trigram

This is a table of type trigram and their frequencies. Use it to search & browse the list to learn more about your study carrel.

trigram	frequency
the number of	666
of the network	401
in order to	341
as well as	311
based on the	308
in the network	288
one of the	242
q q q	240
due to the	216
such as the	206
in terms of	198
the spread of	185
the case of	147
in this paper	146
a number of	144
there is a	143
the effect of	142
the impact of	136
the size of	133
can be used	133
structure of the	129
size of the	125
the effects of	123
on the other	123
the role of	121
the performance of	115
the epidemic threshold	115
the development of	113
a set of	112
according to the	111
the other hand	110
the structure of	108
be used to	107
the evolution of	104
the fact that	101
as shown in	100
the analysis of	99
the use of	96
total number of	95
analysis of the	95
the dynamics of	93
shown in figure	92
the presence of	92
the value of	91
the probability of	89
part of the	89
the total number	88
is based on	88
the importance of	87
the context of	86
in complex networks	85
nodes in the	85
in which the	85
in this case	85
number of nodes	84
of the epidemic	84
some of the	84
large number of	84
in the case	83
in the same	83
in a network	82
most of the	82
the network structure	82
that can be	81
it can be	81
shown in fig	80
in the context	80
the results of	80
is the number	79
number of infected	79
the network is	77
it has been	77
each of the	76
related to the	76
with respect to	75
the fraction of	75
of infectious diseases	75
the degree of	74
with the same	74
wearing condition identification	74
the relationship between	73
the distribution of	72
value of the	72
of the population	71
as a result	71
of the most	70
of the same	70
different types of	70
compared to the	70
of the disease	70
in this section	69
in this work	68
are shown in	67
of a network	66
in social networks	66
the quality of	65
is defined as	64
depends on the	64
to study the	64
is shown in	64
at the same	63
can be seen	63
in the following	62
in the human	62
the problem of	62
the same time	62
in this study	61
there is no	61
changes in the	61
depending on the	61
of complex networks	61
the influence of	60
is given by	60
in addition to	60
well as the	60
performance of the	60
the level of	59
evolution of the	59
distribution of the	59
a large number	59
of the nodes	59
is important to	59
the rest of	58
in the population	58
shown in table	58
need to be	57
the neural network	57
with each other	57
it is not	57
the set of	57
at time t	56
the study of	56
it is also	56
which can be	56
the time of	55
can also be	55
of a node	54
a variety of	54
in other words	54
it is important	54
the existence of	54
function of the	53
nature of the	53
severe acute respiratory	53
likely to be	53
a function of	53
acute respiratory syndrome	52
the sr network	52
we use the	52
is that the	52
social media use	52
rest of the	52
the human brain	51
of the data	51
the probability that	51
to improve the	51
it is possible	51
that it is	51
show that the	51
in the first	50
a network of	50
the input data	50
the majority of	50
of the system	50
internet of things	50
the shannon entropy	50
to each other	50
the nature of	50
of neural networks	50
which is a	49
the network of	49
the process of	49
to understand the	49
found that the	48
social network analysis	48
all rights reserved	48
to evaluate the	48
of the model	47
is one of	47
we find that	47
of social media	47
more likely to	47
community structure in	47
this is a	47
this article is	47
a neural network	47
is used to	46
results of the	46
the ability to	46
figure shows the	46
the network and	46
of the two	46
to reduce the	46
small number of	46
in the field	46
the degree distribution	45
be used for	45
of the infection	45
the form of	45
understanding of the	44
shows that the	44
similar to the	44
a series of	44
in case of	44
values of the	44
of infected individuals	44
of the main	44
the amount of	44
based on a	44
note that the	44
in the literature	43
on the network	43
and the time	43
this means that	43
and can be	43
role in the	43
i is the	43
the end of	43
network can be	43
the average degree	43
number of edges	42
a social network	42
which is the	42
used in the	42
the complex network	42
and it is	42
of infected nodes	42
n is the	42
of the social	42
network representation learning	42
the beginning of	41
the output layer	41
are able to	41
of nodes in	41
a total of	41
of the covid	41
needs to be	41
to have a	41
in the form	41
to identify the	41
as the number	41
because of the	41
the results are	41
i and j	41
we assume that	41
referred to as	41
number of links	40
allows us to	40
one or more	40
between the two	40
convolutional neural network	40
is possible to	40
a result of	40
associated with the	40
terms of the	40
focus on the	40
the sis model	40
the field of	39
are used to	39
the transmission of	39
the integration of	39
during the covid	39
data about the	39
development of the	39
has to be	39
in the study	39
as a function	39
neural network model	39
the emergence of	38
the effectiveness of	38
of the networks	38
epidemic spreading in	38
