mathematics

Article
Analysis of a Predictive Mathematical Model of Weather
Changes Based on Neural Networks
Boris V. Malozyomov 1 , Nikita V. Martyushev 2, * , Svetlana N. Sorokova 3 , Egor A. Efremenkov 3 ,
Denis V. Valuev 4 and Mengxu Qi 2

1 Department of Electrotechnical Complexes, Novosibirsk State Technical University,

630073 Novosibirsk, Russia; [email protected]
2 Department of Materials Science, Tomsk Polytechnic University, 634050 Tomsk, Russia; [email protected]
3 Department of Mechanical Engineering, Tomsk Polytechnic University, 634050 Tomsk, Russia;
[email protected] (S.N.S.); [email protected] (E.A.E.)
4 Yurga Technological Institute (Branch), Tomsk Polytechnic University, 652055 Yurga, Russia; [email protected]
* Correspondence: [email protected]

Abstract: In this paper, we investigate mathematical models of meteorological forecasting based on

the work of neural networks, which allow us to calculate presumptive meteorological parameters of
the desired location on the basis of previous meteorological data. A new method of grouping neural
networks to obtain a more accurate output result is proposed. An algorithm is presented, based
on which the most accurate meteorological forecast was obtained based on the results of the study.
This algorithm can be used in a wide range of situations, such as obtaining data for the operation of
equipment in a given location and studying meteorological parameters of the location. To build this
model, we used data obtained from personal weather stations of the Weather Underground company
and the US National Digital Forecast Database (NDFD). Also, a Google remote learning machine
was used to compare the results with existing products on the market. The algorithm for building
the forecast model covered several locations across the US in order to compare its performance in
different weather zones. Different methods of training the machine to produce the most effective
weather forecast result were also considered.
Citation: Malozyomov, B.V.;
Martyushev, N.V.; Sorokova, S.N.;
Efremenkov, E.A.; Valuev, D.V.; Qi, M.
Keywords: weather mathematical model; forecast; neural network; algorithm for building weather
Analysis of a Predictive Mathematical forecasts
Model of Weather Changes Based on
Neural Networks. Mathematics 2024, MSC: 62M45
12, 480. https://doi.org/10.3390/
math12030480

Academic Editor: Snezhana

1. Introduction
Gocheva-Ilieva
One of the most widespread problems with providing scientific, industrial, and other
Received: 25 December 2023 human activities has been and remains the problem of obtaining information about the
Revised: 1 February 2024 environment at the required time [1]. This problem becomes especially important for the
Accepted: 1 February 2024 support of technical systems (TSs) dependent on meteorological conditions and operating
Published: 2 February 2024 in an autonomous mode. Their parameters cannot be calibrated depending on the meteo-
rological conditions at a certain moment, which makes it necessary to solve the problem
of predicting meteorological conditions over a long period of time [2–4]. While predicting
values for single-parameter technical systems (SPTs) does not pose special problems, multi-
Copyright: © 2024 by the authors.
parameter technical systems (MTSs) require a thoughtful approach to creating an algorithm
Licensee MDPI, Basel, Switzerland.
This article is an open access article
for building a predictive model, which makes this problem very relevant.
distributed under the terms and
The problem of obtaining information about the environment and weather forecasting
conditions of the Creative Commons is of great importance for various spheres of human activity. Scientific research, industrial
Attribution (CC BY) license (https:// operations, agriculture, urban planning, tourism, and many other sectors require accurate
creativecommons.org/licenses/by/ and up-to-date information on weather conditions, climatic changes, and environmental
4.0/). conditions [5–8].

Mathematics 2024, 12, 480. https://doi.org/10.3390/math12030480 https://www.mdpi.com/journal/mathematics

Mathematics 2024, 12, 480 2 of 17

In the past, environmental information was based mainly on observations and data
that were obtained from radio meteorological stations, satellites, sensors, and other de-
vices. Nevertheless, this approach has its limitations due to the large amount of data, the
complexity of data processing, and limitations in space and time [9,10].
Due to the development of neural networks and machine learning, new approaches
to analysing and predicting weather and environmental conditions have become possible.
Neural networks can process large amounts of data and detect complex patterns in weather
events, the climate, and other aspects of the environment. They are able to use these
patterns to create predictive models that can predict weather and other factors with high
accuracy. The application of neural networks in the field of weather and hydrometeorology
can improve the quality of forecasts and provide more accurate and reliable data about
weather conditions. Neural networks are able to take into account complex interactions
between different factors, such as temperature, atmospheric pressure, humidity, and wind,
which allows us to obtain more accurate weather forecasts for short and long periods of
time [11–13].
In addition to weather forecasting, neural networks can be used to analyse and predict
other aspects of the environment, such as air pollution, water quality, vegetation, and
ecosystem health. This enables informed decision-making in environmental protection,
resource management, and other fields of endeavour where environmental information is
an important factor. Determining and predicting weather using neural networks require a
large amount of data for training and tuning models. Therefore, it is important to create
high-quality datasets that include information on past weather events as well as data
collected in real time. In addition, machine learning algorithms and neural networks
need to be further developed and improved to achieve more accurate and reliable results.
Therefore, the application of neural networks in analysing and forecasting weather and the
environment makes it possible to solve complex problems related to the acquisition and
processing of information in various spheres of human activity. This increases not only
the efficiency of scientific and production processes, but also contributes to environmental
conservation and informed decision-making in the field of sustainable development [14–17].
At the moment, various methods and means of obtaining and predicting meteo-
rological data at the desired location of their parameters are known [18–21], including
the following:
1. The study of weather phenomena at the current location using physical laws and the
Weather Research & Forecasting Model available at https://www.mmm.ucar.edu/
models/wrf (accessed on 1 December 2023).
2. Investigation of weather conditions by means of mathematical transformations of
data received from the probes.
3. The use of radar and satellite observations to obtain meteorological information.
Radars and satellites can provide data on wind speed, temperature, atmospheric
pressure, humidity, and other parameters using a variety of sensors and instruments.
4. Application of meteorological balloons and aerological probes that are released into
the atmosphere and equipped with meteorological instruments to measure vertical
profiles of temperature, humidity, pressure, and wind speed. The obtained data help
in weather forecasting and analysing weather conditions.
5. The use of a network of meteorological stations located at various locations. These
stations have meteorological instruments that continuously monitor and record data
on weather conditions, such as temperature, humidity, pressure, and precipitation.
The information from these stations helps in forecasting and analysing the weather at
the location.
6. Utilization of computer weather prediction models that are used to analyse and
forecast meteorological data. These models take into account the physical laws of the
atmosphere, data from meteorological observations, and other parameters to create
forecasts of future weather conditions.
Mathematics 2024, 12, 480 3 of 17

7. Usage of remote sensing, such as LIDAR (laser sensing), and infrared sensing to
measure atmospheric parameters and obtain weather data. These technologies are
based on the analysis of reflected or emitted radiation and can measure temperature,
humidity, cloud cover, and other weather characteristics at the desired location.
All these methods and tools are used together to acquire and forecast meteorological
data at a given location, which is essential for weather forecasting and the planning
of agricultural and urban activities, including safety measures and protection against
weather disasters.
The current number of solutions to the problem of building meteorological forecast
models and analysing them is staggering. One of the most successful, but at the same time
most costly, solutions is the solution provided by Weather Research & Forecasting, which is
reduced to the construction of a mathematical model of physical phenomena and, based
on the current data, the application of a certain location. However, this approach rarely
gives accurate results because it requires a huge number of resources for the solution and
constant calculations taking into account the changing parameters of the system. Such a
solution also demands the same construction of models for neighbouring regions, which
increases the cost of accurate forecasting exponentially, and also does not take into account
the essential factors.
Neural networks use one of the most promising methods of MTS forecasting, which
is mathematical extrapolation, as this method is based only on the statistical analysis of
data [22–24]. The drawback of this approach is the impossibility of extrapolating the further
development of the process for a long time interval. For this purpose, patterns between
input and output values are built, relying on training data. In the case of meteorolog-
ical forecasting, such data are previous observations collected by the personal weather
stations (PWSs).
This paper uses a learning paradigm in which a finite number of neural networks
are trained based on the same data to obtain the desired result. This paradigm is called
grouping neural networks [25–28].
The purpose of this study is to describe, analyse, and compare an algorithm for
building a predictive model for an MTS for a particular location using the example of data
obtained from personal weather stations for certain regions using self-learning machines.
The structure of the paper is as follows. The introduction presents the relevance and
scientific novelty of the analysis of the predictive mathematical model of weather changes
based on neural networks. The second section analyses the structure and application of
artificial neural networks. In the third section, the mathematical models of the used neural
networks are given. The fourth section describes the mathematical model of the grouping
of neural networks proposed for weather forecasts. The fifth section gives an overview of
the performance results of the considered mathematical models. The sixth section presents
the main conclusions of this work.

