CN110414115B - Wavelet neural network tomato yield prediction method based on genetic algorithm - Google Patents
Wavelet neural network tomato yield prediction method based on genetic algorithm Download PDFInfo
- Publication number
- CN110414115B CN110414115B CN201910663789.1A CN201910663789A CN110414115B CN 110414115 B CN110414115 B CN 110414115B CN 201910663789 A CN201910663789 A CN 201910663789A CN 110414115 B CN110414115 B CN 110414115B
- Authority
- CN
- China
- Prior art keywords
- model
- value
- neural network
- input
- output
- 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
- 235000007688 Lycopersicon esculentum Nutrition 0.000 title claims abstract description 35
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 27
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 27
- 230000002068 genetic effect Effects 0.000 title claims abstract description 25
- 238000000034 method Methods 0.000 title claims abstract description 17
- 240000003768 Solanum lycopersicum Species 0.000 title abstract description 33
- 238000013519 translation Methods 0.000 claims abstract description 16
- 238000012549 training Methods 0.000 claims abstract description 11
- 238000012216 screening Methods 0.000 claims abstract description 7
- 238000005457 optimization Methods 0.000 claims abstract description 6
- 238000012545 processing Methods 0.000 claims abstract description 6
- 230000006870 function Effects 0.000 claims description 24
- 230000000694 effects Effects 0.000 claims description 17
- 238000003062 neural network model Methods 0.000 claims description 12
- 238000003973 irrigation Methods 0.000 claims description 6
- 230000002262 irrigation Effects 0.000 claims description 6
- 108090000623 proteins and genes Proteins 0.000 claims description 5
- ZLMJMSJWJFRBEC-UHFFFAOYSA-N Potassium Chemical compound [K] ZLMJMSJWJFRBEC-UHFFFAOYSA-N 0.000 claims description 3
- 230000003044 adaptive effect Effects 0.000 claims description 3
- 239000003337 fertilizer Substances 0.000 claims description 3
- 238000005286 illumination Methods 0.000 claims description 3
- 239000000618 nitrogen fertilizer Substances 0.000 claims description 3
- 239000002686 phosphate fertilizer Substances 0.000 claims description 3
- 229910052700 potassium Inorganic materials 0.000 claims description 3
- 239000011591 potassium Substances 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 230000035772 mutation Effects 0.000 claims description 2
- 241000227653 Lycopersicon Species 0.000 claims 2
- 238000011478 gradient descent method Methods 0.000 claims 1
- 230000005284 excitation Effects 0.000 abstract description 3
- 238000005259 measurement Methods 0.000 abstract description 2
- 238000011160 research Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000004720 fertilization Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 239000013589 supplement Substances 0.000 description 2
- 238000009423 ventilation Methods 0.000 description 2
- 238000003967 crop rotation Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 235000021049 nutrient content Nutrition 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 239000002689 soil Substances 0.000 description 1
- 241000894007 species Species 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
- 235000013311 vegetables Nutrition 0.000 description 1
- 238000009333 weeding Methods 0.000 description 1
Images
Classifications
-
- 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
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Economics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Tourism & Hospitality (AREA)
- Evolutionary Biology (AREA)
- Quality & Reliability (AREA)
- Genetics & Genomics (AREA)
- Physiology (AREA)
- Primary Health Care (AREA)
- Mining & Mineral Resources (AREA)
- Marine Sciences & Fisheries (AREA)
- Animal Husbandry (AREA)
- Agronomy & Crop Science (AREA)
- Operations Research (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
Abstract
The invention discloses a wavelet neural network tomato yield prediction method based on a genetic algorithm, and belongs to the technical field of agricultural information. The method comprises the steps of selecting a certain amount of parameters as input variables, classifying, processing and screening the parameters, and performing optimization training on the parameters by using a genetic algorithm to obtain a proper data population. Then determining the number of nodes of the input layer, the hidden layer and the output layer of the network, replacing an excitation function of a node of the hidden layer of the BP neural network with a wavelet basis function, continuously adjusting the weight by introducing a translation factor and a scale factor, continuously reducing the error of model prediction through multiple times of training and iteration, continuously enabling the output result to be close to an actual measurement value, controlling the error within a reasonable range, improving the prediction precision and function convergence performance of the model, and providing reliable theoretical support for predicting the crop yield in the greenhouse.
