculated for each query key valueConversion to attention weight
Wherein, the attention weight may reflect the importance degree of the ith first memory unit to the model prediction target. LSTM network context information c of t time step after adaptive selection of first memory unit by modeltSee formula (5):
wherein, ctT is a constant value as an input parameter to the LSTM network.
In particular, ctI.e. the adjusted long-term spatio-temporal characteristics, ctTo the LSTM network.
In one embodiment, the second memory module 102 is specifically configured to determine the short-term spatiotemporal features based on the adjusted long-term spatiotemporal features and the historical information.
In particular, the long-term memory context information c processed by the attention unitt-1Historical information v of short time window with target variablet-1And (3) splicing the information to obtain the input information of the LSTM network, which is shown in a formula (6) and a formula (7):
wherein,andis a parameter of the training session that is,denotes vt-1And ct-1Splicing result of fLSTMRepresents the forward propagation process of the LSTM network, andfor representing the output results, i.e., short-term spatio-temporal features, of the LSTM network.
In one embodiment, the output module 104 is specifically configured to sum the long-term spatio-temporal features, the short-term spatio-temporal features, and the initial prediction result to obtain a final prediction result.
Specifically, the obtaining manner of the initial prediction result is shown in formula (8):
wherein,and barIs a parameter of the training session that is,indicating the dependency on historical information vt-T+1,vt-T+2,…,vtSingle step initial prediction results. Obtaining an initial prediction result based on the single-step initial prediction result
The autoregressive module is used for solving the generalization problem of the prediction model, has higher precision on thermal environment prediction of the prediction model, and improves the generalization robustness of the prediction model.
Specifically, for the hidden results output by the TCN model and the RNN model, the long-term spatiotemporal features extracted by the TCN model and the short-term spatiotemporal features output by the RNN model are projected to the output latitude required by the thermal environment prediction task through a set of affine transformations, see formula (9) and formula (10):
wherein, Wtcn∈RH×T、Utcn∈Rp、btcn∈RH、Wrnn∈RH×2qAnd brnn∈RHAre training parameters. M ═ M1,m2,…,mT}∈RT×pRepresents the long-term spatio-temporal characteristics of the TCN model output,representing the short-term spatio-temporal characteristics of the LSTM unit output at time t.
Finally, the local outputs of the TCN model, RNN model and autoregressive module 103 are summed to obtain the multi-step prediction result of the thermal environment, see formula (11). In addition, since the model is smooth and differentiable, the parameters in the predictive model can be trained by a back propagation algorithm, see equation (12):
where Θ represents all the training parameters in the model and N is the batch size in the mini-batch gradient descent algorithm. I | · | purple wind2Representing the final predicted outcome of model predictionAnd real thermal environment data Y(i)L2 norm in between.
In addition, after the training of the prediction model is completed, the prediction accuracy of the model needs to be verified, which is specifically as follows:
in the invention, different evaluation strategies are provided from the aspects of the global performance index and the local performance index. Wherein the predicted effect of multiple future time steps is considered global performance, with emphasis on the overall prediction accuracy of the model, and is evaluated in terms of Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percent Error (MAPE), see equations (13), (14), and (15):
second, the degree of fit of each prediction step (i.e., single step evaluation in multi-step prediction results) is considered a local property, effectively reflecting how well the prediction results match real-world production. The relative square root error (RRSE) and the empirical correlation Coefficient (CORR) are used to measure the local performance of the model, see equations (16) and (17):
wherein,is the model predicted value, y is the true measured value, the above index is defined as follows:
global performance index:
local performance index:
wherein i is a constant and represents any number of data samples, h represents the step size of multi-step prediction of the time series in the prediction model, N is a constant and represents the total amount of the data samples, and Ω represents the set of the data samples and the multi-step prediction result.
Wherein, for RMSE, MAE, MAPE and RRSE indexes, the larger the index, the better; for the CORR index, the smaller the index, the better.
The thermal environment prediction model of the present invention is described in detail below with reference to fig. 3:
firstly, connecting and serializing exogenous variables u and target variables v through a long historical information time window T, and obtaining long-term space-time characteristics m as input of a multilayer TCN modeli。
Then, the [ m ] is1,m2,…,mT]Andinput into attention unit to obtain output c of attention unittWhereinis the hidden layer output, i.e., short-term spatio-temporal features, of the LSTM network.
Further, history information v of the short time window is extractedt-1And ct-1Input to the LSTM network.