2. Analysing the Application of Artificial Neural Networks

2.1. Stages of Neural Network Construction
When building a neural network, several stages can be distinguished, which can be
described in general terms as follows:
- Defining goals and objectives;
- Data collection;
- Data preprocessing;
- Selection of the neural network architecture;
- Determination of hyperparameters;
- Training the network;
- Model evaluation and tuning;
- Model development and optimisation.
 grouping method of neural networks).
At the fourth stage, the neural network is trained by feeding pronormalised data to
the inputs of the mathematical model. The neural network builds functional dependen-
cies between the data. Finally, the output is a function of the dependence of input and
Mathematics 2024, 12, 480 output values in a certain training dataset. 4 of 17
At the fifth stage, the obtained model is checked in terms of accuracy and adequacy.
In forecasting, it is necessary to accurately establish the concept of the necessary accuracy
of the result.
Each ofInthese
this paper,
stages the
has model
its ownispeculiarities
considered adequate
and may when
requireitsadditional
forecast accuracy is
actions and
10%, and the accuracy score is calculated as the percentage ratio between the
experiments to achieve the desired result. To summarise, the main stages of a neural data ob-
tained by theconstruction
network’s model and theare real
showndata.
in Figure 1.

Figure
Figure 1. 1. Main
Main stages
stages ofof a neural
a neural network’s
network’s construction.
construction.
The first step is to collect weather data. The second step is to normalise and discard
2.2. Improving the Prediction Accuracy of Neural Networks
corrupted data obtained from Weather Underground and the US National Digital Forecast
Increasing
Database. Thethe prediction accuracy
normalisation of neuralentirely
process depends networkson istheanchoice
important goal ininthe
of topology the
field
previous step; for example, for RBFN, it is necessary to generate the data in such aused
of machine learning, and various approaches and techniques have been to
way that
achieve this goal.
the training data Furthermore, ways answer.
contain the correct to improve the prediction accuracy of neural net-
works Athave
thebeen
thirdformulated and proposed,
stage, a mathematical including
model the to
is selected following:
build functional dependencies
- between input and output data to obtain the required result.
Increasing the training sample size: A large amount of diverse In thisdata
paper,
canthree
help topologies
a neural
arenetwork
considered (D-PNN, RBFN, and MLPN). A new one is also proposed
learn a wider range of patterns and improve generalisability. This (the may
grouping
in-
method of neural networks).
At the fourth stage, the neural network is trained by feeding pronormalised data to
the inputs of the mathematical model. The neural network builds functional dependencies
between the data. Finally, the output is a function of the dependence of input and output
values in a certain training dataset.
At the fifth stage, the obtained model is checked in terms of accuracy and adequacy.
In forecasting, it is necessary to accurately establish the concept of the necessary accuracy
of the result. In this paper, the model is considered adequate when its forecast accuracy is
10%, and the accuracy score is calculated as the percentage ratio between the data obtained
by the model and the real data.

2.2. Improving the Prediction Accuracy of Neural Networks

Increasing the prediction accuracy of neural networks is an important goal in the field
of machine learning, and various approaches and techniques have been used to achieve
this goal. Furthermore, ways to improve the prediction accuracy of neural networks have
been formulated and proposed, including the following:
- Increasing the training sample size: A large amount of diverse data can help a neural
network learn a wider range of patterns and improve generalisability. This may
involve collecting additional data, augmenting existing data, or using generative
modelling techniques to create new samples [29–31].
- Improving data quality: Data quality is also important. Data preprocessing should
be done to eliminate noise, outliers, missing values, and other problems. In addition,
methods can be applied to eliminate class imbalance or sampling imbalance problems.
- Fitting the network architecture: Choosing the right neural network architecture plays
an important role in achieving high prediction accuracy. By using techniques such
Mathematics 2024, 12, 480 5 of 17

as hyperparameter search or optimisation by automatic machine learning, different

network configurations can be explored, including the number of layers, the number
of neurons in each layer, and activation functions.
- Model regularising: Regularisation techniques can help to control overfitting and
generalise the model better to new data. This includes methods such as L1 or L2
regularisation, neuron or layer dropout, and data dimensionality reduction techniques,
such as principal component analysis (PCA) and stochastic neighbour embedding
with t-distribution (t-SNE).
- Using pretraining: Pretraining can be useful, especially when there are not enough
labelled data available. One can use pretrained models using large unlabelled datasets
(e.g., autoencoders or generative models) and then pretrain them based on smaller
labelled datasets for a specific task.
- Combining models: Sometimes, combining multiple models or using ensembles of
models can significantly improve the forecast accuracy. This can be done by combining
forecasts from several models or by using stacking or bagging techniques.
- Hyperparameter tuning: The hyperparameters of a model also affect its prediction
accuracy. Optimisation techniques, such as grid search or random search, can be used
to find optimal values for the hyperparameters.
- Improving the learning process: Applying various optimisation techniques, such as
stochastic gradient descent with momentum (SGD with momentum), Adam’s method,
and early stopping, can help to train the neural network more efficiently and stably.
Mathematics 2024, 12, x FOR PEER REVIEW 6 of 18
Figure 2 shows the increase in neural network prediction accuracy using generalised
data from the ImageNet Large Scale Visual Recognition Challenge (ILSVRC) project [32].

Figure 2. Increase
Figure 2. Increasein
inneural
neuralnetwork
networkprediction
predictionaccuracy:
accuracy:(a)(a)accuracy
accuracyrate and
rate (b)(b)
and top-5 error
top-5 rate
error in
rate
the ILSVRC.
in the ILSVRC.