Description
Technical Field
The invention belongs to the technical field of agricultural information, and particularly relates to a wavelet neural network tomato yield prediction method based on a genetic algorithm.
Background
The tomatoes are one of main vegetables for facility cultivation in northern areas of China, the cultivation mode of efficient crop rotation is mostly adopted for the tomato cultivation in sunlight greenhouses in the northern areas, and the environmental factors of the greenhouses, the nutrient content of soil and the physiological characteristics of the tomatoes directly determine the yield of the tomatoes. The indoor tomato yield is measured in advance, measures such as deinsectization, weeding and the like can be taken timely for determining a proper planting plan, reasonably arranging fertilizing amount and irrigation amount, and measures such as ventilation, temperature increase, light supplement and the like are taken regularly, so that a good environment is created for healthy growth of tomatoes, and a foundation is laid for high quality and high yield of greenhouse tomatoes. The previous researches on tomato yield prediction are very extensive, but most of the researches are focused on applying a crop mechanism model or a crop growth and development model to predict the tomato yield in a greenhouse, the related parameters are relatively few, and the calculation method has certain limitations. The method for predicting the greenhouse tomato yield by combining the genetic algorithm with the wavelet neural network can improve the model prediction precision by improving and optimizing the model through continuously correcting various parameter variables in the model according to historical data of the past year, is more suitable for the actual environment of the northern sunlight greenhouse, and has a wider application prospect.
Disclosure of Invention
The invention provides a wavelet neural network tomato yield prediction method based on a genetic algorithm, which can predict the tomato yield in a northern energy-saving sunlight greenhouse and can determine an appropriate planting plan.
Wavelet neural network tomato based on genetic algorithmThe yield prediction model is characterized in that input variables of the model comprise the following characteristic parameters: ambient temperature; ambient humidity; irrigation quantity; the input amount of the nitrogen fertilizer; the input amount of phosphate fertilizer; adding potassium fertilizer; CO2 2 Concentration; the intensity of illumination;
the model is built by the following steps:
randomly generating an initial population X m×n ：
n＝s 1 ×s 2 +s 2 ×s 3 +s 2 +s 3 (1)
Wherein: m is the initial population number, n is the individual length, and the individual length represents the number of gene values of each individual and also represents the initial weight number of a neural network; s 1 Number of nodes of input layer; s 2 Number of hidden layer nodes; s 3 Number of output layer nodes;
the genetic algorithm calculates the output error value E of each individual in the initial population i Fitness value f i And evaluating according to the value of the individual fitness value, selecting the individuals with the fitness value within a preset range A in the initial population to enter a sub-population for continuous optimization training:
in the sub-population, the probability of the i-th individual performing crossover or mutation operations is p i According to the crossing rate p c And the rate of variation p m An adaptive function to determine whether the individual needs to be cross-or genetically manipulated:
in the formula: k is a radical of c 、k m Real numbers, f, both being less than 1 c As individual fitness value to be crossed, f m As the individual fitness value to be mutated, f max 、
Respectively a maximum fitness value and an average fitness value in the population,
the convergence degree of the population;
eliminating data with deviation exceeding the annual average value by +/-10% in input variables, optimizing the data by applying a genetic algorithm, considering that the processing effect is better when the dynamic range of the data is less than or equal to +/-5%, and taking the data after screening and optimizing as input data of a wavelet neural network and carrying out subsequent operation;
the processed characteristic parameters are used as the input of a wavelet neural network model, and the greenhouse tomato yield is used as the output, wherein the method comprises the following steps:
the mother wavelet function selected by the model is as follows:
g(x)＝cos(1.75x)exp(-x 2 /2) (7)
and (3) carrying out scale and translation transformation on the mother wavelet function to construct a wavelet basis function:
wherein, a j 、b j Respectively is the scale factor and the translation factor of the jth hidden layer node;
y k for the output of the model:
wherein x is i (I =1,2.., I) is the input to the I-th node of the input layer, y k (j =1,2,... K) is the output of the j-th node of the output layer;
defining the error function E as:
wherein, y k Is the actual output of the kth node of the output layer, t k Target output for kth node of output layer;
the adjustment of the weight, the scale factor and the translation factor comprises the following steps:
wherein, the learning rate: η (η > 0), momentum factor: mu (0 < mu < 1).
