=Paper=
{{Paper
|id=Vol-3899/paper16
|storemode=property
|title=A computerized method for predicting the risk of powdery mildew in wheat based on software analysis of soil and climatic monitoring data
|pdfUrl=https://ceur-ws.org/Vol-3899/paper16.pdf
|volume=Vol-3899
|authors=Grygorii Diachenko,Ivan Laktionov,Dmytro Moroz,Maryna Derzhevetska,Sergii Semenov
|dblpUrl=https://dblp.org/rec/conf/advait/DiachenkoLMDS24
}}
==A computerized method for predicting the risk of powdery mildew in wheat based on software analysis of soil and climatic monitoring data==
https://ceur-ws.org/Vol-3899/paper16.pdf
A computerized method for predicting the risk of
powdery mildew in wheat based on software analysis of
soil and climatic monitoring data⋆
Grygorii Diachenko1, *,†, Ivan Laktionov1 ,†, Dmytro Moroz1 ,†, Maryna Derzhevetska2,†,
Sergii Semenov1 ,†
1 Dnipro University of Technology, av. Dmytra Yavornytskoho 19, UA49005, Dnipro, Ukraine
2Technical University “Metinvest Polytechnic”, Southern Highway 80, UA69008, Zaporizhzhia, Ukraine
Abstract
Today, smart agriculture is one of the core technologies for sustainable development and increasing the
efficiency of open-field crop production enterprises of various sizes and forms of ownership in the face of
changing climate conditions. The development and implementation of computerized methods and
intelligent software and hardware solutions for transforming large volumes of agroclimatic data distributed
in time and space is a relevant and important field for improving the efficiency of information technologies
for agrotechnical applications. In this article, the scientific and applied problem of creating and validating
a computerized method for predicting the probability of occurrence of crop diseases at the pre-symptomatic
stage, which forms the basis of software and hardware components for processing data from
agromonitoring systems based on fog architecture, has been solved. The main results of the research are:
reduction of the number of informative features to five based on the Harris Hawk Optimizer algorithm,
proving the effectiveness of Bagged Trees and Medium Neural Network algorithms in the classification of
Powdery Mildew in Wheat, synthesis and testing of a computer model in Simulink that implements a full
cycle of transformation of agroclimatic monitoring data in predicting the Risk of Powdery Mildew in Wheat.
In addition, prospective directions for further research to improve the efficiency of information
technologies for predicting the probability of crop diseases are substantiated in the article.
Keywords1
Classification, soil and climatic parameters, computerized method, prediction, Powdery Mildew Blumeria
Graminis, feature selection, machine learning
1. Introduction
To date, the principles of digitalization and intellectualization of technological processes are one of
the global trends in improving the efficiency of production processes at enterprises of various
profiles, scales and forms of ownership. Agriculture is one of the strategic sectors of the national
economies of many countries, and therefore requires constant search and implementation of
scientifically substantiated approaches to the sustainable development of agricultural practices.
Smart farming is a key concept relevant to running production processes in agricultural enterprises,
particularly open-field crop production. This concept, in turn, is made possible by introducing
technologies such as: Internet of Things, machine learning and artificial intelligence, remote sensing,
drones and robotics. This approach allows achieving a significant socio-economic, environmental
and technological effect, which consists in: efficient use of material, land and labor and time
AdvAIT-2024: 1st International Workshop on Advanced Applied Information Technologies, December 5, 2024, Khmelnytskyi,
Ukraine - Zilina, Slovakia
∗ Corresponding author.
† These authors contributed equally.
diachenko.g@nmu.one (G. Diachenko); laktionov.i.s@nmu.one (I. Laktionov); dmitriy@moroz.cc (D. Moroz);
Maryna.Derzhevetska@mipolytech.education (M. Derzhevetska); semenov.s.y@nmu.one (S. Semenov)
0000-0001-9105-1951 (G. Diachenko); 0000-0001-7857-6382 (I. Laktionov); 0000-0003-2577-3352 (D. Moroz); 0000-0002-
9952-4992 (M. Derzhevetska); 0000-0002-1244-9687 (S. Semenov)
© 2024 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
resources; increasing the resistance of field crops to changing climatic conditions; increasing yields
and minimizing negative environmental impact [1, 2].
Based on a priori analysis of current statistics on agricultural practices at the global level, it has
been established that cereals (wheat, rice, corn, barley, and others) are the most popular crops in
terms of cultivated areas and specific yields [3, 4]. Over the past decade, wheat has been the leader
among cereals at the national level in terms of cultivated areas: from 5.28 million hectares to 7.1
million hectares. It is also necessary to emphasize that with an increase in cultivated areas, there is
no proportional increase in the yield of cereals [5]. This phenomenon is rooted in the fact that during
the full cycle of cultivation, grain crops are subject to destabilizing effects of physical (changing
agroclimatic conditions) and biological (pests and diseases) factors, which negatively affect the
integral stress resistance and productivity and, as a result, crop yields.
Therefore, an actual scientific and practical task is to develop and implement methods, models,
and software and hardware for predicting the occurrence of crop diseases at the pre-symptomatic
stage in real time, which will allow timely planning and implementation of agrotechnical measures
to increase the stress resistance of crops and preserve the harvest in changing agroclimatic
conditions.