The
Moderntop-five errorneural
artificial rate indicates
networksthe canpercentage
be dividedofinto
testsimple
cases (Figure 2b) where
and complex the
(percep-
correct class was not among the top five predicted classes. For instance,
trons) depending on the number of neuron layers (one or more). They are also divided if a test image
features a Persian
into direct cat andnetworks
and recurrent the top five predicted classes, ranked from highest to lowest
[34–36].
probability, are (Pomeranian (0.4), Mongoose
Direct neural networks are used to send the (0.25), Dingo
signal from(0.15), Persian
the input catoutput
to the (0.1), and
in a
Tabby
straight line, like a train on a track. Recurrent ones allow for the possibility thatthe
cat (0.02)), it is still considered accurate because the real class is within thetop
re-
five predictions.
sulting Withvalue
intermediate ImageNet’s 1000back
can be sent classes, achieving
to the a low
input again error
and rate in the
go through thetop five
neural
predictions is challenging [33].
network from the beginning. The biological brain is a recurrent network, which is the
Modern artificial neural networks can be divided into simple and complex (percep-
reason why it is almost impossible to understand how it works.
trons) depending on the number of neuron layers (one or more). They are also divided into
Each neural network is a data analysis system that is very powerful and accurate but
direct and recurrent networks [34–36].
requires special tuning (training). Moreover, taking into account the self-learning capac-
ity of a neural network, its tuning is not reduced to setting specific parameters according
to which it will work. The principle of training is quite different:
- A certain value is given to the input.
- It is known in advance what value should be at the output.
- In case the output value is different from the desired value, the network is adjusted
Mathematics 2024, 12, 480 6 of 17

Direct neural networks are used to send the signal from the input to the output in
a straight line, like a train on a track. Recurrent ones allow for the possibility that the
resulting intermediate value can be sent back to the input again and go through the neural
network from the beginning. The biological brain is a recurrent network, which is the
reason why it is almost impossible to understand how it works.
Each neural network is a data analysis system that is very powerful and accurate but
requires special tuning (training). Moreover, taking into account the self-learning capacity
of a neural network, its tuning is not reduced to setting specific parameters according to
which it will work. The principle of training is quite different:
- A certain value is given to the input.
- It is known in advance what value should be at the output.
- In case the output value is different from the desired value, the network is adjusted
until the difference is minimised [37].
The second variant of neural network training is the so-called method of error back-
propagation, in which the value obtained at the output, if it differs from the initial one, is
transmitted back through the same neurons through which it came to the output. In the
process of transmission, the weights of these neurons are increased or decreased. Then, a
Mathematics 2024, 12, x FOR PEER REVIEW
new attempt follows, and so on until the result becomes optimal. Figure 3 shows the general
scheme of the neural network training process proposed for predicting weather changes.

Figure 3. Schematic diagram

Figure of the training
3. Schematic diagramprocess of the weather
of the training processneural
of thenetwork.
weather neural network.
3. Mathematical Models of the Used Neural Networks
3. Mathematical Models of the Used Neural Networks
The first neural network is based on the solution of the differential polynomial (the
The neural
differential polynomial first neural network
network is based
(D-PNN). The on the solution
D-PNN of functional
describes the differential polynomi
depen-
dencies of input parameters and their properties, which are often used for forecasting, function
differential polynomial neural network (D-PNN). The D-PNN describes
analysing timependencies of input parameters
series, and revealing and their in
hidden relationships properties, which
data [38–40]. are
This is often used for for
achieved
by templating the dependencies between the data. The main idea of the D-PNN is the[38–40]. T
ing, analysing time series, and revealing hidden relationships in data
approximationachieved by templating
of functions described bythedifferential
dependencies between
equations thedescribe
that data. The
themain idea of the D
relation-
ships between isthethe approximation
input parameters of ofthe
functions
system.described by differential
The relationships betweenequations that descri
the data are
relationships between the input parameters of the system.
described similarly to the Kolmogorov–Gabor polynomial presented in 1971 [2]: The relationships betwe
data are described similarly to the Kolmogorov–Gabor polynomial presented in 19
m m m m m
y = a0 + ∑ ∑ aij xi x j + ∑m∑m∑ aijk xi x j xmk +m. . .m, (1)
i =0 j =0 i =
y = a0 + 0 j=0 k=a0ij xi x j +  aijk xi x j xk + ...,
i =0 j =0 i =0 j =0 k =0

where m is the number of variables, x(x1, x, x23,...) are vectors of input variables, an
a2, a3, ...) are vectors of parameters.
The second neural network is a neural network that uses radial basis functi
activation functions. This network is called the Radial Basis Function Network (R
[41–43]. The RFBN is very popular for function approximation, belt series predictio
Mathematics 2024, 12, 480 7 of 17

where m is the number of variables, x(x1 , x2 , x3 , . . .) are vectors of input variables, and a(a1 ,
a2 , a3 , . . .) are vectors of parameters.
The second neural network is a neural network that uses radial basis functions as acti-
vation functions. This network is called the Radial Basis Function Network (RBFN) [41–43].
The RFBN is very popular for function approximation, belt series prediction, and classifi-
cation. In such networks, it is very important to determine the number of neurons in the
hidden layer, as it strongly affects the complexity of the network and its generalisation
capabilities. In the hidden layer, each neuron has an activating function. The Gaussian
function, which has a parameter controlling the behaviour of the function, is the most
preferable activation function [34].
2
− ( x−b2)
f ( x ) = ae 2c , (2)

where a, b, and c are real numbers.

The third neural network is based on the multilayer perceptron network (MLPN) [35–37].
This type of network is known for requiring the presence of the desired result in the
training dataset. In our case, it is necessary to supply an additional value that will be the
correct result for the input data. In other words, this model relates the input values to
their result using historical weather data to predict the weather in the future, which is very
important in normalising the data [38].
MLPNs have shown the ability to find approximate solutions for extremely complex
problems. In particular, they are a universal function approximator, so they can be success-
fully used in the construction of regression models. Since classification can be considered
as a special case of regression when the output variable is categorical, MLPs can be used to
build classifiers [39,40].
Table 1 summarises the comparative performance of the neural networks proposed
for weather forecasting.

Table 1. Comparative characteristics of neural networks.

Input Output
Principle of Model Construction Mathematical Description
Information Information
D-PNN
Approximation of functions
described by differential equations m m m m m x(1), x(2), x(3),
y = a0 + ∑ ∑ aij xi x j + ∑ ∑ ∑ aijk xi x j xk + . . . x(n + 1)
that describe relationships between x(4) . . . x(n)
i =0 j =0 i =0 j =0 k =0
input parameters of the system
RBFN
Approximation of the unknown
solution by means of functions of a nk nk x(1), x(2), x(3),
yi (t) = wi0 + ∑ λij vl (t) + ∑ φ( v(t) − c j (t) ) x(n + 1)
special kind, whose arguments x(4) . . . x(n)
j =1 j =1
are distant
MLPN
Approximation of the unknown n x(1), x(2), x(3),
yi (t) = ϕ( ∑ wij x kj + bi ) x(n + 1)
solution using nonlinear functions x(4) . . . x(n)
j =1

4. Description of the Mathematical Model of Neural Network Grouping

Neural network grouping, also known as a neural network ensemble or a neural net-
work committee, is a technique in which multiple independent neural networks are trained
on the same task in order to improve the prediction quality or system reliability [41–43].
The application of neural network clustering has several advantages. First, different
neural networks can capture different aspects of the data or extract different characteristics.
By combining the predictions of these neural networks, the diverse aspects of the data
Mathematics 2024, 12, 480 8 of 17

