CN108133282B - Dendrobium officinale growth environment prediction method - Google Patents

Dendrobium officinale growth environment prediction method Download PDF

Info

Publication number
CN108133282B
CN108133282B CN201711282227.XA CN201711282227A CN108133282B CN 108133282 B CN108133282 B CN 108133282B CN 201711282227 A CN201711282227 A CN 201711282227A CN 108133282 B CN108133282 B CN 108133282B
Authority
CN
China
Prior art keywords
error
momentum
state
rate
learning rate
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201711282227.XA
Other languages
Chinese (zh)
Other versions
CN108133282A (en
Inventor
丁金婷
谢翻翻
屠杭垚
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hangzhou City University
Original Assignee
Zhejiang University City College ZUCC
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Zhejiang University City College ZUCC filed Critical Zhejiang University City College ZUCC
Priority to CN201711282227.XA priority Critical patent/CN108133282B/en
Publication of CN108133282A publication Critical patent/CN108133282A/en
Application granted granted Critical
Publication of CN108133282B publication Critical patent/CN108133282B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • 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)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Economics (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • General Health & Medical Sciences (AREA)
  • Tourism & Hospitality (AREA)
  • Health & Medical Sciences (AREA)
  • Marketing (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Business, Economics & Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Biomedical Technology (AREA)
  • Mining & Mineral Resources (AREA)
  • Animal Husbandry (AREA)
  • Agronomy & Crop Science (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Primary Health Care (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Artificial Intelligence (AREA)
  • Marine Sciences & Fisheries (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention relates to a method for predicting the growth environment of dendrobium officinale, which comprises a growth environment prediction model and an improved BP neural network, wherein the growth environment prediction model comprises the step of normalizing acquired parameters to ensure that the neural network has good fitting property on a sample, and the improved BP neural network comprises the steps of processing an input error signal e (n) by a first-stage fuzzy controller and processing the change rate of the input error by a second-stage fuzzy controller
Figure DDA0001497687480000011
And (6) processing. The invention has the beneficial effects that: the invention provides an improved BP neural network algorithm for predicting the growing environment of plants, the improved BP neural network overcomes the defect of low convergence speed of the traditional BP neural network, and simultaneously, the step length is changed, so that self-adaption step length changing is carried out according to the growing environment by utilizing fuzzy control, and the convergence speed is improved.