In the present-day world practice, there is a significant number of high-quality research and
developments of information and computer technologies for agrotechnical purposes to detect the
parameters and characteristics of the processes of occurrence and progression of crop diseases based
on various data collection technologies [6], in particular: obtaining and analyzing graphic images
from satellites [7, 8] and UAVs [9, 10], as well as collecting and processing data from ground-based
sensor networks [11, 12].
One of the main tasks, which is the focus of many relevant studies in developing intelligent
information technologies for predictive monitoring of crops, is the precise and reliable analysis of
observation results. To date, machine learning methods have gained considerable popularity in
solving problems of intellectual analysis of large amounts of data when creating intelligent
information technologies for various applied fields [13, 14]. In agriculture, such technologies are used
for intelligent processing of agroclimatic data distributed in time and space [15, 16], as these
approaches allow aggregating, analyzing, and interpreting large amounts of measurement data with
subsequent automatic support for making management decisions to optimize agrotechnical
procedures.
The perspectives of integrating sensor networks for agroclimatic monitoring and machine
learning methods have been proven by the authors of scientific studies on: the introduction of
precision farming systems [17], analysis of promising practices for managing agrotechnical
processes and resources [18], accounting for the impact of changing climatic conditions on crop
cultivation regimes [19], and others.
Thus, the results of the analysis of the current state of scientific and applied research prove the
potential and effectiveness of creating and implementing information technologies for predictive
monitoring of crop diseases based on online measurements of soil and climatic parameters with their
subsequent processing by software based on machine learning algorithms. Consequently, the current
research task is to develop computer components for complex intelligent processing of
agromonitoring measurement data that implements a full cycle of data transformation (primary
statistical processing, selection of informative features, and predictive analytics) within an integrated
hardware and software architecture, taking into account the specifics of detected diseases and
agroclimatic growing conditions for specific types and periods of grain crops.
Therefore, the aim of the article is to further develop information technologies for agrotechnical
purposes to predict the risk of occurrence of grain crop diseases (Powdery Mildew Blumeria graminis
in wheat) at the pre-symptomatic stage through the development and research of computerized
method of intelligent analysis of measured data of agroclimatic monitoring utilizing machine
learning models. The object of research is information processes of software analysis of agroclimatic
data distributed in time and space. The subject of research is methods and computer models of
complex predictive processing of agroclimatic monitoring data.
Accordingly, researching the development and validation of computerized methods and software
components of predictive transformation and analysis of agroclimatic measurement data during the
creation of information technologies for agrotechnical purposes to increase the stress resistance of
grain crops is an actual scientific and practical task.
2. Materials and methods
Two primary software environments were used in this research: Python 3.10.12 and MATLAB
R2024a. Preliminary data analysis was performed in Python using libraries such as Pandas and
NumPy. MATLAB was used to train the classifier model with the possibility of further generating
code for microcontrollers. For this purpose, the Simulink and Statistics and Machine Learning
toolboxes were used.
In Figure 1, a generalized structure of the research in this article is illustrated.
Figure 1: Generalized structure of the research.
The data for the research was obtained using professional Metos weather stations from Pessl
Instruments via the FieldClimate IoT platform, access to which was provided by Metos Ukraine LLC.
The experimental data reflects the results of monitoring soil and climatic parameters collected in two
agroclimatic zones of Ukraine from September 2022 to September 2023:
northern steppe of Ukraine: a zone characterized by an arid and very warm climate. The
hydrothermal coefficient varies from 0.7 to 1.0, and the annual sum of temperatures ranges
from 2900 °C to 3300 °C;
forest-steppe of Ukraine: insufficiently humid and warm zone with a hydrothermal
coefficient of 1.0 to 1.3 and an annual sum of temperatures from 2500 °C to 2900 °C.
The data sample from the northern steppe zone (Dnipro region) consisted of 8656 records with a
sampling interval of 1 hour and 14 attributes. Similarly, the data sample from the forest-steppe zone
(Cherkasy region) consisted of 8687 records with the same sampling interval and number of
attributes. Both samples have the probability of occurrence of Powdery Mildew Blumeria graminis
as the target function for further analysis and modeling. A description of all attributes included in
the two data sets is given in Table 1.
Table 1
Soil and climatic attributes present in the dataset
SI. No Attribute Units Datatype Description
1 DT - Continuous Date and time
2 AT °C Continuous Air temperature
3 DP °C Continuous Dew point
4 SR Wt/m2 Continuous Solar radiation
5 VPD kPa Continuous Vapor pressure-deficit
6 RH % Continuous Relative humidity of the air
7 PR mm Continuous Precipitation within one hour
Leaf wetness. If the leaves were wet
8 LW min Discrete
during the last hour (60), otherwise (0)
9 WS m/s Continuous Wind speed
10 WG m/s Continuous Wind direction
11 WD m/s Continuous Wind gust
12 ST °C Continuous Soil temperature
13 ET mm Continuous Evapotranspiration
Risk of Powdery Mildew Blumeria
14 PMBG % Continuous
graminis disease in range from 0 to 100
During the preliminary data analysis, all rows containing missing values were removed to avoid
distortion of the results and to ensure the correctness of the modeling. There were two such rows
and given that ‘PMBG’ is calculated once a day, their removal does not affect the value of the target
function. In addition, for the initial analysis of the data in Python, the describe() function from the
Pandas library was used to obtain statistical information about the numerical characteristics of the
data, such as mean, standard deviation, minimum and maximum values, and quartiles. These
indicators allowed assessing data distribution and identifying possible anomalies and general trends
in the dataset. The statistical indicators obtained as a result of using describe() for the combined
sample from the two regions are shown in Table 2.