More preferably, the error value is expressed as a percentage, and the fitness value is in the range of 0 to 1, a = f i ＞0.667。
The invention provides a wavelet neural network tomato yield prediction method based on a genetic algorithm, which is used for classifying, processing and screening data and optimally training the data by applying the genetic algorithm to obtain a proper data population. Then determining the number of nodes of the network input layer, the hidden layer and the output layer, replacing the excitation function of the hidden layer node of the BP neural network with a wavelet basis function, continuously adjusting the weight by introducing a translation factor and a scale factor, continuously reducing the error of model prediction through multiple times of training and iteration, enabling the output result to be continuously close to the measured value, and controlling the error within a reasonable range.
Drawings
FIG. 1 is a wavelet neural network structure diagram of a wavelet neural network tomato yield prediction method based on genetic algorithm provided by the present invention;
FIG. 2 is a graph of the predicted effect of the GA-WNN model;
FIG. 3 is a graph of the percentage error of the GA-WNN model;
FIG. 4 is a WNN model prediction effect graph;
FIG. 5 is a WNN model error percentage curve;
FIG. 6 is a BP neural network model prediction effect;
FIG. 7 is a BP neural network model error percentage curve;
FIG. 8 is a flow chart of a wavelet neural network tomato yield prediction method based on a genetic algorithm provided by the invention.
Detailed Description
An embodiment of the present invention will be described in detail below with reference to the accompanying drawings, but it should be understood that the scope of the present invention is not limited to the embodiment.
The tomato yield prediction model establishing process comprises the following steps: firstly, classifying, processing and screening data and optimally training the data by using a genetic algorithm to obtain a proper data population. Then, the number of nodes of the network input layer, the hidden layer and the output layer is determined, a wavelet basis function is used for replacing an excitation function of a BP neural network hidden layer node, a translation factor and a scale factor are introduced to continuously adjust the weight, the error of model prediction is continuously reduced through multiple training and iteration, the output result is continuously close to an actual measurement value, the error is controlled within a reasonable range, and the specific flow is shown in fig. 8.
The GA-WNN algorithm specifically comprises the following steps:
the model selects the following parameters as input variables: (1) ambient temperature; (2) ambient humidity; (3) irrigation quantity; (4) the input amount of the nitrogen fertilizer; (5) the input amount of phosphate fertilizer; (6) adding potassium fertilizer; (7) the concentration of CO 2; (8) illumination intensity, tomato yield data was used as the output variable for this model. Firstly, a genetic algorithm is used as a prepositive means of a WNN algorithm, original data are optimized to obtain data with a small dynamic range (the dynamic range of the data is less than or equal to +/-5%), and the data after screening and optimization are used as initial values of a subsequent WNN algorithm. And then, improving the traditional BP neural network by applying a wavelet function, combining the characteristics of multi-resolution analysis of the wavelet function with the self-adaptive capacity of the neural network, constructing a wavelet neural network model, and continuously correcting and perfecting the wavelet neural network model, so that the prediction precision and the function convergence performance of the model are greatly improved, and a reliable theoretical support is provided for predicting the crop yield in the greenhouse.