Further, history information [ v ] of the long time window is stored1,v2,…,vT]The initial prediction result is input to the autoregressive module 103, and the initial prediction result output by the autoregressive module 103 is obtained.
And finally, inputting the long-term space-time characteristics, the short-term space-time characteristics and the initial prediction result into the output module 104 respectively, and obtaining a final prediction result through the output module 104.
In addition, through practical verification in a fish and vegetable symbiotic base, the thermal environment prediction model disclosed by the invention realizes excellent results in multi-step prediction of circulating water temperature. In the prediction tasks of 6-hour, 12-hour and 24-hour time spans, the RMSE indexes respectively achieve the accuracy improvement of 15.40%, 13.93% and 22.15%. In particular, the advantages in long-term prediction of thermal environments are more pronounced. The multi-scale memory structure gives the prediction model time sequence insight from coarse granularity to fine granularity, and can completely learn the rule in the thermal environment time sequence data. In addition, different components used by the method can effectively mine the relevant characteristics of time and space in the thermal environment data, so that the prediction model can acquire a sufficient time and space mode, and the accuracy of the prediction task is effectively improved.
The invention provides a thermal environment prediction model and a thermal environment prediction method based on agricultural facilities, which can obtain a prediction result of a thermal environment through the thermal environment prediction model, wherein the thermal environment prediction model comprises the following steps: the device comprises a first memory module, a second memory module, an autoregressive module and an output module. The invention carries out feature extraction on time series information corresponding to a thermal environment through a first memory module to obtain long-term space-time features of the thermal environment, wherein the time series information comprises: the thermal environment prediction model is effectively based on the relevant parameters of the thermal environment, and obtains the long-term space-time characteristics of the thermal environment; the second memory module extracts the short-term space-time characteristics of the thermal environment based on the long-term space-time characteristics of the thermal environment, so that the short-term space-time characteristics of the thermal environment are obtained; the autoregressive module is used for determining an initial prediction result based on the historical information of the target parameter; the output module is used for outputting the final prediction result of the thermal environment based on the long-term space-time characteristics, the short-term space-time characteristics and the initial prediction result.
The embodiment of the invention provides a thermal environment prediction method based on a thermal environment prediction model of agricultural facilities, wherein the thermal environment prediction method described below and the thermal environment prediction model described above are correspondingly referred to each other, and the specific implementation of the method is shown in fig. 4:
step 401, performing feature extraction on the time series information corresponding to the thermal environment to obtain long-term spatiotemporal features of the thermal environment.
Wherein the time series information includes: an environmental parameter and a target parameter of the thermal environment.
In one embodiment, the long-term spatiotemporal features are adjusted and the short-term spatiotemporal features are determined based on the adjusted long-term spatiotemporal features and historical information.
In step 403, an initial prediction result is determined based on the history information of the target parameter.
In one embodiment, the long-term spatiotemporal features, the short-term spatiotemporal features, and the initial prediction result are summed to obtain a final prediction result.
Fig. 5 illustrates a physical structure diagram of an electronic device, which may include, as shown in fig. 5: a processor (processor)501, a communication Interface (Communications Interface)502, a memory (memory)503, and a communication bus 504, wherein the processor 501, the communication Interface 502, and the memory 503 are configured to communicate with each other via the communication bus 504. Processor 501 may call logic instructions in memory 503 to perform a thermal environment prediction method comprising: performing feature extraction on time series information corresponding to the thermal environment to obtain long-term space-time features of the thermal environment, wherein the time series information comprises: an environmental parameter and a target parameter of the thermal environment; extracting short-term space-time characteristics of the thermal environment based on the long-term space-time characteristics; determining an initial prediction result based on historical information of the target parameter; and outputting a final prediction result of the thermal environment based on the long-term space-time characteristics, the short-term space-time characteristics and the initial prediction result.
In addition, the logic instructions in the memory 503 may be implemented in the form of 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 may be embodied in the form of a software product, which is stored in a storage medium and includes 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 stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions which, when executed by a computer, enable the computer to perform the thermal environment prediction method provided by the above methods, the method comprising: performing feature extraction on time series information corresponding to the thermal environment to obtain long-term space-time features of the thermal environment, wherein the time series information comprises: an environmental parameter and a target parameter of the thermal environment; extracting short-term space-time characteristics of the thermal environment based on the long-term space-time characteristics; determining an initial prediction result based on historical information of the target parameter; and outputting a final prediction result of the thermal environment based on the long-term space-time characteristics, the short-term space-time characteristics and the initial prediction result.