Table 2
Descriptive statistics of the data collected (combined data of two regions)
count mean std min 25% 50% 75% max
AT 10.6 9.7 -10.9 2.7 9.8 18.2 37
DP 5.3 7.7 -16.5 -0.3 5.5 11.5 25
SR 135.2 215.7 0 0 3 192 1059
VPD 0.5 0.6 0 0.1 0.2 0.7 5
RH 72.6 17.2 19 61 76 87 100
PR 0.1 0.3 0 0 0 0 15.2
LW 17343 8.6 21 0 0 0 0 60
WS 3.5 1.8 0 2.1 3.2 4.6 13.4
WG 6.1 2.7 0.3 4.1 5.8 7.8 19.1
WD 180.2 107.3 1 83 180 278 360
ST 11.9 8.9 -6.5 3.9 11 19.1 35.7
ET 0.1 0.1 0 0 0 0.1 1.3
PMBG 12.2 19.2 0 0 0 20 70
The average ‘AT’ value of 10.6 °C indicates a moderate climate in the region. The range of values
from 10.9°C to 37°C indicates the presence of both cold and very warm periods. The ‘AT’ values are
centered around 9.8°C (median), with the bottom quartile (25%) falling within 2.7°C and the top
quartile (75%) falling within 18.2°C, indicating a significant temperature variation. The average ‘RH’
value is 72.6%, indicating generally humid conditions. The humidity varies widely, with a minimum
of 19% and a maximum of 100%, with most values concentrated between 61% and 87%. A mean ‘PR’
value of 0.1 mm indicates low precipitation. In fact, most of the records show no precipitation, as the
median and 25th quartile are 0. The maximum value of 15.2 mm indicates significant but rare
precipitation. The average ‘LW’ value is 8.6 minutes per hour, indicating predominantly dry
conditions. 75% of the ‘LW’ values are 0, which means that the leaves remain dry most of the time.
75% of ‘ET’ values below 0.1 mm indicate generally low evapotranspiration. The average ‘PMBG’
value is 12.2%, indicating a low average risk of developing the disease. The maximum ‘PMBG’ is 70%,
which indicates a significant risk in certain periods. At the same time, the median of 0% indicates
that a significant part of the sample had no or low risk of the disease, but the 75th quartile at 20%
shows that there are periods with an increased risk of developing the disease.
The data in Table 2 shows moderate climatic conditions with relatively low precipitation and
moderate temperatures. The risk of developing Powdery Mildew is generally low, although periods
of higher risk require attention.
The risk of Powdery Mildew varies by 10% on average during the day under favorable conditions
[20]. Thus, it was decided to aggregate the hourly measurement results to a daily resolution.
Additionally, eight attributes were introduced: ‘FVT12S’ – the number of hours when the
temperature ranges from 12°C to 21°C; ‘TL16S’ – the number of hours when the temperature is below
15°C; ‘TG21S’ – the number of hours when the temperature exceeds 21°C; ‘FVT16S’ – the number of
hours when the temperature ranges from 16°C to 21°C; ‘TG25S’ – the number of hours when the
temperature exceeds 25°C; ‘SR_count’ – the number of hours when solar radiation was greater than
0; ‘RHG85S’ – the number of hours when relative humidity was greater than or equal to 85%;
‘LW_count’ – the number of hours when the leaves were wet.
Taking into account the above transformations, the columns of the final table are renamed
according to the predefined names stored in the variable INPUT_SUMMARY_COLUMNS =
[‘AT_mean’, ‘FVT12S’, ‘TL16S’, ‘TG21S’, ‘FVT16S’, ‘TG25S’, ‘DP_mean’, ‘SR_sum’, ‘SR_count’,
‘VPD_mean’, ‘RH_mean’, ‘RHG85S’, ‘PR_sum’, ‘LW_count’, ‘WS_mean’, ‘WD_mean’, ‘WG_mean’,
‘ST_mean’, ‘ET_sum’, ‘PMBG_mean’], and the calculations correspond to the Python code (see
Appendix A). The difference between the value of the disease risk for the previous day and the
current day is calculated using diff(). Then, based on this difference, a new attribute ‘PMBG_class’ is
created in which the class of risk change is stored: 1 – risk has increased, 2 – risk has decreased, 0 –
risk has not changed.
The data was then split into training and test samples in a 70:30 ratio. 70% of the data was used
to train the models, and 30% to evaluate their performance.