The specific operation process is as follows:
(1) And (4) screening and optimizing the selected 8 parameters by using a genetic algorithm. First, an initial population X is randomly generated m×n The method comprises the steps of optimizing original data, wherein m is an initial population number (since 8 variables are selected, m =8 is selected), the length n of each individual represents the number of gene values of each individual and the initial weight number of a neural network (the number of hidden layer nodes is continuously changed along with network training, the weight number is changed accordingly, initially 8 input layer nodes, 5 hidden layer nodes, 1 output layer node and the initial weight number is 46), and the gene values in the individual correspond to the initial weight of the neural network one by one. The research adopts a real number coding mode to code the gene value, can avoid a decoding process and improve the training efficiency.
n＝s 1 ×s 2 +s 2 ×s 3 +s 2 +s 3 (1)
In the formula: n is the individual length; s 1 Number of nodes of input layer; s 2 Number of nodes of hidden layer; s 3 To be transportedThe number of egress nodes.
The genetic algorithm calculates the output error value E of each individual in the initial population i Fitness value f i Evaluating according to the value of the individual fitness value, selecting the fitness value in the initial population to be larger (the error value is expressed by percentage, the fitness value is in the range of 0-1, and f i Values > 0.667 are regarded as values with better fitness) enter the sub-population to continue optimization training.
In the sub-population, the probability that the ith individual selected will be crossed or mutated is p i And according to the crossing rate p c And the rate of variation p m Adaptive function to determine whether the individual needs to be cross-or genetically manipulated, p c And p m Will be adapted to the value f according to the individual i The fitness is changed, so that the population is always diversified.
In the formula: k is a radical of c 、k m Real numbers, f, both being less than 1 c As individual fitness value to be crossed, f m As the individual fitness value to be mutated, f max 、
Respectively a maximum fitness value and an average fitness value in the population,
the convergence of the population. According to the genetic algorithm, the original data is screened and optimized, the large deviation is eliminated (a variable exceeds the annual average value +/-10% and is regarded as a large deviation value, the value is restricted by the upper limit and the lower limit, and is specified as a boundary value for subsequent calculation), the data is optimized by applying the genetic algorithm, the dynamic range of the data can be reduced, the processing effect is good when the dynamic range of the data is less than or equal to +/-5%, and the screened and optimized data is used as input data of a wavelet neural network for subsequent calculation.
(2) The 8 characteristic parameters processed and optimized in the previous step are used as input of a Wavelet Neural Network (WNN) model, the yield of the greenhouse tomatoes is used as output, and the structure of the network is important for the prediction precision and stability of the model. When the model is constructed, the structure of the model is considered firstly, so that the model can meet the test requirements, the stability and the reliability of the model are ensured, then the problem existing in the error function analysis is applied, the model is continuously optimized, and the prediction precision is improved.
The mother wavelet function selected by the model is as follows:
g(x)＝cos(1.75x)exp(-x 2 /2) (7)
and (3) carrying out scale and translation transformation on the mother wavelet function to construct a wavelet basis function:
wherein, a j 、b j Scale factors and translation factors of the jth hidden layer node, respectively. y is k For the output of the model:
x i (I =1,2.., I) -input of the ith node of the input layer, y k (j =1,2.., k) -output of the jth node of the output layer.
The stability and the prediction precision of the model are improved by continuously adjusting the weight, the scale factor and the translation factor in the model, so that the model is more suitable for the actual field environment and is continuously perfected and optimized. Defining the error function E as:
wherein, y k Is the actual output of the kth node of the output layer, t k Is the target output of the kth node of the output layer. The adjustment of the weight, the scale factor and the translation factor comprises the following steps:
wherein, the learning rate: η (η > 0), momentum factor: mu (0 < mu < 1). Parameters in the formula are continuously adjusted, so that the prediction precision of the model is improved, MATLAB software is adopted for simulation, and the prediction effect of the model is verified.
And then, the stability and the prediction precision of the model are improved by continuously adjusting parameters such as GA parameters, BP parameters, weights, scale factors, translation factors and the like in the model, so that the model is more suitable for the actual field environment, and the model is continuously perfected and optimized.