can be taken into account and the generalisability of the model can be improved. Second,
grouping neural networks can help with robustness in terms of noise and outliers. If
one neural network makes an error using certain examples, other neural networks can
compensate for this error and improve the overall prediction accuracy [44–47].
The grouping of neural networks can be implemented in various ways:
- Bagging: This method involves training a set of independent neural networks on sub-
sets of training data obtained by selecting bootstrap examples. The predictions from
each neural network are then combined, for example, by majority voting or averaging.
- Boosting: Unlike bagging, boosting builds a sequence of neural networks, each of
which learns to correct the errors of previous networks. At each iteration, the weights
of the samples are adjusted to focus on the examples where the previous networks
made a mistake. Hence, subsequent neural networks focus on complex and poorly
classified examples [48].
- Stacking: This method is used when the predictions of several neural networks become
inputs to another model (a meta-model) that produces the final prediction. In this way,
the meta-model is trained to combine and utilise the predictions of different neural
networks to produce the best overall result.
Each of these approaches has its own characteristics and is suitable for different
scenarios and tasks. The choice of a particular neural network clustering method depends
on data characteristics, the required accuracy, computational resources, and other factors.
Neural network clustering is a powerful tool used to improve the prediction accuracy and
increase the confidence in decision-making. This method is actively used in various fields
including computer vision, natural language processing, and speech recognition [49].
Grouping neural networks can be used to predict weather parameters, such as tem-
perature, humidity, and wind speed. This is an important task that helps with planning
and decision-making in various fields including agriculture, energy, urban planning, and
tourism. One way of applying neural network grouping for weather forecasting is by
using the bagging method. In this case, different neural networks are trained on different
subsets of the original data with different characteristics (e.g., different time intervals and
geographical areas). The forecasts from each neural network are then combined to produce
the final weather forecast. This approach can help to account for different features and
nuances of weather conditions in different areas [50–52].
When applying neural network boosting, a sequence of neural networks learns to
predict weather parameters by correcting the errors of previous neural networks. At each
iteration, the sample weights or errors of the previous networks are used to emphasise
areas where the previous neural networks made an error. In this way, more complex and
difficult-to-predict situations can be handled by subsequent neural networks. When neural
network stacking is applied to weather forecasting, forecasts from multiple neural networks
become inputs to a meta-model that produces the final weather forecast. The meta-model
is trained to combine and utilise the forecasts from different neural networks to produce
the best overall result [53,54].
An important aspect in weather forecasting is the use of various input features, such
as data from weather stations, satellite observations, geographical data, and historical data.
Grouping neural networks allows us to combine information from these different sources,
which can improve the quality and accuracy of forecasts [55].
All these methods of grouping neural networks for weather forecasting require large
amounts of data and computational resources. In addition, it is important to carefully select
the structure and parameters of neural networks and train them on a sufficient amount of
diverse data to achieve good results [56,57].
The development of more powerful computing resources and the advancement of
deep learning have made grouping neural networks increasingly popular and successful in
the field of weather forecasting. However, it is always necessary to consider the nature of
the complexity of weather processes and the limitations of modelling. Weather forecasts by
Mathematics 2024, 12, 480 9 of 17

their nature remain tentative and there remains the possibility that an error will appear in
the result, especially in the case of long-term forecasts.
In view of this, neural network grouping is defined by an approach to building a self-
learning machine in which a finite number of neural networks is trained to solve the same
task. This approach originates from a paper by Hansen and Salamon [58], which shows
that the neural network system can be significantly improved by the grouping approach,
which means that the predictions produced by such a machine are much more accurate. In
general, the grouping of neural networks can be divided into two steps: training several
neural networks and then combining and processing the predictions of each. The result
of such a system is the averaged value of the outputs of each neural network separately,
combined with a function describing the comparative deviation of values obtained at the
Mathematics 2024, 12, x FOR PEER REVIEW 10 of 18
training stage relative to each other. The results of such systems significantly improve the
accuracy of predictions [59,60]. In this paper, a new approach for training these systems
will be considered. The weighting coefficients are proportional to the corresponding output
neural network
values. The essenceproduces
of themore accurate
approach is toforecasts.
determineLet us consider
which an example.
neural network One may
produces more
suppose forecasts.
accurate that there Let
areus two neural an
consider networks
example. thatOnehavemaytosuppose
performthat a simple
there classification
are two neural
task. If thethat
networks input
haveis to
1, perform
then the aoutput
simpleisclassification
1; if the input is 0,If then
task. the output
the input is 1, thenis 0.the
Let the
output
neural
is networks’
1; if the input isoutput
0, thenbe the0.6 and 0.9,
output is 0.respectively,
Let the neural at anetworks’
certain step. output In this
be 0.6case,
andthe 0.9,
second machine
respectively, at a receives muchInmore
certain step. reliable
this case, thedata because
second machine0.9 isreceives
closer to 1. more reliable
much
data Backpropagation
because 0.9 is closer networks
to 1. set the initial weighting factors randomly to reduce the
standard deviation [61,62].
Backpropagation The difference
networks in initial
set the initial weighting
weighting factorscoefficients
randomly gives different
to reduce the
results. Therefore, grouping neural networks integrates these independent
standard deviation [61,62]. The difference in initial weighting coefficients gives different networks to
improveTherefore,
results. the generalisation
grouping ability.
neuralThis methodintegrates
networks also guarantees an increase innetworks
these independent accuracyto
in terms of
improve thea generalisation
standard deviation [63–66].
ability. This method also guarantees an increase in accuracy in
termsInofthis paper, we
a standard propose[63–66].
deviation a grouping of nonlinear leading networks created using a
constructive algorithm.
In this paper, In constructive
we propose a grouping algorithms,
of nonlinear theleading
numbernetworksof neurons in hidden
created using a
layers is initially
constructive small and
algorithm. then gradually
In constructive increases.
algorithms, theHence,
numberinofconstructive
neurons in hidden algorithms,
layers
theinitially
is skills acquired
small andby thegradually
then network increases.
before increasing
Hence, inthe number ofalgorithms,
constructive neurons are thepre-
skills
served.
acquired by the network before increasing the number of neurons are preserved.
Constructive
Constructive algorithms
algorithmsdiffer differinintheir
theirrules
rulesfor
forsetting
settingparameter
parametervalues valuesininthe thenewnew
neurons added to the the network:
network:
- Parameter
Parameter values
valuesare arerandom
randomnumbers
numbersfrom fromaagiven
givenrange;
range;
- Values
Values ofof synaptic
synaptic weights
weightsof ofthe
thenew
newneuron
neuronare aredetermined
determinedby bysplitting
splittingone oneofofthe
the
old
old neurons.
neurons.
The first rule
rule does
doesnot notrequire
requiresignificant
significantcomputation,
computation, but butitsits
useuseleadsleads to increase
to an an in-
crease
in in the value
the value of theof the error
error function
function after after
eacheach addition
addition of a of
new a new neuron.
neuron. AsAs a result
a result of of
the
the random
random assignment
assignment of parameter
of parameter values
values of new
of new neurons,
neurons, a redundancy
a redundancy in the
in the num-of
number
ber of hidden
hidden layer neurons
layer neurons may appear.
may appear. Neuron Neuron
splitting splitting
is devoid is devoid
of these oftwothese two disad-
disadvantages.
vantages.
The essenceTheofessence of thealgorithm
the splitting splitting algorithm is illustrated
is illustrated in Figurein 4. Figure 4.

Vector of
Figure 4. Vector of hidden
hidden layer
layerneuron
neuronweights
weightsand
andchanges
changescorresponding
correspondingtotoindividual
individual training
training
examples.

Figure 4 shows the vector of weights of the hidden layer neuron at some training
step and the vectors of weight changes corresponding to individual training examples.
The change vectors have two preferential directions and form a region in space that is
significantly different from the spherical one. The essence of the algorithm is to identify
Mathematics 2024, 12, 480 10 of 17

Figure 4 shows the vector of weights of the hidden layer neuron at some training
step and the vectors of weight changes corresponding to individual training examples.
The change vectors have two preferential directions and form a region in space that is
significantly different from the spherical one. The essence of the algorithm is to identify and
split such neurons. As a result of splitting, there are two neurons instead of the one initial
neuron in the network. The first of these neurons has a vector of weights, which is the sum
of the vector of weights of the original neuron, and vectors of changes in the weights of
one of the preferential directions. The summation of the vectors of weight changes in the
other preferential direction and the vector of weights of the original neuron results in the
synaptic weights of the second new neuron [64].
It is necessary to split neurons whose change vectors have two preferential directions
because the presence of such neurons leads to oscillations during training by backpropa-
gation. When training the method with an integral error function, the presence of such
neurons leads to a hit in the local minimum with a large error value.
The splitting algorithm includes the construction of a covariance matrix of vectors of
changes in synaptic weights and the calculation of eigenvectors and eigenvalues of the
obtained matrix using the iterative Oja algorithm, according to which stochastic gradient
lifting and Gramm–Schmidt orthogonalisation are performed [65].
It is necessary to consider a single neural network that is trained on some dataset.
Let x be the input vector that appears for the first time in this network and d be the
desired outcome. Values x and d represent the realisation of a random vector X and a
random variable D, respectively. Let F(x) be the input–output function realised using the
network. Then,