After performing these steps, the imbalance of classes in the target variable of the training dataset
was detected (class 1 – 6.6%, class 2 – 6.4%, class 0 – 87%). This can lead to the model learning to
favor a more common class, ignoring important features of less represented classes, which ultimately
degrades the overall performance of the classifier model. Two approaches were used to address this
problem: ‘undersampling’ and ‘oversampling’. The ‘undersampling’ approach is about reducing the
number of instances of the majority class to achieve a balance with the minority. This allows the
model to better account for the minority, as the number of samples of all classes becomes
proportional. To implement this approach, the pandas.DataFrame.sample() function was used, which
allows randomly selecting instances from the majority class. Figure 2(a) shows a comparison of the
original dataset before dividing it into training and test samples and the data after resampling. The
second approach is ‘oversampling’, which implies increasing the number of minority samples to
achieve balance with the majority. One of the most common methods for this is the Synthetic
Minority Over-sampling Technique (SMOTE) [21, 22]. This method generates synthetic minority
samples by interpolating between real samples. It works by selecting a few nearest neighbors for
each minority pattern and creating new patterns based on these neighbors. Figure 2(b) shows a
comparison of the original training set and its version after SMOTE.
700 500
600 original 400 original train
sampled smote
500
300
400
300 200
200
100
100
0 0
0 1 2 0 1 2
(a) (b)
Figure 2: (a) comparison of original and resampled preprocessed dataset across three categories, (b)
comparison of original train data and resampled train data with SMOTE across three categories.
These two approaches help to achieve a balance in the data and improve the overall quality of the
model, preventing it from being biased towards the class with more instances and increasing
classification accuracy for less represented classes.
The next step in the process of preparing data before training classification models is scaling,
including standardization and normalization. Scaling is an important step in data processing because
different features can have different scales, which can negatively affect the performance of machine
learning models, especially those based on distance or gradient methods.
The mean and standard deviation are computed from the training data, and these values are then
used to standardize the training data. The same mean and standard deviation from the training set
are also applied to standardize the test data. This ensures that both datasets are transformed
consistently, allowing for accurate evaluation of the performance of the model. The standardization
is performed in MATLAB Classification Learner App [23] automatically before training the models.
The last step is to extract significant features. To do this, the 'Feature Selection Wrapper Class'
algorithm from a GitHub toolbox was used [24]. In the research, the Harris Hawk Optimizer (HHO)
method was used to extract significant features. This feature selection method mimics the behavior
and joint hunting strategy of Harris's hawks when chasing prey.
The algorithm proposed by Heidari and co-authors [25] is based on the swarm approach, does
not use gradients, and includes evolutionary optimization elements. HHO consists of two main
phases, exploration and exploitation, which alternate in time to find the best parameters.
This algorithm has a high convergence rate and excellent local search capabilities, which makes
it effective for feature selection problems. In this research, the HHO algorithm proposed five
significant features (‘TL16S’, ‘FVT16S’, ‘DP_mean’, ‘ST_mean’, ‘ET_sum’), which were extracted for
further analysis and model training.
After feature selection, the training set was used to train various Simulink-compatible machine-
learning classifiers. Types of classifiers chosen and their relevant hyperparameters [26, 27] are
summarized in Table 3.
Table 3
Models and Hyperparameters used for training and testing
Model Type Preset Hyperparameters
Tree Fine Tree Maximum number of splits: 100; Split criterion: Gini's
diversity index; Surrogate decision splits: Off
Tree Medium Tree Maximum number of splits: 20; Split criterion: Gini's
diversity index; Surrogate decision splits: Off
Tree Coarse Tree Maximum number of splits: 4; Split criterion: Gini's
diversity index; Surrogate decision splits: Off
Discriminant Linear Discriminant Covariance structure: Full
Discriminant Quadratic Covariance structure: Full
Discriminant
Efficient Efficient Logistic Learner: Logistic regression; Solver: Auto;
Logistic Regression Regularization: Auto; Regularization strength (Lambda):
Regression Auto; Relative coefficient tolerance (Beta tolerance):
0.0001; Multiclass coding: One-vs-One
Efficient Efficient Linear SVM Learner: SVM; Solver: Auto; Regularization: Auto;
Linear SVM Regularization strength (Lambda): Auto; Relative
coefficient tolerance (Beta tolerance): 0.0001; Multiclass
coding: One-vs-One
Naive Bayes Gaussian Naive Distribution name for numeric predictors: Gaussian;
Bayes Distribution name for categorical predictors: Not
Applicable
Naive Bayes Kernel Naive Bayes Distribution name for numeric predictors: Kernel;
Distribution name for categorical predictors: Not
Applicable; Kernel type: Gaussian; Support: Unbounded;
Standardize data: Yes
SVM Linear SVM Kernel function: Linear; Kernel scale: Automatic; Box
constraint level: 1; Multiclass coding: One-vs-One;
Standardize data: Yes
SVM Quadratic SVM Kernel function: Quadratic; Kernel scale: Automatic;
Box constraint level: 1; Multiclass coding: One-vs-One;