In the test, 2010-2015 data are selected to train the model, and 2016-2018 data are used to test the actual prediction effect of the GA-WNN model. The network is trained by using data of 6 years, model parameters such as weight, scale factor, translation factor, learning rate, momentum factor and iteration times of the model are continuously adjusted, the model with a better prediction effect is obtained by continuously improving and optimizing the model after comparing test results and calculating errors for multiple times, the error is in a reasonable range (the error percentage is within 5 percent), the difference between a predicted value and an actual measured value is very small, accurate prediction is basically realized on the yield of the greenhouse tomatoes, and the prediction effect of the GA-WNN model is shown as follows. The result shows that the model prediction value has a good tracking effect on the measured value, the model has high stability, and the greenhouse tomato yield can be effectively predicted. The experimental data and experimental results are shown in the following graph:
in order to verify the prediction effect of the model, the GA-WNN model is compared with the WNN model and the BP neural network model, and the comparison condition of the prediction effect is shown in the following table. The analysis shows that the predicted values and the actual values of the two models have certain errors, the average relative error of the GA-WNN model is 0.66%, the average relative error of the WNN model is 1.02%, the average relative error of the BP neural network model is 2.42%, and the prediction effect of the GA-WNN model is ideal. The GA-WNN model has a convergence speed superior to that of the WNN and BP neural network models, and the optimal prediction effect is achieved through 208 steps. In conclusion, the GA-WNN model constructed by combining the genetic algorithm, the wavelet analysis and the BP neural network has high convergence rate and high prediction precision, so that the accurate prediction of the tomato yield is realized, and a basis can be provided for reasonably arranging crop species in a greenhouse, making management plans such as irrigation and fertilization and the like.
According to the research, the GA-WNN model is used for predicting the tomato yield in the northern sunlight greenhouse, factors influencing the greenhouse tomato yield are comprehensively considered, parameters are processed and screened by means of genetic algorithms and the like, then the characteristic vectors of data are extracted to serve as model input, a greenhouse tomato yield prediction model is constructed, weight adjustment is carried out by using scale factors and translation factors in a wavelet neural network, and the prediction accuracy of the model is improved. Simulation results show that the average relative error of the GA-WNN model is 0.66%, the average relative error of the WNN model is 1.02%, the average relative error of the BP neural network model is 2.42%, the prediction effect of the GA-WNN model is superior to that of the WNN and BP neural network models, the optimal prediction effect is achieved after 208 steps, and the GA-WNN model is proved to have good practical application value. Compared with the traditional WNN and BP neural network models, the GA-WNN model has the advantages of small average relative error, higher precision and stronger stability, is more consistent with greenhouse tomato yield prediction, can provide basis for making tomato planting decisions in a greenhouse, provides certain theoretical support for measures such as irrigation, fertilization, light supplement and ventilation adopted during tomato planting, and plays a role in promoting improvement and optimization of a greenhouse system.
The above disclosure is only for a few specific embodiments of the present invention, however, the present invention is not limited to the above embodiments, and any variations that can be made by those skilled in the art are intended to fall within the scope of the present invention.