Description

Dendrobium officinale growth environment prediction method
Technical Field
The invention relates to a prediction method of a plant growth environment, in particular to a prediction method of a dendrobium officinale growth environment.
Background
The dendrobium officinale is a traditional rare Chinese medicinal material in China, and has the pharmacological effects of enhancing immunity, resisting tumors, resisting oxidation, benefiting the liver and the stomach, reducing blood sugar, nourishing yin, promoting the production of body fluid and the like. Because the growth environment of the dendrobium officinale is special, the natural reproduction rate is low, the dendrobium officinale grows slowly and the like, the supply of the dendrobium officinale is always seriously insufficient, and the important factors influencing the growth of the dendrobium officinale are illumination, temperature, humidity and the like, so that the real-time monitoring and prediction of the growth environment can ensure that the counter-measures are taken in advance under bad conditions, and unnecessary loss is avoided.
At present, the BP neural network has good nonlinear fitting characteristics, and is widely applied to solving the plant growth environment. If prediction of a large-area plant growth process is required, a large amount of data is required for effective learning many times, so that the learning time and space consumption are large.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provide a rapid and efficient method for predicting the growth environment of dendrobium officinale. The purpose of the invention is realized by the following technical scheme:
the method for predicting the growth environment of the dendrobium officinale comprises the following steps:
step one, a growth environment prediction model:
firstly, the obtained parameters are normalized, so that the neural network has good fitting property to the sample:
Figure BDA0001497687460000011
wherein x' is a normalized value, x is a current value, and xmaxIs the maximum value of the sample, xminIs the sample minimum;
setting the number of nodes of an input layer of a neural network as A, wherein A represents the number of main factors influencing the growth environment of the dendrobium officinale, namely soil temperature, soil humidity, air temperature, air humidity, illumination, carbon dioxide and time; the number of nodes of an output layer is B, wherein B represents the number of main factors influencing the prediction of the growth environment of the dendrobium officinale, and A is generally set to be B; the hidden layer node C is dependent on the requirements of input and output;
in order to shorten the training time of the network, a serial input mode is adopted, N training data with known samples are input, and the (N + 1) th data are used for training for a teacher; after finishing, the input data are sequentially moved backwards by one data to serve as the next group of input data, the (N + 2) th data serve as a guide to train, and the like; therefore, the prediction rule of the neural network can be mastered, and the prediction model is as follows:
Dn+1＝F(Dn，Dn-1，Dn-2，Dn-3，...，Dn-m)
F(Dn) Representing a function of predictive mapping of historical data through a neural network, DnA value representing an environmental parameter at time n;
step two, the improved BP neural network comprises the following steps:
firstly, processing data, namely processing data with larger errors, and secondly, processing other data errors by using a secondary fuzzy controller;
1) the first-stage fuzzy controller processes an input error signal e (n): inputting the error caused by the convergence of the neural network into a first-stage fuzzy controller, comparing the error with the last error e (n-1), and judging whether the difference (e (n) -e (n-1)) of the errors is convergent or divergent; the method is characterized in that the method comprises the following steps of dividing a set membership function into seven states of error high-speed rising PB1, error medium-speed rising PM1, error low-speed rising PS1, error stable 01, error low-speed falling NS1, error medium-speed falling NM1 and error high-speed falling NB1, carrying out fuzzy processing, and correspondingly changing learning rate and momentum according to a designed rule base so as to adjust the step size:
if e (n) is in a PB1 state, the learning rate is appropriately reduced, the momentum is set to 0.01, the iteration is cancelled, and the previous iteration is returned;
if e (n) is in a PM1 state, the learning rate and the momentum are properly reduced, and the next iteration is continued;
if e (n) is in PS1 state, then the learning rate and momentum are properly reduced, and the next iteration is continued;
if e (n) is in a state of 01, the learning rate and the momentum are kept unchanged, and the next iteration is continued;
if e (n) is NS1 state, then increasing learning rate and momentum properly, continuing next iteration;
sixthly, if e (n) is NM1 state, the learning rate and momentum are increased properly, and the next iteration is continued;
if e (n) is the state of NB1, the learning rate and the momentum are properly increased, and the next iteration is continued;
2) rate of change of input error by two-stage fuzzy controller
Figure BDA0001497687460000021