Standardize data: Yes
SVM Cubic SVM Kernel function: Cubic; Kernel scale: Automatic; Box
constraint level: 1; Multiclass coding: One-vs-One;
Standardize data: Yes
SVM Fine Gaussian SVM Kernel function: Gaussian; Kernel scale: 0.56; Box
constraint level: 1; Multiclass coding: One-vs-One;
Standardize data: Yes
SVM Medium Gaussian Kernel function: Gaussian; Kernel scale: 2.2; Box
SVM constraint level: 1; Multiclass coding: One-vs-One;
Standardize data: Yes
SVM Coarse Gaussian Kernel function: Gaussian; Kernel scale: 8.9; Box
SVM constraint level: 1; Multiclass coding: One-vs-One;
Standardize data: Yes
KNN Fine KNN Number of neighbors: 1; Distance metric: Euclidean;
Distance weight: Equal; Standardize data: Yes
KNN Medium KNN Number of neighbors: 10; Distance metric: Euclidean;
Distance weight: Equal; Standardize data: Yes
KNN Coarse KNN Number of neighbors: 100; Distance metric: Euclidean;
Distance weight: Equal; Standardize data: Yes
KNN Cosine KNN Number of neighbors: 10; Distance metric: Cosine;
Distance weight: Equal; Standardize data: Yes
KNN Cubic KNN Number of neighbors: 10; Distance metric: Minkowski
(cubic); Distance weight: Equal; Standardize data: Yes
KNN Weighted KNN Number of neighbors: 10; Distance metric: Euclidean;
Distance weight: Squared inverse; Standardize data: Yes
Ensemble Boosted Trees Ensemble method: AdaBoost; Learner type: Decision
tree; Maximum number of splits: 20; Number of learners:
30; Learning rate: 0.1; Number of predictors to sample:
Select All
Ensemble Bagged Trees Ensemble method: Bag; Learner type: Decision tree;
Maximum number of splits: 99; Number of learners:
30; Number of predictors to sample: Select All
Ensemble RUSBoosted Trees Ensemble method: RUSBoost; Learner type: Decision
tree; Maximum number of splits: 20; Number of
learners: 30; Learning rate: 0.1; Number of predictors
to sample: Select All
Neural Narrow Neural Number of fully connected layers: 1; First layer size:
Network Network 10; Activation: ReLU; Iteration limit: 1000;
Regularization strength (Lambda): 0; Standardize
data: Yes
Neural Medium Neural Number of fully connected layers: 1; First layer size:
Network Network 25; Activation: ReLU; Iteration limit: 1000;
Regularization strength (Lambda): 0; Standardize
data: Yes
Neural Wide Neural Number of fully connected layers: 1; First layer size:
Network Network 100; Activation: ReLU; Iteration limit: 1000;
Regularization strength (Lambda): 0; Standardize
data: Yes
Neural Bilayered Neural Number of fully connected layers: 2; First layer size:
Network Network 10; Second layer size: 10; Activation: ReLU; Iteration
limit: 1000; Regularization strength (Lambda): 0;
Standardize data: Yes
Neural Trilayered Neural Number of fully connected layers: 3; First layer size:
Network Network 10; Second layer size: 10; Third layer size: 10;
Activation: ReLU; Iteration limit: 1000;
Regularization strength (Lambda): 0; Standardize
data: Yes
Based on the described methodology for assessing the risk of Powdery Mildew Blumeria graminis
in wheat, the research steps of this article were presented in the form of a structural algorithmic
scheme, as shown in Figure 3.
When training the models, the five-fold cross-validation methodology was used to assess their
performance [28]. This approach allows for efficient use of available data and reduces the risk of
model overtraining.
The data is divided into five subsets of equal size. At each iteration, four subsets are used to train
the model, and one is used to test it. As a result, average accuracy rates are obtained, which gives a
more stable and reliable assessment of model quality on different chunks of the dataset.
3. Results and discussion
After training and testing the classification models, the following results were obtained, as shown in
Table 4. For each model, training was performed using undersampling and oversampling data
balancing approaches, in particular, the accuracy rates on the validation (Val.) and test (Test) datasets
were compared.
Figure 3: Machine learning pipeline to predict PMBG risk using soil and climatic monitoring data.
Table 4
The summary of the findings from the various machine learning models applied in this study with
respect to the Harris Hawks algorithm
Model Undersample, Undersample, Oversample, Oversample,
Acc. % (Val.) Acc. % (Test) Acc. % (Val.) Acc. % (Test)
Fine Tree 74 65 84 73
Medium Tree 74 65 83 73
Coarse Tree 74 56 76 65
Linear Discriminant 69 58 75 62
Quadratic Discriminant 60 65 82 64
Efficient Logistic Regression 69 58 74 62
Efficient Linear SVM 68 58 76 59
Gaussian Naive Bayes 65 65 72 63
Kernel Naive Bayes 76 70 76 56
Linear SVM 68 60 76 60
Quadratic SVM 72 67 87 76
Cubic SVM 70 63 90 80
Fine Gaussian SVM 69 58 91 78
Medium Gaussian SVM 73 58 84 71
Coarse Gaussian SVM 61 49 73 58
Fine KNN 72 60 90 75
Medium KNN 68 58 86 71
Coarse KNN 39 19 76 63
Cosine KNN 57 60 85 69
Cubic KNN 68 58 86 72
Weighted KNN 75 63 89 72
Boosted Trees 48 65 84 72
Bagged Trees 80 65 88 74
RUSBoosted Trees 71 63 83 75
Narrow Neural Network 70 63 88 73
Medium Neural Network 75 58 91 80
Wide Neural Network 69 70 90 78
Bilayered Neural Network 69 56 90 71
Trilayered Neural Network 69 60 89 76
The tree models (Fine Tree, Medium Tree, Coarse Tree) showed the best accuracy when using
oversampling, with the best accuracy for Fine Tree (84% on validation and 73% on testing). Coarse
Tree significantly reduced the accuracy on the undersampled test set (56%). SVM models had stable
results, especially when using oversampling. For example, Quadratic SVM and Cubic SVM
demonstrated high accuracy rates (up to 90% on the validation set). Fine KNN and Weighted KNN
achieved the best results in oversampling, showing 90% and 89% accuracy, respectively, on validation.