Claims (2)
1. A wavelet neural network tomato yield prediction model based on a genetic algorithm is characterized in that input variables of the model comprise the following characteristic parameters: ambient temperature; ambient humidity; irrigation quantity; the input amount of the nitrogen fertilizer; the input amount of phosphate fertilizer; adding potassium fertilizer; CO2 2 Concentration; the illumination intensity;
the model is built by the following steps:
randomly generating an initial population X m×n ：
n＝s 1 ×s 2 +s 2 ×s 3 +s 2 +s 3 (1)
Wherein: m is the initial population number, n is the individual length, and the individual length represents the number of gene values of each individual and also represents the initial weight number of a neural network; s 1 Number of nodes of input layer; s 2 Number of nodes of hidden layer; s 3 The number of nodes of the output layer;
the genetic algorithm calculates the output error value E of each individual in the initial population i Fitness value f i And evaluating according to the value of the individual fitness value, selecting the individuals with the fitness value within a preset range A in the initial population to enter a sub-population for continuous optimization training:
in the sub-population, the probability of the i-th individual performing crossover or mutation operations is p i According to the crossing rate p c And the rate of variation p m An adaptive function to determine whether the individual needs to be cross-or genetically manipulated:
in the formula: k is a radical of c 、k m Are all less thanReal number of 1, f i Adapting the value for the individual; f. of c As individual fitness value to be crossed, f m As the individual fitness value to be mutated, f max 、
Respectively a maximum fitness value and an average fitness value in the population,
the convergence degree of the population;
eliminating data with deviation exceeding the annual average value by +/-10% in input variables, optimizing the data by applying a genetic algorithm, considering that the processing effect is better when the absolute value of the dynamic range of the data is within 5%, and taking the data after screening and optimizing as input data of a wavelet neural network and carrying out subsequent operation;
the processed characteristic parameters are used as input of a wavelet neural network model, and the greenhouse tomato yield is used as output, wherein the method comprises the following steps:
the mother wavelet function selected by the model is as follows:
g(x)＝cos(1.75x)exp(-x 2 /2) (7)
x is an independent variable of the mother wavelet function and represents an input variable of the yield prediction model;
and (3) carrying out scale and translation transformation on the mother wavelet function to construct a wavelet basis function:
wherein, a j 、b j Respectively is the scale factor and the translation factor of the jth hidden layer node;
y k for the output of the model:
wherein x is i (I =1,2.., I) is the input to the I-th node of the input layer, y k (j =1,2,... K) is the output of the j-th node of the output layer;
the predicted yield value can be obtained according to the formula (9) and compared with the actual yield value, so as to calculate the prediction error of the network, and an error function E is defined as:
wherein, y k Is the actual output of the kth node of the output layer, t k Target output of the kth node of the output layer;
adjusting the weight W from the input layer to the hidden layer according to the gradient descent method ji And weight W from hidden layer to output layer kj Scale factor a j (n + 1), translation factor b j (n + 1), the formula is as follows:
in equations (11) - (14), wkj (n + 1) and Wji (n + 1) represent the current time weight, wkj (n) and Wji (n) represent the last time weight, and the learning rate: η (η > 0), momentum factor: mu (0 < mu < 1).
2. The method as claimed in claim 1, wherein the error value is expressed as a percentage, the fitness value is in the range of 0 to 1, and A = f i ＞0.667。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910663789.1A CN110414115B (en) | 2019-07-13 | 2019-07-13 | Wavelet neural network tomato yield prediction method based on genetic algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910663789.1A CN110414115B (en) | 2019-07-13 | 2019-07-13 | Wavelet neural network tomato yield prediction method based on genetic algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110414115A CN110414115A (en) | 2019-11-05 |
| CN110414115B true CN110414115B (en) | 2023-01-20 |
Family
ID=68362450
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910663789.1A Active CN110414115B (en) | 2019-07-13 | 2019-07-13 | Wavelet neural network tomato yield prediction method based on genetic algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110414115B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113158470B (en) * | 2020-11-25 | 2022-09-23 | 中国石油大学(华东) | Oil reservoir automatic history fitting system and method based on transfer learning |
| CN113379188B (en) * | 2021-05-06 | 2022-10-25 | 贵州省烟草公司贵阳市公司 | Tobacco crop rotation planting method and system based on Internet of things |