the input layer	38
the frequency of	38
properties of the	38
networks can be	38
the construction of	38
a measure of	38
of the proposed	38
state of the	38
that there is	38
included in the	38
the concept of	38
defined as the	38
the application of	38
is able to	38
the adjacency matrix	37
for each of	37
of the total	37
to the same	37
and so on	37
networks in the	37
may not be	37
means that the	37
convolutional neural networks	37
respect to the	37
the result of	37
used for the	37
is likely to	37
the data from	37
have been proposed	36
data from the	36
forward and backward	36
depend on the	36
corresponds to the	36
can be found	36
is consistent with	36
dynamics of the	36
of social networks	36
complex network theory	36
the contact network	36
the sum of	36
is characterized by	36
at d p	36
the diffusion of	36
description of the	35
backward disruption propagation	35
quality of the	35
proteins in the	35
the risk of	35
the ratio of	35
it is necessary	35
the structure and	35
are based on	35
in response to	35
on the number	35
the values of	35
in contrast to	35
units in the	35
for a given	35
relationship between the	35
the united states	35
protected by copyright	34
to determine the	34
the proportion of	34
the formation of	34
of the three	34
authors proposed a	34
the basis of	34
information about the	34
this can be	34
article is protected	34
the units in	34
degree of the	34
more than one	34
of the number	34
is protected by	34
the choice of	34
to be the	34
that the network	34
time of exposure	34
in such a	34
the social network	34
so that the	34
to be a	33
an increase in	33
network structure and	33
the accuracy of	33
representation of the	33
it is a	33
characterized by a	33
a collection of	33
are likely to	33
the purpose of	33
a small number	33
we found that	33
spread of infectious	33
the definition of	33
spread of the	33
wireless sensor networks	33
of the human	33
all of the	33
of an epidemic	33
can lead to	33
we propose a	33
average number of	32
the road network	32
we need to	32
behavior of the	32
results show that	32
to the number	32
of node i	32
involved in the	32
the whole network	32
is necessary to	32
at least one	32
network and the	32
an example of	32
the behavior of	32
public research institutions	32
be able to	32
used as a	32
deep neural networks	32
version of the	32
can be applied	31
for the analysis	31
condition identification network	31
in recent years	31
the robustness of	31
compared with the	31
in the data	31
characteristics of the	31
it is worth	31
result of the	31
beginning of the	31
this is the	31
use of the	31
the efficiency of	31
was used to	31
that of the	31
is due to	31
and backward disruption	31
the identification of	31
the information content	31
node in the	31
network inference methods	31
has been shown	31
the topology of	31
the average number	31
nodes and edges	30
in a population	30
have been developed	30
is the total	30
number of individuals	30
machine learning techniques	30
to analyze the	30
for this purpose	30
the rate of	30
of the first	30
results in the	30
find that the	30
node i is	30
a power law	30
the perspective of	30
is the most	30
of these networks	30
to find the	30
that are not	30
the state of	30
there are no	30
have to be	29
impact on the	29
as long as	29
each time step	29
individuals in the	29
proportional to the	29
at the beginning	29
innovation and financing	29
the possibility of	29
with regard to	29
consistent with the	29
have shown that	29
the data set	29
a case study	29
been used to	29
the density of	29
leading to the	29
to be more	29
of nodes and	29
in the future	29
caused by the	29
the detection of	29
the difference between	29
artificial neural networks	29
neural networks in	29
there are several	29
the eighteenth century	29
it is the	29
be found in	29
a network with	29
the other two	28
by means of	28
the properties of	28
we show that	28
and analysis of	28
average degree of	28
from the perspective	28
an overview of	28
the lack of	28
the strength of	28
in the previous	28
with a probability	28
there is an	28
nodes of the	28
a network is	28
a consequence of	28
in social media	28
in relation to	28
in the present	28
this paper is	28
in the next	28
the model is	28
showed that the	28
to investigate the	28
indicates that the	28
tail risk network	28
of this paper	28
on a network	28
refers to the	28
basic reproduction number	28
to predict the	28
in the social	28
of the node	28
the design of	28
a and b	28
end of the	28
in the number	28
and in the	28
as a consequence	27
as in the	27
it would be	27
in this article	27
have the same	27
equal to the	27
stored in the	27
the giant component	27
same number of	27
point of view	27
on social media	27
the community structure	27
in the second	27
from the same	27
the same number	27
of surface science	27
with the highest	27
it may be	27
average path length	27
on the basis	27
fraction of the	27
are more likely	27
number of connections	27
a range of	27
fact that the	27
a node is	27
it to the	27
on complex networks	27
the interactions between	27
is organized as	27
in this way	27
we consider the	27