And (3) processing: converting the error processed by the first-stage fuzzy controller into an error change rate, inputting the error change rate into the second-stage fuzzy controller, dividing the error into seven states of high-speed rising PB2 of the error rate, medium-speed rising PM2 of the error rate, low-speed rising PS2 of the error rate, stable 02 of the error rate, low-speed falling NS2 of the error rate, medium-speed falling NM2 of the error rate and high-speed falling NB2 of the error rate through set membership functions, and changing the learning rate and momentum correspondingly again so as to adjust the step size:
if delta is in a PB2 state, the learning rate is properly reduced, momentum is set to be 0.01, the iteration is cancelled, and the previous iteration is returned;
if delta is in a PM2 state, the learning rate and momentum are properly reduced, and the next iteration is continued;
if delta is in a PS2 state, properly reducing the learning rate and momentum, and continuing to perform the next iteration;
if the delta is in the 02 state, keeping the learning rate and the momentum unchanged, and continuing to iterate the next time;
if delta is NS2 state, then increasing learning rate and momentum properly, continuing next iteration;
sixthly, if the delta is the NM2 state, the learning rate and the momentum are properly increased, and the next iteration is continued;
if delta is NB2 state, then the learning rate and momentum are properly increased and the next iteration is continued.
The invention has the beneficial effects that: the invention provides an improved BP neural network algorithm for predicting the growing environment of plants, the improved BP neural network overcomes the defect of low convergence speed of the traditional BP neural network, and simultaneously, the step length is changed, so that self-adaption step length changing is carried out according to the growing environment by utilizing fuzzy control, and the convergence speed is improved.
Drawings
FIG. 1 is a schematic diagram of an inventive neural network prediction model;
FIG. 2 is a schematic diagram of a fuzzy controller design;
FIG. 3 is a schematic diagram of membership function (the membership design of the first and second fuzzy controllers is the same);
FIG. 4 is a schematic diagram of an improved model;
FIG. 5 is a graph of the effect of the fit;
fig. 6 is a prediction effect graph.
Detailed Description
The present invention will be further described with reference to the following examples. The following examples are provided only to aid understanding of the present invention. Modifications to the invention can be made by one skilled in the art without departing from the principles of the invention and are intended to be within the scope of the claims.
For prediction, the plant growth environment is monitored, data detected by a sensor (the period is 1 hour) are input into an improved BP neural network for prediction, and then alarm processing is carried out.
Firstly, the obtained parameters are normalized, so that the neural network has good fitting property to the sample:
Figure BDA0001497687460000031
wherein x' is a normalized value, x is a current value, and xmaxIs the maximum value of the sample, xminIs the sample minimum.
Setting the number of nodes of an input layer of a neural network as A, wherein A represents the number of main factors influencing the growth environment of the dendrobium officinale, namely soil temperature, soil humidity, air temperature, air humidity, illumination, carbon dioxide and time; the number of nodes of an output layer is B, and B represents the number of main factors influencing the prediction of the growth environment of the dendrobium officinale (generally, A is set to be B); the hidden layer node C is dependent on the requirements of the input and output.
In order to shorten the training time of the network, a serial input mode is adopted, N training data with known samples are input, and the (N + 1) th data are used for training for a teacher. After finishing, the input data are sequentially moved backwards by one data to be the next group of input data, the N +2 th data are taken as the instructor for training, and so on. Therefore, the prediction rule of the neural network can be mastered, and the prediction model is as follows:
Dn+1＝F(Dn，Dn-1，Dn-2，Dn-3，...，Dn-m，)
F(Dn) Representing a function of predictive mapping of historical data through a neural network, DnRepresenting the value of the environmental parameter at time n.
Step two, the improved BP neural network comprises the following steps:
firstly, processing data with larger errors, preventing the errors caused by randomness from influencing the whole prediction process, and secondly, processing other data errors by utilizing a secondary fuzzy controller.
1) The first-stage fuzzy controller processes an input error signal e (n): inputting the error caused by the convergence of the neural network into a first-stage fuzzy controller, comparing the error with the last error e (n-1), and judging whether the difference (e (n) -e (n-1)) of the errors is convergent or divergent; the method is characterized in that the method comprises the following steps of dividing a set membership function into seven states of error high-speed rising PB1, error medium-speed rising PM1, error low-speed rising PS1, error stable 01, error low-speed falling NS1, error medium-speed falling NM1 and error high-speed falling NB1, carrying out fuzzy processing, and correspondingly changing learning rate and momentum according to a designed rule base so as to adjust the step size:
if e (n) is in a PB1 state, the learning rate is appropriately reduced, the momentum is set to 0.01, the iteration is cancelled, and the previous iteration is returned;
if e (n) is in a PM1 state, the learning rate and the momentum are properly reduced, and the next iteration is continued;
if e (n) is in PS1 state, then the learning rate and momentum are properly reduced, and the next iteration is continued;
if e (n) is in a state of 01, the learning rate and the momentum are kept unchanged, and the next iteration is continued;
if e (n) is NS1 state, then increasing learning rate and momentum properly, continuing next iteration;
sixthly, if e (n) is NM1 state, the learning rate and momentum are increased properly, and the next iteration is continued;
if e (n) is NB1 state, then the learning rate and momentum are increased appropriately, and the next iteration is continued.
2) Rate of change of input error by two-stage fuzzy controller
Figure BDA0001497687460000051
And (3) processing: converting the error processed by the first-stage fuzzy controller into an error change rate, inputting the error change rate into the second-stage fuzzy controller, dividing the error into seven states of high-speed rising PB2 of the error rate, medium-speed rising PM2 of the error rate, low-speed rising PS2 of the error rate, stable 02 of the error rate, low-speed falling NS2 of the error rate, medium-speed falling NM2 of the error rate and high-speed falling NB2 of the error rate through set membership functions, and changing the learning rate and momentum correspondingly again so as to adjust the step size:
if delta is in a PB2 state, the learning rate is properly reduced, momentum is set to be 0.01, the iteration is cancelled, and the previous iteration is returned;
if delta is in a PM2 state, the learning rate and momentum are properly reduced, and the next iteration is continued;
if delta is in a PS2 state, properly reducing the learning rate and momentum, and continuing to perform the next iteration;
if the delta is in the 02 state, keeping the learning rate and the momentum unchanged, and continuing to iterate the next time;
if delta is NS2 state, then increasing learning rate and momentum properly, continuing next iteration;
sixthly, if the delta is the NM2 state, the learning rate and the momentum are properly increased, and the next iteration is continued;
if delta is NB2 state, then the learning rate and momentum are properly increased and the next iteration is continued.
The design of membership function for error and error rate is shown in FIG. 3, and can be processed as seven-value signals of { -3, -2, -1, 0, 1, 2, 3} which respectively represent seven states of PB1/PB2, PM1/PM2, PS1/PS2, 01/02, NS1/NS2, NM1/NM2, and NB1/NB 2. The same approach of the first-level fuzzy rule base design and the second-level fuzzy rule base design is adopted here as shown in table 1.
TABLE 1 learning Rate and momentum Change rule base for error/error Rate states
Figure BDA0001497687460000052
The final improved BP neural network model structure is shown in FIG. 4, firstly, the parameters of the growing environment of the Dendrobium officinale Kimura et Migo are input into the BP neural network, then the predicted and actual measurement errors are input into the first-level fuzzy controller, the error of the secondary calculation is derived and input into the second-level fuzzy controller, further the momentum and the learning rate are adaptively changed, and finally the result is output into the neural network, so that the change of the growing environment is predicted.
In order to test the effect of the improved BP neural network, the traditional BP neural network and other improved BP neural networks, the detected data of the same group of dendrobium officinale growing environment are respectively input into each network, and the result prediction comparison is shown in table 2 below.
TABLE 2 prediction results
Figure BDA0001497687460000061
The result shows that the BP neural network has good error convergence effect and higher speed.
And (4) predicting results: the quality of a neural network is evaluated, and the method can be seen from both the fitting and the prediction of the neural network. The quality of the fitting is determined by the training method, and the predicted result is determined by the quality of the fitting. If the parameter settings are not reasonable enough to cause overfitting, the accuracy of the prediction will be degraded. As shown in fig. 5 and fig. 6, which are respectively a fitting graph and a prediction graph, 1920 groups of data obtained in 20 days are sampled at equal intervals (period 1 hour), 225 groups of data are obtained after error data are eliminated, the front 125 groups of data are used as training data, and the rear 100 groups of data are used as detection data, wherein a solid line represents actual measurement data, and a dotted line represents a predicted value thereof, so that the model can predict a growth environment parameter change 1 hour ahead, and accordingly, corresponding measures are taken for the growth environment parameter change.