ED [(F(x) − E[D|X = x])D2 ] = BD (F(x)) + VD (F(x)), (3)

where E[D|X = x] is the mathematical expectation, BD (F(x)) is the square of the bias

BD (F(x)) = (ED [F(x)] − E[D|X = x])2 , (4)

and VD (F(x)) is the difference:

VD (F(x)) = ED [(F(x) − ED [F(x)])2 ]. (5)

The expectation ED on the set D is called the set covering the distribution of all training
data, such as input and output values, and the distribution of all initial conditions. There
are several ways to individually train a neural network and several ways to group the
output data. This paper assumes that the networks have the same configurations, but their
training starts from different initial conditions. An average simple grouping is used to
combine the outputs of a group of neural networks. Let ψ be the set of all initial conditions
and FI (x) be the average input–output value of the network functions. Then, by analogy
with Equation (3), we obtain

Eψ [(FI (x) − E[D|X = x])2 ] = Bψ (F(x)) + Vψ (F(x)), (6)

where Bψ (F(x)) is the square of the deviation defined by the set ψ:

Bψ (F(x)) = (Eψ [F(x)] − E[D|X = x])2 (7)

and Vψ (F(x)) is the corresponding difference:

Vψ (F(x)) = Eψ [(EI (x)Eψ [F(x)])2 ]. (8)

The mathematical expectation Eψ is taken for the set ψ.

Mathematics 2024, 12, 480 11 of 17

From the definition of set D, we can think of it as the set of initial conditions ψ and the
remaining set is denoted by D′ . By analogy with Equation (3), we obtain:

ED′ [(FI (x) − E[D|X = x])2 ] = BD′ (EI (x)) + VD′ (F(x)), (9)

where BD′ (FI (x)) is the square of the deviation defined by the set D′ :

BD′ (FI (x) = (ED′ [EI (x)] − E[D|X = x])2 (10)

and VD′ (FI (x)) is the corresponding difference:

VD′ = ED′ [(EI (x) − ED′ [EI (x)])2 ]. (11)

From the definition of sets D, ψ, and D′ , it is obvious that:

ED′ [FI (x)] = ED [F(x)]. (12)

It follows that Equation (10) can be rewritten in an equivalent form:

VD′ (FI (x)) = (ED [F(x)] − E[D|X = x])2 = BD (F(x)). (13)

Having the difference VD′ (FI (x)) from Equation (11), since the difference in a random
variable is equivalent to the RMS value of a random variable, let us subtract its square
of deviation:
VD′ (FI (x)) = ED′ [(F(x))2 ] − (ED′ [F(x)])2 = ED′ [(F(x))] (14)
or
VD (FI (x)) = ED [(F(x))2 ] − (ED [F(x)])2 . (15)
It is worth considering that the RMS value of function F(x) on set D must be greater
than or equal to the RMS value of function FI (x) on set D′ .

ED [(F(x))2 ] ≥ ED′ [(FI (x))]. (16)

Provided that there are Equations (14) and (15), we have:

VD′ (FI (x)) ≤ VD (F(x)). (17)

Hence, based on Equations (13) and (17), two conclusions can be drawn:
1. The bias of function FI (x) referring to multiple classification systems is exactly the
same as the bias of function F(x) referring to a single neural network.
2. The difference in function FI (x) is smaller than that in function F(x).

5. Review of the Results of the Considered Mathematical Models

The grouping of linear neural networks has shown improved performance with
respect to individual networks. This paper presents the grouping of nonlinear anticipatory
networks generated by a constructive algorithm. A similar approach has been applied to
the previously discussed D-PNN, RBGN, and MLP.
Statistically, data from Weather Underground personal weather stations and the US
National Digital Forecast Database (NDFD) were used to build this model. In addition,
Google’s remote learning machine was used to compare the results with existing products
on the market.
As a result, the input parameters were historical data obtained from the PWS, and the
result of training is a function of the dependence of these parameters on each other. Then,
when entering sufficient data in the request for a forecast into the future, we will obtain
the data we need. In order to build a forecast based on the above-mentioned mathematical
Mathematics 2024, 12, 480 12 of 17

models, we need to arrange the obtained data in the form of an array of enumerated
parameters (Table 2).

Table 2. Example of input data for training.

Temp City Daytime Day_of_Year Year Humidity% Wind Speed kph Pressure in mBar Weather Conditions
13.2 1 3 1 2022 18 6 1300 3
14.6 1 4 1 2022 19 7 1150 4
... ... ... ... ... ... ... ... ...

In this case, since the model operates using numerical data, cities and weather descrip-
tions were assigned specific indices, e.g., 1 in the city column means Los Angeles. In this
study, several training methods were applied by manipulating the input values. Table 3
shows the average deviations of the predicted temperature values.

Table 3. Average deviations of predicted temperature values.

Mathematics 2024, 12, x FOR PEER REVIEW

Time of Year Assemble D-PNN RBFN 13 of 18
MLPN
Winter 12% 18% 15% 16%
Spring 10% 16% 14% 13%
Summer Summer
8% 8%14% 14% 16% 16% 11%11%
Autumn 9% 13% 16% 12%
Autumn 9% 13% 16% 12%

Let us determine
determine how the models
models work
work for
for different
different seasons.
seasons. For this purpose,
purpose, we
enter data
data for
for 55years
yearsand
andcompare
comparethetheobtained
obtaineddata
datawith
withreal values
real forfor
values different periods
different peri-
(0:00–23:00 on 1on
ods (0:00–23:00 January 2022,2022,
1 January 0:00–23:00 on 1on
0:00–23:00 April 2022,2022,
1 April 0:00–23:00 on 1 on
0:00–23:00 July 2022,2022,
1 July and
0:00–23:00 on 1 October 2022). Figure 5 shows the predicted air temperature values
and 0:00–23:00 on 1 October 2022). Figure 5 shows the predicted air temperature values obtained
by each method
obtained by eachand the grouping
method method for
and the grouping 1 January
method for 12022.
January 2022.

Figure 5.
Figure 5. Graph
Graph of
of forecast
forecast values
values obtained
obtained by
by each
each method
method as
as of
of 11 January
January 2022.
2022.

buildingthe
When building theforecasting
forecastingsystem,
system,itsitsmain
main parameters
parameters were
were determined
determined (fore-
(forecast
cast horizon,
horizon, forecast
forecast period,
period, and forecast
and forecast interval).
interval). The forecast
The forecast period
period is theisbasic
the basic unit
unit of of
time
time for which the forecast is formed. The forecast horizon is the number of time
for which the forecast is formed. The forecast horizon is the number of time periods in the periods
in the future
future that
that the the forecast
forecast covers.covers. For example,
For example, there isthere is a forecast
a forecast for 10 weeks
for 10 weeks into theinto the
future
future with data for each week. In this case, the period is one week and the horizon is ten
weeks. The forecast interval is the frequency with which a new forecast is generated.
Usually, the forecast interval coincides with the forecast period. In this case, the forecast
is revised each period using the last period’s claim and other current information as a
baseline for the revised forecast. In our case, the forecast theoretical horizon is 1 year.
Mathematics 2024, 12, 480 13 of 17

with data for each week. In this case, the period is one week and the horizon is ten weeks.
The forecast interval is the frequency with which a new forecast is generated. Usually, the
forecast interval coincides with the forecast period. In this case, the forecast is revised each
period using the last period’s claim and other current information as a baseline for the
revised forecast. In our case, the forecast theoretical horizon is 1 year.
This work presents a grouping of nonlinear leading networks created using a con-
structive algorithm. A similar approach was applied to the previously considered D-PNN,
RBGN, and MLP. Therefore, the created method of neural network clustering has higher
accuracy in all seasons and has an acceptable accuracy of 91%. The pro-forecast accuracy
of the other linear neural networks showed the following results of maximum–maximum
error: for MLPN it is 13%, for D-PNN it is 15%, and for RBFN it is 11%. The computation
time required to obtain a monthly forecast is 4 min. So, predicting the meteorological
forecast of temperature changes for a year would take 48 min. Hence, it is possible to realise
a forecasting model in a real-time system.
The forecast quality of the weather neural network model has increased by about
2 percent in the last two years (2021–2022). Table 4 shows the rankings of temperature
prognostic models in 2021–2022 for different countries, with a forecast accuracy of 180 h.
To date, the accuracy of most existing predictive weather models is limited to 120 h, while
we managed to increase this forecast period to 180 h.