number of susceptible	27
estimation of the	27
we used the	27
with community structure	27
that there are	27
that have been	27
the nodes in	27
of infectious disease	26
each of these	26
of the graph	26
can be represented	26
made of small	26
there are many	26
organized as follows	26
model for the	26
liu et al	26
of the current	26
been shown to	26
is an important	26
network is a	26
networks and the	26
conditions of the	26
is related to	26
has also been	26
infectious disease transmission	26
according to their	26
of the community	26
on the data	26
in networks with	26
supported by the	26
wide range of	26
determined by the	26
but it is	26
of the time	26
the need for	26
a wide range	26
the coefficient of	26
the spreading of	26
the absence of	26
the network representation	26
of the world	26
to address the	26
the most important	26
the supply chain	26
at each time	26
closely related to	26
into account the	26
study of the	26
the transmission rate	26
in figure a	26
is equal to	26
the network topology	26
to the network	26
the training set	25
the propagation of	25
we want to	25
if there is	25
of the contact	25
taking into account	25
of the road	25
in the last	25
network analysis of	25
in a number	25
we conclude that	25
financing and innovation	25
the introduction of	25
the creation of	25
due to its	25
review of the	25
be seen in	25
in some cases	25
illustrated in fig	25
we use a	25
there are two	25
as part of	25
the ability of	25
as described in	25
probability that a	25
of all the	25
of the input	25
social network size	25
the complexity of	25
for this reason	25
many of the	25
the ppi network	25
contribute to the	25
to deal with	25
parts of the	25
of the experiment	25
model to study	25
network with a	25
of the other	25
to estimate the	25
knowledge of the	25
have been used	25
of the original	24
be considered as	24
to be used	24
relative to the	24
friends of friends	24
between eo and	24
difference between the	24
addition to the	24
in the past	24
within the network	24
of the paper	24
as the data	24
number of contacts	24
derived from the	24
by using the	24
the random network	24
the likelihood of	24
as an example	24
as it is	24
characterized by the	24
information on the	24
from the first	24
the sir model	24
seems to be	24
we do not	24
the availability of	24
of a disease	24
followed by the	24
the paper is	24
world health organization	24
that in the	24
at the end	24
in the face	24
a subset of	24
which in turn	24
offline social network	24
network in the	24
spread of disease	24
social big data	24
of this study	24
in the lungs	24
of the neural	24
is not a	24
of each node	24
the outcome of	24
the face of	24
in the us	24
the goal of	24
due to their	24
the latent space	24
target data stream	23
to simulate the	23
can be defined	23
complexity of the	23
accuracy of the	23
response to the	23
used to identify	23
the basic reproduction	23
the output of	23
from the input	23
and the number	23
in a large	23
network is the	23
to make the	23
also known as	23
length of the	23
the network was	23
cases at d	23
be applied to	23
a combination of	23
wang et al	23
is the probability	23
the reliability of	23
between two nodes	23
were used to	23
model can be	23
of the real	23
probability of a	23
a complex network	23
present in the	23
important role in	23
of the brain	23
majority of the	23
to obtain a	23
respiratory syndrome coronavirus	23
nodes in a	23
this is not	23
for the network	23
corresponds to a	23
science and technology	23
to describe the	23
to model the	23
a comparison of	23
lead to the	23
properties of a	23
the characteristics of	23
during the training	23
in this context	23
the relevance of	23
role of social	23
increase in the	23
the area of	23
networks have been	22
serve as a	22
is stored in	22
to explore the	22
lead to a	22
we focus on	22
network model to	22
neurons and synapses	22
a group of	22
to understand how	22
to detect the	22
an important role	22
the largest eigenvalue	22
we refer to	22
that the proposed	22
weights of the	22
can be considered	22
on the one	22
a lack of	22
in the presence	22
case of the	22
the speed of	22
while there is	22
tend to be	22
to control the	22
paper is organized	22
of the degree	22
effect of the	22
the type of	22
anxiety and depression	22
the potential to	22
the network to	22
found to be	22
in the limit	22
be used as	22
protein interaction networks	22
information content of	22
for a specific	22
a review of	22
known as the	22
the after peak	22
for all the	22
in the graph	22
to the fact	22
on the same	22
leads to a	22
the understanding of	22
edges in the	21
indicate that the	21
a factor of	21
the immune response	21
chen et al	21
to compute the	21
it should be	21
approaches have been	21
network of networks	21
the range of	21
an innovation system	21
and in a	21
is represented by	21
they do not	21
to be able	21
of the problem	21
is determined by	21
the interplay between	21
to this end	21
the training process	21
is the case	21
an analysis of	21
as can be	21
expressed in the	21
networks with community	21
each node is	21
considered to be	21
is used in	21
in this sense	21