In yet another aspect, the present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program that when executed by a processor is implemented to perform the thermal environment prediction methods provided above, the method comprising: performing feature extraction on time series information corresponding to the thermal environment to obtain long-term space-time features of the thermal environment, wherein the time series information comprises: an environmental parameter and a target parameter of the thermal environment; extracting short-term space-time characteristics of the thermal environment based on the long-term space-time characteristics; determining an initial prediction result based on historical information of the target parameter; and outputting a final prediction result of the thermal environment based on the long-term space-time characteristics, the short-term space-time characteristics and the initial prediction result.
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. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the 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 (10)
1. An agricultural facility-based thermal environment prediction model, comprising:
the first memory module is used for extracting the characteristics of time series information corresponding to the thermal environment to obtain the long-term space-time characteristics of the thermal environment, wherein the time series information comprises: an environmental parameter and a target parameter of the thermal environment;
a second memory module for extracting short-term spatiotemporal features of the thermal environment based on the long-term spatiotemporal features;
the autoregressive module is used for determining an initial prediction result based on the historical information of the target parameter;
an output module to output a final prediction result of the thermal environment based on the long-term spatiotemporal features, the short-term spatiotemporal features, and the initial prediction result.
2. The agricultural facility-based thermal environment prediction model of claim 1, wherein the second memory module comprises: an attention unit and at least two second memory units;
the attention unit is used for connecting the first memory module and the second memory module, adjusting the long-term space-time characteristics, and transmitting the adjusted long-term space-time characteristics to the second memory unit.
3. The agricultural facility-based thermal environment prediction model of claim 2, wherein the first memory module comprises at least two first memory units;
the attention unit is specifically configured to, after obtaining the long-term spatiotemporal feature, calculate an attention weight corresponding to each of the first memory units based on the long-term spatiotemporal feature and the short-term spatiotemporal feature, adjust the long-term spatiotemporal feature based on the attention weight, and transmit the adjusted long-term spatiotemporal feature to the second memory unit.
4. The agricultural facility-based thermal environment prediction model of claim 3, wherein the second memory unit is configured to determine the short-term spatiotemporal features based on the adjusted long-term spatiotemporal features and the historical information.
5. The agricultural facility-based thermal environment prediction model according to any one of claims 2-4, wherein the second memory module is a Recurrent Neural Network (RNN), and the second memory unit comprises: long and short term memory networks LSTM.
6. The agricultural facility-based thermal environment prediction model according to any one of claims 3-4, wherein the first memory unit comprises: time domain convolutional network TCN.
7. The agricultural facility-based thermal environment prediction model of claim 1, wherein the output module is specifically configured to sum the long-term spatiotemporal features, the short-term spatiotemporal features, and the initial prediction results to obtain the final prediction results.
8. A thermal environment prediction method based on the thermal environment prediction model of the agricultural facility as defined in any one of claims 1 to 7, comprising:
performing feature extraction on time series information corresponding to the thermal environment to obtain long-term space-time features of the thermal environment, wherein the time series information comprises: an environmental parameter and a target parameter of the thermal environment;
extracting short-term spatiotemporal features of the thermal environment based on the long-term spatiotemporal features;
determining an initial prediction result based on the historical information of the target parameter;
outputting a final prediction result of the thermal environment based on the long-term spatiotemporal features, the short-term spatiotemporal features, and the initial prediction result.
9. The thermal environment prediction method of claim 8, wherein the extracting short-term spatiotemporal features of the thermal environment based on the long-term spatiotemporal features comprises:
and adjusting the long-term spatiotemporal features, and determining the short-term spatiotemporal features based on the adjusted long-term spatiotemporal features and the historical information.