Table 4. Ranking of temperature predictive models in 2021–2022. Forecast accuracy at 180 h.

Accuracy Years 2021/2022, %

№ Country
Assemble D-PNN RBFN MLPN
1 USA 89.1/91.4 75.65/77.3 79.21/80.99 77.4/79.2
2 Canada 87.3/90.1 73.9/76.5 77.43/80.1 75.6/78.3
3 Europe 88.7/91.6 74.8/77.35 78.32/80.99 76.5/79.2
4 UK 89.4/90.0 75.65/76.5 79.21/80.1 77.4/78.3
5 Australia 87.2/90.3 73.95/76.5 77.43/80.1 75.6/78.3
6 Germany 88.8/90.7 74.8/76.5 78.32/80.1 76.5/78.3
7 Japan 86.2/88.0 73.1/74.8 76.54/78.4 74.8/76.6
8 China 87.9/88.2 73.95/74.8 77.43/78.3 75.6/76.6
9 Korea 88.3/90.8 74.8/76.5 78.32/80.1 76.5/78.3
10 Russia 86.5/90.2 73.1/76.5 76.54/80.1 74.8/78.3

The neural network grouping method is an approach that combines the forecasts of
multiple neural networks to produce a more accurate and stable weather forecast. This
method uses an ensemble of neural networks that are trained based on historical weather
data and other relevant parameters. One of the key aspects of this method is the use of
different training algorithms for each neural network in the ensemble. This can include
different neural network architectures, various hyperparameters, and various optimisation
algorithms. This approach allows for a variety of models that can better capture the
characteristics and complexities of different weather seasons.
To achieve a higher forecast accuracy in all weather seasons, the neural network
clustering method can apply the averaging or weighting of the contributions of each neural
network. This allows us to smooth out possible errors in individual models and take into
account different forecast scenarios for different weather seasons. One of the advantages of
the neural network clustering method is its ability to adapt to changes in weather conditions.
The ability to train and combine models based on current and updated data allows the
method to respond quickly to weather changes and provide high forecast accuracy even in
new seasons.
Of course, achieving such forecast accuracies requires the preparation of high-quality
data and the careful tuning of parameters. An optimal choice of models, the optimisation of
hyperparameters, and the use of a large amount of data allow us to achieve the best results.
It should be noted that the neural network clustering method may have the potential
Mathematics 2024, 12, 480 14 of 17

for application in other areas requiring accurate forecasting, such as financial markets,
transportation systems, and energy. Its ability to combine forecasts from multiple models
may be useful for improving the forecast accuracy in other areas as well. In this work,
we determined that the neural network clustering method is a powerful tool for weather
forecasting that provides higher forecast accuracy in all weather seasons and is able to adapt
to changes in weather conditions. The application of this method can lead to more reliable
and accurate weather forecasts in different areas and has significant practical application.
Therefore, models built using the proposed neural network grouping method can
account for a larger number of dependent variables than that of individual neural models,
which ultimately improves their accuracy. The grouping of neural networks is an effective
tool for working with satellite data in the form of large amounts of weather data, which
was shown by the example of the classification of forecast values of objects (Figure 5).
Currently, we are developing clustered weather models for such phenomena as snow,
cloudiness, ice, and short-term precipitation forecasting based on the applied method
of neural network clustering using satellite data and numerical predictive models. All
approaches described in this paper were tested using specially generated datasets by
experienced decoding specialists. The current algorithms accepted in real operational
practice were also compared. The successful results from testing the models presented in
this paper allowed us to implement them in the operations of the Novosibirsk weather
centre of SIC “Climat”. There were almost no limitations in applying our model in different
locations and weather zones in terms of data, as the proposed model involves the processing
of large datasets.

6. Conclusions
In this paper, an algorithm for the construction of meteorological forecast models using
a grouping of neural networks was developed and investigated. The algorithm for building
a mathematical model for predicting future states of meteorological system parameters
based on differential polynomials, radial basis functions, multilayer perceptrons, and
groupings of neural networks was considered.
The model built using the proposed method of neural network clustering allowed us
to consider a larger number of dependent variables than that of individual neural models,
which ultimately improved their accuracy. Clustering neural networks is an effective and
promising method for processing meteorological information in the form of large datasets
on various weather factors, including snow, cloudiness, ice, and short-term precipitation,
based on the applied method of neural network clustering and the creation of numerical
prognostic models.
Based on the numerical experiment, it can be concluded that the combination of
mathematical modelling and “correct” input data related to weather phenomena can make
the meteorological forecast model more accurate. In the future, we will continue to study
meteorological forecast models in order to not only improve the order and accuracy of
input data but also change the mathematical basis for building the model itself.

Author Contributions: Conceptualization, B.V.M. and N.V.M.; methodology, S.N.S. and E.A.E.;
software, M.Q.; validation, S.N.S. and E.A.E.; formal analysis, S.N.S. and E.A.E.; investigation, D.V.V.;
resources, D.V.V.; data curation, D.V.V.; writing—original draft preparation, B.V.M. and N.V.M.;
writing—review and editing, B.V.M. and N.V.M.; visualization, M.Q. All authors have read and
agreed to the published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: The data presented in this study are available from the corresponding
authors upon reasonable request.
Conflicts of Interest: The authors declare no conflicts of interest.
Mathematics 2024, 12, 480 15 of 17