Medium and Wide Neural Networks showed the highest accuracy (91% and 90%, respectively) with
oversampling, indicating their ability to efficiently process more balanced data. The tree-based
methods (Boosted Trees, Bagged Trees, RUSBoosted Trees) also performed well, especially Bagged
Trees, with 88% accuracy on the validation and 74% on the oversampled test set. Thus, the use of
oversampling generally improved the performance of the models, especially for SVMs, KNNs, and
neural networks.
To analyze the performance of the models in more detail, the Confusion Matrix for the training
and test data, as well as the ROC curves for the three classes, were constructed, as shown in Figure
4 and Figure 5. This analysis was performed for two models: Bagged Trees (on undersampled data)
and Medium Neural Network (on oversampled data).
The Confusion Matrix for both models showed similar results. For the Bagged Trees model on the
undersampled data, it can be seen that the model generally copes with the classification of classes,
although there are some errors in the classification of smaller classes. The Medium Neural Network
model, on the other hand, showed better results on oversampled data, reducing the number of
misclassifications, especially for less represented classes.
The ROC curves for both models show high sensitivity and specificity for each of the three classes.
Both models perform well for most classes, with the Medium Neural Network on oversampled data
showing slightly higher performance in terms of area under the curve (AUC) for class ‘2’.
1
0 21 3 7 0 11 1 5
0.8
True Positive Rate
True Class
True Class
0.6
1 26 4 1 2 12 4
0.4
0 (AUC = 0.8778)
0 Model Operating Point
1 (AUC = 0.8456)
0.2
1 Model Operating Point
2 3 3 33 2 2 1 5 2 (AUC = 0.8018)
2 Model Operating Point
0
0 1 2 0 1 2 0 0.2 0.4 0.6 0.8 1
Predicted Class Predicted Class False Positive Rate
(a) (b) (c)
Figure 4: Results for Bagged Trees with undersampled data (a) Validation confusion matrix, (b) Test
confusion matrix, (c) Test ROC curve.
1
0 382 10 51 0 160 7 21
0.8
True Positive Rate
True Class
True Class
0.6
1 11 425 7 1 6 9 5
0.4
0 (AUC = 0.8431)
0 Model Operating Point
1 (AUC = 0.8432)
0.2
1 Model Operating Point
2 29 12 402 2 4 6 2 (AUC = 0.8216)
2 Model Operating Point
0
0 1 2 0 1 2 0 0.2 0.4 0.6 0.8 1
Predicted Class Predicted Class False Positive Rate
(a) (b) (c)
Figure 5: Results for Medium Neural Network with oversampled data (a) Validation confusion
matrix, (b) Test confusion matrix, (c) Test ROC curve.
The Bagged Trees model trained on undersampled data was then exported [29] to the Simulink
environment (Figure 6) to implement a computerized model for automatic data processing to predict
the Risk of Powdery Mildew in Wheat.
label
U
Y x To Workspace
From Workspace Idx1
1
score
Selector
ClassificationEnsemble Predict
SelectedFeatureIndex
Figure 6: Simulink model for Bagged Trees with undersampled data.
An example of the model output in Figure 6 for the northern steppe scenario in the form of time
graphs is shown in Figure 7, comparing actual and predicted data on the risk of Powdery Mildew
Blumeria graminis in wheat.
80
predicted
70 dataset
60
50
PMBG risk, %
40
30
20
10
0
0 50 100 150 200 250 300 350
Days
Figure 7: Results of modeling in Simulink for Bagged Trees with undersampled data for the period
from September 2022 to September 2023.
The analysis of time graphs shows that the predicted data from the model is able to reproduce the
main trends of growth and reduction of disease risk in the relevant time periods. Although there are
some discrepancies between the actual and predicted values at certain points, in general, the model
reflects risk behavior well and can be used for operational monitoring and decision-making in open-
field conditions.
4. Conclusion
In this research, the relevant scientific and practical task has been solved by developing and
researching the computerized method for predicting the risk of powdery mildew in wheat based on
software analysis of soil and climatic monitoring data. The research allowed for the further
development of information technologies for agrotechnical purposes to predict the risk of occurrence
of grain crop diseases (Powdery Mildew Blumeria graminis in wheat) at the pre-symptomatic stage.
The main results of the research include:
1. Using undersampling and oversampling methods to solve the problem of class imbalance in the
training sample.
2. The application of feature selection algorithms, in particular Harris Hawk Optimizer, reduced
the number of features to five.