Claims (1)

1. A method for predicting the growth environment of dendrobium officinale is characterized by comprising the following steps:
step one, a growth environment prediction model:
firstly, the obtained parameters are normalized, so that the neural network has good fitting property to the sample:
Figure FDA0001497687450000011
wherein x' is a normalized value, x is a current value, and xmaxIs the maximum value of the sample, xminIs the sample minimum;
setting the number of nodes of an input layer of a neural network as A, wherein A represents the number of main factors influencing the growth environment of the dendrobium officinale, namely soil temperature, soil humidity, air temperature, air humidity, illumination, carbon dioxide and time; the number of nodes of an output layer is B, wherein B represents the number of main factors influencing the prediction of the growth environment of the dendrobium officinale, and A is generally set to be B; the hidden layer node C is dependent on the requirements of input and output;
in order to shorten the training time of the network, a serial input mode is adopted, N training data with known samples are input, and the (N + 1) th data are used for training for a teacher; after finishing, the input data are sequentially moved backwards by one data to serve as the next group of input data, the (N + 2) th data serve as a guide to train, and the like; therefore, the prediction rule of the neural network can be mastered, and the prediction model is as follows:
Dn+1＝F(Dn，Dn-1，Dn-2，Dn-3，...，Dn-m)
F(Dn) Representing a function of predictive mapping of historical data through a neural network, DnA value representing an environmental parameter at time n;
step two, the improved BP neural network comprises the following steps:
firstly, processing data, namely processing data with larger errors, and secondly, processing other data errors by using a secondary fuzzy controller;
1) the first-stage fuzzy controller processes an input error signal e (n): inputting the error caused by the convergence of the neural network into a first-stage fuzzy controller, comparing the error with the last error e (n-1), and judging whether the difference (e (n) -e (n-1)) of the errors is convergent or divergent; the method is characterized in that the method comprises the following steps of dividing a set membership function into seven states of error high-speed rising PB1, error medium-speed rising PM1, error low-speed rising PS1, error stable 01, error low-speed falling NS1, error medium-speed falling NM1 and error high-speed falling NB1, carrying out fuzzy processing, and correspondingly changing learning rate and momentum according to a designed rule base so as to adjust the step size:
if e (n) is in a PB1 state, the learning rate is appropriately reduced, the momentum is set to 0.01, the iteration is cancelled, and the previous iteration is returned;
if e (n) is in a PM1 state, the learning rate and the momentum are properly reduced, and the next iteration is continued;
if e (n) is in PS1 state, then the learning rate and momentum are properly reduced, and the next iteration is continued;
if e (n) is in a state of 01, the learning rate and the momentum are kept unchanged, and the next iteration is continued;
if e (n) is NS1 state, then increasing learning rate and momentum properly, continuing next iteration;
sixthly, if e (n) is NM1 state, the learning rate and momentum are increased properly, and the next iteration is continued;
if e (n) is the state of NB1, the learning rate and the momentum are properly increased, and the next iteration is continued;
2) rate of change of input error by two-stage fuzzy controller
Figure FDA0001497687450000021
And (3) processing: converting the error processed by the first-stage fuzzy controller into error change rate, inputting the error change rate into the second-stage fuzzy controller, and dividing the error into PB2 with high-speed rise of error rate, PM2 with medium-speed rise of error rate, PS2 with low-speed rise of error rate and error through the set membership functionSeven states of rate stabilization 02, error rate low speed drop NS2, error rate medium speed drop NM2 and error rate high speed drop NB2, corresponding changes are made again to the learning rate and momentum, so that the step size is adjusted:
if delta is in a PB2 state, the learning rate is properly reduced, momentum is set to be 0.01, the iteration is cancelled, and the previous iteration is returned;
if delta is in a PM2 state, the learning rate and momentum are properly reduced, and the next iteration is continued;
if delta is in a PS2 state, properly reducing the learning rate and momentum, and continuing to perform the next iteration;
if the delta is in the 02 state, keeping the learning rate and the momentum unchanged, and continuing to iterate the next time;
if delta is NS2 state, then increasing learning rate and momentum properly, continuing next iteration;
sixthly, if the delta is the NM2 state, the learning rate and the momentum are properly increased, and the next iteration is continued;
if delta is NB2 state, then the learning rate and momentum are properly increased and the next iteration is continued.
CN201711282227.XA 2017-12-07 2017-12-07 Dendrobium officinale growth environment prediction method Active CN108133282B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711282227.XA CN108133282B (en) 2017-12-07 2017-12-07 Dendrobium officinale growth environment prediction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711282227.XA CN108133282B (en) 2017-12-07 2017-12-07 Dendrobium officinale growth environment prediction method

Publications (2)

Publication Number Publication Date
CN108133282A CN108133282A (en) 2018-06-08
CN108133282B true CN108133282B (en) 2021-03-30

Family

ID=62390151

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711282227.XA Active CN108133282B (en) 2017-12-07 2017-12-07 Dendrobium officinale growth environment prediction method

Country Status (1)

Country Link
CN (1) CN108133282B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109324503B (en) * 2018-08-28 2022-02-15 南京理工大学 Multilayer neural network motor system control method based on robust integration
CN112666246A (en) * 2020-12-17 2021-04-16 贵州中医药大学 Method for screening dendrobium officinale cultivated by imitating wild rock fissure epiphytic growth
CN118247532A (en) * 2024-05-27 2024-06-25 贵州航天智慧农业有限公司 Intelligent monitoring and regulating method and system for plant growth environment