| CN113439520A (en) * | 2021-07-21 | 2021-09-28 | 中国农业科学院农业环境与可持续发展研究所 | Intelligent decision-making method and system for crop irrigation and fertilization |
| CN115907366B (en) * | 2022-11-15 | 2023-09-22 | 淮阴工学院 | Agricultural product growth environment optimal regulation and control method and equipment based on flamingo algorithm |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106651012A (en) * | 2016-12-02 | 2017-05-10 | 东华大学 | Crop transpiration prediction method based on improved extreme learning machine |
| CN109345508A (en) * | 2018-08-31 | 2019-02-15 | 北京航空航天大学 | A kind of Assessing Standards For Skeletal method based on two stages neural network |
| CN109359741A (en) * | 2018-09-27 | 2019-02-19 | 华南师范大学 | A kind of wastewater treatment influent quality timing variations intelligent Forecasting |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10698918B2 (en) * | 2013-11-20 | 2020-06-30 | Qliktech International Ab | Methods and systems for wavelet based representation |
| US10260319B2 (en) * | 2016-02-08 | 2019-04-16 | Rs Energy Group Topco, Inc. | Method for estimating oil/gas production using statistical learning models |
-
2019
- 2019-07-13 CN CN201910663789.1A patent/CN110414115B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106651012A (en) * | 2016-12-02 | 2017-05-10 | 东华大学 | Crop transpiration prediction method based on improved extreme learning machine |
| CN109345508A (en) * | 2018-08-31 | 2019-02-15 | 北京航空航天大学 | A kind of Assessing Standards For Skeletal method based on two stages neural network |
| CN109359741A (en) * | 2018-09-27 | 2019-02-19 | 华南师范大学 | A kind of wastewater treatment influent quality timing variations intelligent Forecasting |
Non-Patent Citations (1)
| Title |
|---|
| 《作物产量预测的时间序列神经网络模型》;魏周会等;《节水灌溉》;20061231;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110414115A (en) | 2019-11-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110414115B (en) | Wavelet neural network tomato yield prediction method based on genetic algorithm | |
| CN106373022B (en) | BP-GA-based greenhouse crop planting efficiency condition optimization method and system | |
| CN106842923B (en) | Greenhouse environment multi-factor coordination control method based on crop physiology and energy consumption optimization | |
| CN107909149A (en) | A kind of Temperature in Greenhouse Forecasting Methodology based on Genetic BP Neutral Network | |
| CN107832892B (en) | Multi-block seed selection decision optimization method based on combinatorial optimization | |
| CN107145941B (en) | Method for dynamically acquiring light demand quantity in real time based on optimal light quality and photon flux density | |
| CN107169621A (en) | A kind of Dissolved Oxygen in Water Forecasting Methodology and device | |
| Morimoto et al. | AI approaches to identification and control of total plant production systems | |
| CN107390754A (en) | Intelligent plant growth environment adjustment system and method based on Internet of Things cloud platform | |
| CN113282122A (en) | Commercial building energy consumption prediction optimization method and system | |
| CN107329511A (en) | Based on the vegetable aquaculture luminous environment high efficiency regulatory method and system between suitable root warm area | |
| CN110472840A (en) | A kind of agricultural water conservancy dispatching method and system based on nerual network technique | |
| CN110070228A (en) | BP neural network wind speed prediction method for neuron branch evolution | |
| Valenzuela et al. | Pre-harvest factors optimization using genetic algorithm for lettuce | |
| CN114859734B (en) | Greenhouse environment parameter optimization decision method based on improved SAC algorithm | |
| Lyu et al. | Multi-objective winter wheat irrigation strategies optimization based on coupling AquaCrop-OSPy and NSGA-III: A case study in Yangling, China | |
| An et al. | A simulator-based planning framework for optimizing autonomous greenhouse control strategy | |
| CN105913326B (en) | Constrain the Crop growing stage model Cultivar parameter optimization method of sex knowledge and the tactful Genetic Algorithm Fusion of elite individual | |
| Li et al. | Determining optimal CO2 concentration of greenhouse tomato based on PSO-SVM | |
| CN114282702A (en) | Soil conditioning time sequence prediction method and system based on IGA-BP neural network | |
| CN117575094A (en) | Crop yield prediction and optimization method and device based on digital twin | |
| CN113221447A (en) | Soil humidity prediction method for optimizing BP neural network based on improved genetic algorithm | |
| Chen et al. | Multi-objective optimal regulation model and system based on whole plant photosynthesis and light use efficiency of lettuce | |
| CN109934400B (en) | Rain collecting, regulating and deficiency crop water demand prediction method based on improved neural network | |
| Chen et al. | A water-saving irrigation decision-making model for greenhouse tomatoes based on genetic optimization TS fuzzy 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 |