3. Classification models such as Bagged Trees and Medium Neural Network performed well on
both validation and test datasets, demonstrating good generalizability.
4. The export of the Bagged Trees model to the Simulink environment and the subsequent
generation of program code for microcontroller devices allows it to be used for real-world
prediction and control in agroclimatic systems based on fog architecture.
To improve the results, future research should pay attention to the following aspects:
1. Improving the parameters of algorithms such as Bagged Trees and Neural Networks by further
tuning hyperparameters, which can lead to even more accurate results.
2. Involvement of new data sources, such as information on field cultivation, which can improve
the recognition of conditions that contribute to the occurrence and development of diseases.
3. Further integration of the model with IoT platforms for automatic real-time monitoring can
provide more up-to-date and accurate data for predicting disease risks.
Acknowledgments
This research was carried out as part of the scientific project ‘Development of software and hardware
of intelligent technologies for sustainable crop production in wartime and post-war’ funded by the
Ministry of Education and Science of Ukraine at the expense of the state budget (state registration
number 0124U000289).
Declaration on Generative AI
The authors have not employed any Generative AI tools.
References
[1] I. Laktionov, G. Diachenko, V. Kashtan, A. Vizniuk, V. Gorev, K. Khabarlak, Y. Shedlovska, A
Comprehensive Review of Recent Approaches and Hardware-Software Technologies for
Digitalisation and Intellectualisation of Open-Field Crop Production: Ukrainian Case Study in
the Global Context, Computers and Electronics in Agriculture 225 (2024) 1–31.
doi:10.1016/j.compag.2024.109326.
[2] A. Soussi, E. Zero, R. Sacile, D. Trinchero, M. Fossa, Smart Sensors and Smart Data for Precision
Agriculture: A Review, Sensors 24 (8) (2024) 1–32. doi: 10.3390/s24082647.
[3] USDA: U.S. Department of Agriculture, 2024. URL: https://www.usda.gov/.
[4] FAO: Food and Agriculture Organization of the United Nations, 2024. URL:
https://www.fao.org/home/en/.
[5] FAOSTAT: Food and Agriculture Organization of the United Nations, 2024. URL:
https://www.fao.org/faostat/en/#data/QCL.
[6] Q. Zheng, W. Huang, Q. Xia, Y. Dong, H. Ye, H. Jiang, S. Chen, S. Huang, Remote Sensing
Monitoring of Rice Diseases and Pests from Different Data Sources: A Review, Agronomy 13 (7)
(2023) 1–18. doi: 10.3390/agronomy13071851.
[7] P. Karmakar, S.W. Teng, M. Murshed, S. Pang, Y. Li, H. Lin, Crop monitoring by multimodal
remote sensing: A review, Remote Sensing Applications: Society and Environment 33 (2024) 1–
15. doi: 10.1016/j.rsase.2023.101093.
[8] A. San Bautista, D. Fita, B. Franch, S. Castiñeira-Ibáñez, P. Arizo, M.J. Sánchez-Torres, I. Becker-
Reshef, A. Uris, C. Rubio, Crop Monitoring Strategy Based on Remote Sensing Data (Sentinel-2
and Planet), Study Case in a Rice Field after Applying Glycinebetaine, Agronomy 12 (3) (2022)
1–23. doi: 10.3390/agronomy12030708.
[9] L. Kouadio, M. El Jarroudi, Z. Belabess, S.-E. Laasli, M.Z.K. Roni, I.D.I. Amine, N. Mokhtari,
F. Mokrini, J. Junk, R. Lahlali, A Review on UAV-Based Applications for Plant Disease Detection
and Monitoring, Remote Sensing 15 (17) (2023) 1–23. doi: 10.3390/rs15174273.
[10] A. Abbas, Z. Zhang, H. Zheng, M.M. Alami, A.F. Alrefaei, Q. Abbas, S.A.H. Naqvi, M.J. Rao,
W.F.A. Mosa, Q. Abbas, A. Hussain, M.Z. Hassan, L. Zhou, Drones in Plant Disease Assessment,
Efficient Monitoring, and Detection: A Way Forward to Smart Agriculture,
Agronomy 13 (6) (2023) 1–26. doi: 10.3390/agronomy13061524.
[11] S. Wang, P. Qi, W. Zhang, X. He, Development and application of an intelligent
plant protection monitoring system, Agronomy 12 (5) (2022) 1–15.
[12] I. Laktionov, G. Diachenko, V. Koval, M. Yevstratiev, Computer-oriented model
for Network Aggregation of Measurement Data in IoT monitoring of soil and climatic
parameters of agricultural crop production enterprises, Baltic Journal of Modern
Computing 11 (3) (2023) 500–522. doi: 10.22364/bjmc.2023.11.3.09.
[13] O. Kovalchuk, A Machine Learning Cluster Model for the Decision-Making Support in Criminal
Justice, Computer Systems and Information Technologies 3 (2023) 51–58. doi: 10.31891/csit-
2023-3-6.
[14] O. Pavlova, V. Alekseiko, The Concept of an Information System for Forecasting the
Temperature Regime of the Earth’s Surface Based on Machine Learning, Computer Systems and
Information Technologies 2 (2024) 6–13. doi: 10.31891/csit-2024-2-1.