References
1. Baldauf, M.; Seifert, A.; Forstner, J.; Majewski, D.; Raschendorfer, M.; Reinhardt, T. Operational convective-scale numerical
weather prediction with the COSMO model: Description and sensitivities. Mon. Wea. Rev. 2011, 139, 3887–3905. [CrossRef]
2. Brisson, E.; Blahak, U.; Lucas-Picher, P.; Purr, C.; Ahrens, B. Contrasting lightning projection using the lightning potential index
adapted in a convection-permitting regional climate model. Clim. Dyn. 2021, 57, 2037–2051. [CrossRef]
3. Voitovich, E.V.; Kononenko, R.V.; Konyukhov, V.Y.; Tynchenko, V.; Kukartsev, V.A.; Tynchenko, Y.A. Designing the Optimal
Configuration of a Small Power System for Autonomous Power Supply of Weather Station Equipment. Energies 2023, 16, 5046.
[CrossRef]
4. Kleyko, D.; Rosato, A.; Frady, E.P.; Panella, M.; Sommer, F.T. Perceptron Theory Can Predict the Accuracy of Neural Networks.
IEEE Trans. Neural Netw. Learn. Syst. 2023, 1–15. [CrossRef] [PubMed]
5. Stephan, K.; Schraff, C. Improvements of the operational latent heat nudging scheme used in COSMO-DE at DWD. COSMO
Newsl. 2008, 9, 7–11.
6. Steppeler, J.; Doms, G.; Schaettler, U.; Bitzer, H.-W.; Gassmann, A.; Damrath, U.; Gregoric, G. Meso-gamma scale forecasts using
the nonhydrostatic model LM. Meteorol. Atmos. Phys. 2003, 82, 75–96. [CrossRef]
7. Armenta, M.; Jodoin, P.-M. The Representation Theory of Neural Networks. Mathematics 2021, 9, 3216. [CrossRef]
8. Bengio, Y.; Goodfellow, I.; Courville, A. Deep Learning; MIT Press: Cambridge, MA, USA, 2015.
9. Meng, L.; Zhang, J. IsoNN: Isomorphic Neural Network for Graph Representation Learning and Classification. arXiv 2019,
arXiv:1907.09495. [CrossRef]
10. Sorokova, S.N.; Efremenkov, E.A.; Qi, M. Mathematical Modeling the Performance of an Electric Vehicle Considering Various
Driving Cycles. Mathematics 2023, 11, 2586. [CrossRef]
11. Dozat, T. Incorporating Nesterov momentum into Adam. In Proceedings of the ICLR Workshop, San Juan, Puerto Rico, 2–4 May
2016; Volume 1, pp. 2013–2016.
12. Xie, S.; Kirillov, A.; Girshick, R.; He, K. Exploring Randomly Wired Neural Networks for Image Recognition. In Proceedings of
the IEEE/CVF International Conference on Computer Vision, Seoul, Republic of Korea, 27 October–2 November 2019.
13. He, K.; Zhang, X.; Ren, S.; Sun, J. Delving Deep into Rectifiers: Surpassing Human-level Performance on ImageNet Classification.
arXiv 2015, arXiv:1502.01852. Available online: https://arxiv.org/pdf/1502.01852.pdf (accessed on 1 December 2023).
14. Kukartsev, V.V.; Martyushev, N.V.; Kondratiev, V.V.; Klyuev, R.V.; Karlina, A.I. Improvement of Hybrid Electrode Material
Synthesis for Energy Accumulators Based on Carbon Nanotubes and Porous Structures. Micromachines 2023, 14, 1288. [CrossRef]
15. He, K.; Zhang, X.; Ren, S.; Sun, J. Deep Residual Learning for Image Recognition. In Proceedings of the IEEE Conference on
Computer Vision and Pattern Recognition, Las Vegas, NV, USA, 27–30 June 2016; pp. 770–778.
16. Kalnay, E. Atmospheric Modelling, Data Assimilation and Predictability; Cambridge University Press: Cambridge, UK, 2003.
17. Kazakova, E.; Rozinkina, I.; Chumakov, M. Verification of results of the working technology SNOWE for snow water equivalent
and snow density fields determination as initial data for COSMO model. COSMO Newsl. 2016, 16, 25–36.
18. Chernykh, N.; Mikhalev, A.; Dmitriev, V.; Tynchenko, V.; Shutkina, E. Comparative Analysis of Existing Measures to Reduce
Road Accidents in Western Europe. In Proceedings of the 2023 22nd International Symposium INFOTEH-JAHORINA, INFOTEH,
Sarajevo, Bosnia and Herzegovina, 15–17 March 2023. [CrossRef]
19. Krasnopolsky, V.M.; Lin, Y. A neural network nonlinear multimodel ensemble to improve precipitation forecasts over continental
US. Adv. Meteorol. 2012, 2012, 649450. [CrossRef]
20. Marzban, C. A neural network for post-processing model output: ARPS. Mon. Wea. Rev. 2003, 131, 1103–1111. [CrossRef]
21. Warner, T.T. Numerical Weather and Climate Prediction; Cambridge University Press: Cambridge, UK, 2010.
22. Ye, C.; Zhao, C.; Yang, Y.; Fermuller, C.; Aloimonos, Y. LightNet: A Versatile, Standalone Matlab-based Environment for Deep
Learning. arXiv 2016. Available online: https://arxiv.org/pdf/1605.02766.pdf (accessed on 1 December 2023).
23. Zurada, J.M. Introduction to Artificial Neural Systems; PWS: New York, NY, USA, 1992.
24. Goyal, M.; Goyal, R.; Lall, B. Learning Activation Functions: A New Paradigm of Understanding Neural Networks. arXiv 2019,
arXiv:1906.09529.
25. Burgers, G.; Van Leeuwen, P.J.; Evensen, G. Analysis scheme in the ensemble Kalman filter. Mon. Weather. Rev. 1998, 126, 1719–1724.
[CrossRef]
26. Dey, R.; Chakraborty, S. Convex-hull & DBSCAN clustering to predict future weather. In Proceedings of the 2015 International
Conference and Workshop on Computing and Communication (IEMCON), Vancouver, BC, Canada, 15–17 October 2015; pp. 1–8.
[CrossRef]
27. Saima, H.; Jaafar, J.; Belhaouari, S.; Jillani, T.A. Intelligent methods for weather forecasting: A review. In Proceedings of the 2011
National Postgraduate Conference, Perak, Malaysia, 19–20 September 2011; pp. 1–6. [CrossRef]
28. Salman, A.G.; Kanigoro, B.; Heryadi, Y. Weather forecasting using deep learning techniques. In Proceedings of the International
Conference on Advanced Computer Science and Information Systems (ICACSIS), Depok, Indonesia, 10–11 October 2015;
pp. 281–285. [CrossRef]
29. Singh, N.; Chaturvedi, S.; Akhter, S. Weather forecasting using machine learning algorithm. In Proceedings of the 2019
International Conference on Signal Processing and Communication (ICSC), Noida, India, 7–9 March 2019; pp. 171–174. [CrossRef]
Mathematics 2024, 12, 480 16 of 17