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6128609A (en) * 1997-10-14 2000-10-03 Ralph E. Rose Training a neural network using differential input
CN104835103A (en) * 2015-05-11 2015-08-12 大连理工大学 Mobile network health evaluation method based on neural network and fuzzy comprehensive evaluation
CN105701280A (en) * 2016-01-05 2016-06-22 浙江大学城市学院 Southern America white-leg shrimp pond culture water quality prediction method
CN107255920A (en) * 2017-06-21 2017-10-17 武汉理工大学 PID control method and apparatus and system based on network optimization algorithm

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9798751B2 (en) * 2013-10-16 2017-10-24 University Of Tennessee Research Foundation Method and apparatus for constructing a neuroscience-inspired artificial neural network

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6128609A (en) * 1997-10-14 2000-10-03 Ralph E. Rose Training a neural network using differential input
CN104835103A (en) * 2015-05-11 2015-08-12 大连理工大学 Mobile network health evaluation method based on neural network and fuzzy comprehensive evaluation
CN105701280A (en) * 2016-01-05 2016-06-22 浙江大学城市学院 Southern America white-leg shrimp pond culture water quality prediction method
CN107255920A (en) * 2017-06-21 2017-10-17 武汉理工大学 PID control method and apparatus and system based on network optimization algorithm

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
Treatment of missing data using neural networks and genetic algorithms;ABDELLA M;《Proceedings of the international joint conference on neural networks （IJCNN）》;20051231;全文 *
模糊方法改进的反向传输神经网络预测南美白对虾养殖的水质;丁金婷;《浙江大学学报（农业与生命科学版）》;20170122;全文 *

Also Published As

Publication number Publication date
CN108133282A (en) 2018-06-08

Similar Documents

Publication Publication Date Title
CN108133282B (en) Dendrobium officinale growth environment prediction method
CN101576734B (en) Dissolved oxygen control method based on dynamic radial basis function neural network
CN108898215B (en) Intelligent sludge bulking identification method based on two-type fuzzy neural network
CN110687800B (en) Data-driven self-adaptive anti-interference controller and estimation method thereof
CN108920812B (en) Machining surface roughness prediction method
WO2023088212A1 (en) Online unit load prediction method based on ensemble learning
CN111182564B (en) Wireless link quality prediction method based on LSTM neural network
CN110826791A (en) Hybrid wind power prediction method based on long-time and short-time memory neural network
CN105354620A (en) Method for predicting fan generation power
CN110119086B (en) Tomato greenhouse environmental parameter intelligent monitoring device based on ANFIS neural network
CN101639902B (en) Modeling method of support vector machine (SVM)-based software measurement instrument in biological fermentation process
CN102411308A (en) Adaptive control method of dissolved oxygen (DO) based on recurrent neural network (RNN) model
CN106920014A (en) A kind of short-term load forecasting method and device
CN204595644U (en) Based on the aluminum-bar heating furnace temperature of combustion automaton of neural network
CN107609718A (en) Method and system for predicting dissolved oxygen in aquaculture water
CN114036813A (en) Greenhouse temperature and humidity method controlled by particle swarm BP neural network PID
CN108710964A (en) A kind of prediction technique of Fuzzy time sequence aquaculture water quality environmental data
CN101963785B (en) On-line control method for filtering process of oxidation mother liquor in production of purified terephthalic acid
CN111125907B (en) Sewage treatment ammonia nitrogen soft measurement method based on hybrid intelligent model
CN113219871A (en) Curing room environmental parameter detecting system
CN103792845A (en) Method and system for carbohydrate supplementation speed optimal control in fermentation process of aureomycin
CN109408896B (en) Multi-element intelligent real-time monitoring method for anaerobic sewage treatment gas production
CN116661517B (en) Compound microbial fertilizer fermentation temperature intelligent regulation and control system based on thing networking
CN104503226A (en) Wireless Wi-Fi remote monitoring system based on multi-sensor information fusion in fermentation chamber environment
CN111754033A (en) Non-stationary time sequence data prediction method based on recurrent 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
GR01 Patent grant
GR01 Patent grant
TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20220707

Address after: 310015 No. 51, Huzhou street, Hangzhou, Zhejiang

Patentee after: HANGZHOU City University

Address before: 310015 No. 50 Huzhou Street, Hangzhou City, Zhejiang Province

Patentee before: Zhejiang University City College