[15] K.G. Liakos, P. Busato, D. Moshou, S. Pearson, D. Bochtis, Machine Learning in Agriculture: A
Review, Sensors 18 (8) (2018) 1–29. doi: 10.3390/s18082674.
[16] D. Tamayo-Vera, X. Wang, M. Mesbah, A Review of Machine Learning Techniques in
Agroclimatic Studies, Agriculture 14 (3) (2024) 1–19. doi: 10.3390/agriculture14030481.
[17] C.E. Hachimi, S. Belaqziz, S. Khabba, B. Sebbar, D. Dhiba, A. Chehbouni, Smart Weather Data
Management Based on Artificial Intelligence and Big Data Analytics for Precision Agriculture,
Agriculture 13 (1) (2023) 1–22. doi: 10.3390/agriculture13010095.
[18] S.O. Araújo, R.S. Peres, J.C. Ramalho, F. Lidon, J. Barata, Machine Learning Applications in
Agriculture: Current Trends, Challenges, and Future Perspectives, Agronomy 13 (12) (2023) 1–27.
[19] J. Šuljug, J. Spišić, K. Grgić, D. Žagar, A Comparative Study of Machine Learning Models for
Predicting Meteorological Data in Agricultural Applications, Electronics 13 (16) (2024) 1–20.
[20] J. Bradley, G. Thomas, Wheat powdery mildew epidemiology and crop management options. In:
GRDC Update papers, 2019. URL: https://grdc.com.au/resources-and-publications/grdc-update-
papers/tab-content/grdc-update-papers/2019/02/wheat-powdery-mildew-epidemiology-and-
crop-management-options.
[21] I. Dey, V. Pratap, A Comparative Study of SMOTE, Borderline-SMOTE, and ADASYN
Oversampling Techniques using Different Classifiers. 2023 3rd International Conference on
Smart Data Intelligence (ICSMDI), Trichy, India, (2023) 294–302.
[22] SMOTE – Synthetic Minority Over-sampling Technique. URL: https://imbalanced-
learn.org/stable/references/generated/imblearn.over_sampling.SMOTE.html#id1.
[23] Classification Learner App, 2024. URL: https://www.mathworks.com/help/stats/classification-
learner-app.html.
[24] J. Too, Jx-WFST: A Wrapper Feature Selection Toolbox, 2024. URL:
https://github.com/JingweiToo/Wrapper-Feature-Selection-Toolbox.
[25] A.A. Heidari, S. Mirjalili, H. Faris, I. Aljarah, M. Mafarja, H. Chen, Harris hawks optimization:
Algorithm and applications, Future Generation Computer Systems, 97 (2019) 849-872. doi:
10.1016/j.future.2019.02.028.
[26] Train Classification Models in Classification Learner App. URL:
https://www.mathworks.com/help/stats/train-classification-models-in-classification-learner-
app.html.
[27] M. Patlak, M. Çunkaş, U. Taskiran, The Innovative Approach to Real-Time Detection of Fuel
Types Based on Ultrasonic Sensor and Machine Learning. Arabian Journal for Science and
Engineering 49 (2024) 16571–16591. doi: 10.1007/s13369-024-09092-5.
[28] K. Pal, B.V. Patel, Data Classification with k-fold Cross Validation and Holdout Accuracy
Estimation Methods with 5 Different Machine Learning Techniques. In: 2020 Fourth
International Conference on Computing Methodologies and Communication (ICCMC), Erode,
India, (2020) 83–87. doi: 10.1109/ICCMC48092.2020.ICCMC-00016.
[29] Export Classification Model to Predict New Data, 2024. URL:
https://www.mathworks.com/help/stats/export-classification-model-for-use-with-new-data.html.
Appendix A. Python code for daily summarized data
def read_df_with_daily_summary(csv_path: str) -> pd.DataFrame:
df = pd.read_csv(csv_path)
df["DT"] = pd.to_datetime(df["DT"], dayfirst=True)
df.set_index("DT", inplace=True)
# Resample data by day to calculate the required values
df_daily_summary = df.resample("D").agg(
{
"AT": [
"mean",
lambda x: ((x >= 12) & (x <= 21)).sum(),
lambda x: (x < 15).sum(),
lambda x: (x > 21).sum(),
lambda x: ((x >= 16) & (x <= 21)).sum(),
lambda x: (x >= 25).sum(),
],
"DP": "mean",
"SR": ["sum", lambda x: (x > 0).sum()],
"VPD": "mean",
"RH": ["mean", lambda x: (x >= 85).sum()],
"PR": "sum",
"LW": lambda x: (x > 0).sum(),
"WS": "mean",
"WD": "mean",
"WG": "mean",
"ST": "mean",
"ET": "sum",
"PMBG": "mean",
} )
df_daily_summary.columns = INPUT_SUMMARY_COLUMNS
df_daily_summary["PMBG_class"] = (
df_daily_summary["PMBG_mean"]
.diff()
.apply(lambda x: 1 if x > 0 else 2 if x < 0 else 0)
)
# Reset the index for a clean DataFrame
df_daily_summary.reset_index(inplace=True)
df_daily_summary.drop(columns=["DT", "PMBG_mean"], inplace=True)
return df_daily_summary