30. Sela, J.G. The Implementation of the Sigma Pressure Hybrid Coordinate into the GFS. Office Note (National Centers for
Environmental Prediction (U.S.)). 2009; p. 461. Available online: https://repository.library.noaa.gov/view/noaa/11401 (accessed
on 1 December 2023).
31. Weather Research & Forecasting Model (WRF) Mesoscale & Microscale Meteorology Laboratory. NCAR. Available online:
https://www.mmm.ucar.edu/models/wrf (accessed on 1 December 2023).
32. Yonekura, K.; Hattori, H.; Suzuki, T. Short-term local weather forecast using dense weather station by deep neural network. In
Proceedings of the IEEE International Conference on Big Data (Big Data), Seattle, WA, USA, 10–13 December 2018; pp. 1683–1690.
[CrossRef]
33. Bauer, P.; Thorpe, A.; Brunet, G. The quiet revolution of numerical weather prediction. Nature 2015, 525, 47–55. [CrossRef]
34. Buschow, S.; Friederichs, P. Local dimension and recurrent circulation patterns in long-term climate simulations. arXiv 2018,
1803, 11255. [CrossRef]
35. C3S: ERA5: Fifth Generation of ECMWF Atmospheric Reanalyses of the Global Climate, Copernicus Climate Change Service
Climate Data Store (CDS). 2017. Available online: https://cds.climate.copernicus.eu/cdsapp#!/home (accessed on 7 June 2019).
36. Compo, G.P.; Whitaker, J.S.; Sardeshmukh, P.D.; Matsui, N.; Allan, R.J.; Yin, X.; Gleason, B.E.; Vose, R.S.; Rutledge, G.; Bessemoulin,
P.; et al. The twentieth century reanalysis project. Q. J. R. Meteor. Soc. 2011, 137, 1–28. [CrossRef]
37. Coors, B.; Paul Condurache, A.; Geiger, A. Spherenet: Learning spherical representations for detection and classification in
omnidirectional images. In Proceedings of the European Conference on Computer Vision (ECCV), Munich, Germany, 8–14
September 2018; pp. 518–533.
38. Dee, D.P.; Uppala, S.M.; Simmons, A.; Berrisford, P.; Poli, P.; Kobayashi, S.; Andrae, U.; Balmaseda, M.; Balsamo, G.; Bauer, P.;
et al. The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Q. J. R. J. R. Meteor. Soc. 2011,
137, 553–597. [CrossRef]
39. Dueben, P.D.; Bauer, P. Challenges and design choices for global weather and climate models based on machine learning. Geosci.
Model Dev. 2018, 11, 3999–4009. [CrossRef]
40. Faranda, D.; Messori, G.; Yiou, P. Dynamical proxies of North Atlantic predictability and extremes. Sci. Rep. 2017, 7, 41278.
[CrossRef] [PubMed]
41. Faranda, D.; Messori, G.; Vannitsem, S. Attractor dimension of time-averaged climate observables: Insights from a low-order
ocean-atmosphere model. Tellus A 2019, 71, 1–11. [CrossRef]
42. Fraedrich, K.; Jansen, H.; Kirk, E.; Luksch, U.; Lunkeit, F. The Planet Simulator: Towards a user-friendly model. Meteorol. Z. 2005,
14, 299–304. [CrossRef]
43. Freitas, A.C.M.; Freitas, J.M.; Todd, M. Hitting time statistics and extreme value theory. Probab. Theory Rel. 2010, 147, 675–710.
[CrossRef]
44. Tynchenko, V.; Kurashkin, S.; Murygin, A.; Bocharov, A.; Seregin, Y. Software for optimization of beam output during electron
beam welding of thin-walled structures. Procedia Comput. Sci. 2022, 200, 843–851. [CrossRef]
45. Kukartsev, V.V.; Tynchenko, V.S.; Bukhtoyarov, V.V.; Wu, X.; Tyncheko, Y.A.; Kukartsev, V.A. Overview of Methods for Enhanced
Oil Recovery from Conventional and Unconventional Reservoirs. Energies 2023, 16, 4907. [CrossRef]
46. Krasnopolsky, V.M.; Fox-Rabinovitz, M.S. Complex hybrid models combining deterministic and machine learning components
for numerical climate modelling and weather prediction. Neural Netw. 2006, 19, 122–134. [CrossRef]
47. Filina, O.A.; Sorokova, S.N.; Efremenkov, E.A.; Valuev, D.V.; Qi, M. Stochastic Models and Processing Probabilistic Data for
Solving the Problem of Improving the Electric Freight Transport Reliability. Mathematics 2023, 11, 4836. [CrossRef]
48. Krasnopolsky, V.M.; Fox-Rabinovitz, M.S.; Belochitski, A.A. Using ensemble of neural networks to learn stochastic convection
parameterisations for climate and numerical weather prediction models from data simulated by a cloud resolving model. Adv.
Artif. Neural Syst. 2013, 2013, 485913. [CrossRef]
49. Lorenz, E.N. Deterministic nonperiodic flow. J. Atmo. Sci. 1963, 20, 130–141. [CrossRef]
50. Das, P.; Manikandan, S.; Raghavendra, N. Holomorphic aspects of moduli of representations of quivers. Indian J. Pure Appl. Math.
2019, 50, 549–595. [CrossRef]
51. Konyukhov, V.Y.; Oparina, T.A.; Zagorodnii, N.A.; Efremenkov, E.A.; Qi, M. Mathematical Analysis of the Reliability of Modern
Trolleybuses and Electric Buses. Mathematics 2023, 11, 3260. [CrossRef]
52. Filina, O.A.; Tynchenko, V.S.; Kukartsev, V.A.; Bashmur, K.A.; Pavlov, P.P.; Panfilova, T.A. Increasing the Efficiency of Diagnostics
in the Brush-Commutator Assembly of a Direct Current Electric Motor. Energies 2024, 17, 17. [CrossRef]
53. Frankle, J.; Carbin, M. The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks. In Proceedings of the
International Conference on Learning Representations, New Orleans, LA, USA, 6–9 May 2019.
54. Nooteboom, P.D.; Feng, Q.Y.; López, C.; Hernández-García, E.; Dijkstra, H.A. Using network theory and machine learning to
predict El Nino. Earth Syst. Dynam. 2018, 9, 969–983. [CrossRef]
55. O’Gorman, P.A.; Dwyer, J.G. Using Machine Learning to Parameterize Moist Convection: Potential for Modeling of Climate,
Climate Change, and Extreme Events. J. Adv. Model. Earth Syst. 2018, 10, 2548–2563. [CrossRef]
56. Golik, V.I.; Brigida, V.; Kukartsev, V.V.; Tynchenko, Y.A.; Boyko, A.A.; Tynchenko, S.V. Substantiation of Drilling Parameters for
Undermined Drainage Boreholes for Increasing Methane Production from Unconventional Coal-Gas Collectors. Energies 2023,
16, 4276. [CrossRef]
Mathematics 2024, 12, 480 17 of 17

57. Volneikina, E.; Kukartseva, O.; Menshenin, A.; Tynchenko, V.; Degtyareva, K. Simulation-Dynamic Modeling of Supply Chains
Based on Big Data. In Proceedings of the 2023 22nd International Symposium INFOTEH-JAHORINA, INFOTEH 2023, Sarajevo,
Bosnia and Herzegovina, 15–17 March 2023. [CrossRef]
58. Poli, P.; Hersbach, H.; Dee, D.P.; Berrisford, P.; Simmons, A.J.; Vitart, F.; Laloyaux, P.; Tan, D.G.H.; Peubey, C.; Thépaut, J.-N.; et al.
ERA-20C: An Atmospheric Reanalysis of the Twentieth Century. J. Clim. 2016, 29, 4083–4097. [CrossRef]
59. Semenova, E.; Tynchenko, V.; Chashchina, S.; Suetin, V.; Stashkevich, A. Using UML to Describe the Development of Software
Products Using an Object Approach. In Proceedings of the 2022 IEEE International IOT, Electronics and Mechatronics Conference
(IEMTRONICS), Toronto, ON, Canada, 1–4 June 2022. [CrossRef]
60. Scher, S. Toward Data-Driven Weather and Climate Forecasting: Approximating a Simple General Circulation Model with Deep
Learning. Geophys. Res. Lett. 2018, 45, 12616–12622. [CrossRef]
61. Martyushev, N.V.; Malozyomov, B.V.; Sorokova, S.N.; Efremenkov, E.A.; Valuev, D.V.; Qi, M. Review Models and Methods for
Determining and Predicting the Reliability of Technical Systems and Transport. Mathematics 2023, 11, 3317. [CrossRef]
62. Scher, S. Videos for Weather and climate forecasting with neural networks: Using GCMs with different complexity as studyground.
Zenodo 2019. [CrossRef]
63. Scher, S. Code and data for Weather and climate forecasting with neural networks: Using GCMs with different complexity as
study-ground. Zenodo 2019. [CrossRef]
64. Martyushev, N.V.; Malozyomov, B.V.; Kukartsev, V.V.; Gozbenko, V.E.; Konyukhov, V.Y.; Mikhalev, A.S.; Kukartsev, V.A.;
Tynchenko, Y.A. Determination of the Reliability of Urban Electric Transport Running Autonomously through Diagnostic
Parameters. World Electr. Veh. J. 2023, 14, 334. [CrossRef]
65. Scher, S.; Messori, G. Predicting weather forecast uncertainty with machine learning. Q. J. R. Meteor. Soc. 2018, 144, 2830–2841.
[CrossRef]
66. Schneider, T.; Lan, S.; Stuart, A.; Teixeira, J. Earth System Modeling 2.0: A Blueprint for Models That Learn from Observations
and Targeted High-Resolution Simulations. Geophys. Res. Lett. 2017, 44, 12396–12417. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.