10. The method of predicting a thermal environment according to claim 9, wherein outputting a final prediction result of the thermal environment based on the long-term spatiotemporal features, the short-term spatiotemporal features, and the initial prediction result comprises:
and summing the long-term space-time characteristics, the short-term space-time characteristics and the initial prediction result to obtain the final prediction result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110950819.4A CN113570156A (en) | 2021-08-18 | 2021-08-18 | Thermal environment prediction model and thermal environment prediction method based on agricultural facilities |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110950819.4A CN113570156A (en) | 2021-08-18 | 2021-08-18 | Thermal environment prediction model and thermal environment prediction method based on agricultural facilities |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113570156A true CN113570156A (en) | 2021-10-29 |
Family
ID=78172123
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110950819.4A Pending CN113570156A (en) | 2021-08-18 | 2021-08-18 | Thermal environment prediction model and thermal environment prediction method based on agricultural facilities |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113570156A (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109492830A (en) * | 2018-12-17 | 2019-03-19 | 杭州电子科技大学 | A kind of mobile pollution source concentration of emission prediction technique based on space-time deep learning |
CN111524535A (en) * | 2020-04-30 | 2020-08-11 | 杭州电子科技大学 | Feature fusion method for speech emotion recognition based on attention mechanism |
CN111785301A (en) * | 2020-06-28 | 2020-10-16 | 重庆邮电大学 | Residual error network-based 3DACRNN speech emotion recognition method and storage medium |
CN112365040A (en) * | 2020-11-03 | 2021-02-12 | 哈尔滨工业大学 | Short-term wind power prediction method based on multi-channel convolution neural network and time convolution network |
CN112988538A (en) * | 2021-03-12 | 2021-06-18 | 山东英信计算机技术有限公司 | Artificial intelligence development platform monitoring alarm data prediction method, device and medium |
-
2021
- 2021-08-18 CN CN202110950819.4A patent/CN113570156A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109492830A (en) * | 2018-12-17 | 2019-03-19 | 杭州电子科技大学 | A kind of mobile pollution source concentration of emission prediction technique based on space-time deep learning |
CN111524535A (en) * | 2020-04-30 | 2020-08-11 | 杭州电子科技大学 | Feature fusion method for speech emotion recognition based on attention mechanism |
CN111785301A (en) * | 2020-06-28 | 2020-10-16 | 重庆邮电大学 | Residual error network-based 3DACRNN speech emotion recognition method and storage medium |
CN112365040A (en) * | 2020-11-03 | 2021-02-12 | 哈尔滨工业大学 | Short-term wind power prediction method based on multi-channel convolution neural network and time convolution network |
CN112988538A (en) * | 2021-03-12 | 2021-06-18 | 山东英信计算机技术有限公司 | Artificial intelligence development platform monitoring alarm data prediction method, device and medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Jiang et al. | Prediction of house price based on the back propagation neural network in the keras deep learning framework | |
CN111563706A (en) | Multivariable logistics freight volume prediction method based on LSTM network | |
CN112541124B (en) | Method, apparatus, device, medium and program product for generating a multitasking model | |
CN113283588B (en) | Near-shore single-point wave height forecasting method based on deep learning | |
CN111091196B (en) | Passenger flow data determination method and device, computer equipment and storage medium | |
CN113065013B (en) | Image annotation model training and image annotation method, system, equipment and medium | |
CN109189922B (en) | Comment evaluation model training method and device | |
CN111882157A (en) | Demand prediction method and system based on deep space-time neural network and computer readable storage medium | |
CN117575663A (en) | Fitment cost estimation method and system based on deep learning | |
CN115545334A (en) | Land use type prediction method, land use type prediction device, electronic device, and storage medium | |
CN114970357A (en) | Energy-saving effect evaluation method, system, device and storage medium | |
CN110852415A (en) | Vegetation index prediction method, system and equipment based on neural network algorithm | |
CN116706907B (en) | Photovoltaic power generation prediction method based on fuzzy reasoning and related equipment | |
CN117557375A (en) | Transaction evaluation method and related device based on virtual power plant | |
CN115392594B (en) | Electrical load model training method based on neural network and feature screening | |
CN117372070A (en) | Method, device, computer equipment and storage medium for determining property market trend | |
CN111768218B (en) | Method and device for processing user interaction information | |
CN116975686A (en) | Method for training student model, behavior prediction method and device | |
CN117172355A (en) | Sea surface temperature prediction method integrating space-time granularity context neural network | |
CN116703466A (en) | System access quantity prediction method based on improved wolf algorithm and related equipment thereof | |
CN113570156A (en) | Thermal environment prediction model and thermal environment prediction method based on agricultural facilities | |
CN115984653A (en) | Method for constructing dynamic intelligent container commodity identification model | |
CN117035038A (en) | Model pruning method, device, equipment and storage medium | |
CN109726690B (en) | Multi-region description method for learner behavior image based on DenseCap network | |
Wang | Forecast model of TV show rating based on convolutional neural network |
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 |