social networks in	21
cannot be predicted	21
to provide a	21
which means that	21
given by the	21
there has been	21
results in a	21
this is because	21
the center of	21
social networks and	21
features of the	21
defined by the	21
in line with	21
in temporal networks	21
set of nodes	21
the original network	21
shown to be	21
targeted by sars	21
through the network	21
effects of the	21
data can be	21
of network structure	21
overview of the	21
networks based on	21
the one hand	21
to measure the	21
be defined as	21
shown that the	21
of the different	21
two types of	21
input data and	21
especially in the	21
been applied to	21
proteins targeted by	21
from the network	21
of the algorithm	21
as a whole	21
and machine learning	21
by a factor	21
a sequence of	21
are displayed in	21
more and more	21
close to the	21
effect on the	21
the combination of	21
of a new	21
can only be	20
is defined by	20
for each node	20
of the inter	20
in a given	20
in the process	20
in which a	20
the stability of	20
the infection rate	20
in our case	20
based on their	20
in the model	20
and the network	20
transmit it to	20
t is the	20
the need to	20
of the challenges	20
do not have	20
the tail risk	20
in the current	20
of the stochastic	20
network with the	20
galaz et al	20
of susceptible nodes	20
are associated with	20
we define the	20
out of the	20
is difficult to	20
to solve the	20
largest eigenvalue of	20
the first and	20
in the range	20
results showed that	20
is associated with	20
generated by the	20
depression and anxiety	20
to that of	20
to a large	20
gene expression data	20
affected by the	20
may be a	20
greater than the	20
models have been	20
time of the	20
has been used	20
is used for	20
at the time	20
focused on the	20
for the same	20
the management of	20
take into account	20
layer and a	20
learning techniques in	20
of exposure to	20
around the world	20
structure in the	19
probability of the	19
we compare the	19
goal is to	19
the relation between	19
the order of	19
the increase of	19
online social networks	19
pair of nodes	19
of the sr	19
such as a	19
are interested in	19
the correlation between	19
the course of	19
in the two	19
the description of	19
to consider the	19
to achieve a	19
that using the	19
is that while	19
the entire network	19
birth and death	19
is a big	19
experimental data about	19
the connection between	19
to the ability	19
which has been	19
apply machine learning	19
impact of the	19
the desired output	19
the granular material	19
protein interaction network	19
the function of	19
output of the	19
challenges in the	19
with the prediction	19
the peak of	19
using the network	19
a survey of	19
has been a	19
observed in the	19
of severe acute	19
in the scale	19
result with the	19
nodes that are	19
the steady state	19
it does not	19
social media and	19
the middle of	19
we investigate the	19
of experimental data	19
the notion of	19
fraction of nodes	19
found in the	19
a model for	19
members of the	19
is set to	19
in the area	19
the basis for	19
there are also	19
we observe that	19
the port community	19
relationships such as	19
increases with the	19
and control of	19
more hidden layers	19
on temporal networks	19
we can find	19
structure and the	19
the advantage of	19
of the giant	19
the real world	19
to increase the	19
taken into account	19
to note that	19
in the output	19
differences in the	19
the spreading dynamics	19
we show the	19
can be estimated	19
used in this	19
between any two	19
for the purpose	19
of the virus	19
a lot of	19
this kind of	19
what is the	19
supply chain network	19
the outbreak of	19
in each layer	19
as the zipf	19
the implementation of	18
scientists and engineers	18
the most common	18
nodes with a	18
number of important	18
the cost of	18
big amount of	18
in the united	18
is not the	18
makes the network	18
assumed to be	18
the challenges in	18
collective dynamics of	18
depicted in fig	18
is smaller than	18
of ly e	18
the impacts of	18
may lead to	18
the help of	18
in random networks	18
of aconitine alkaloids	18
comparison of the	18
to represent complex	18
the hidden layer	18
in the sense	18
middle of the	18
is similar to	18
on the spread	18
number of neighbors	18
in old at	18
the network more	18
a better understanding	18
by the network	18
view of the	18
between the contact	18
and aboav scaling	18
robustness of the	18
to assess the	18
law degree distribution	18
importance of the	18
proposed or protocol	18
a big amount	18
order to predict	18
for infectious disease	18
meta path detection	18
network has three	18
consequence of the	18
the proposed or	18
to the output	18
with the network	18
we are interested	18
predicted from the	18
aboav scaling laws	18
to process the	18
be interpreted as	18
the network dynamics	18
a study of	18
delivered to the	18
that a node	18
corresponding to the	18
in the public	18
of the following	18
the proposed algorithm	18
is composed of	18
role of the	18
case of a	18
by the number	18
some of these	18
such as surface	18
has not been	18
of the eighteenth	18
to create a	18
the selection of	18
the data about	18
structure and function	18
as soon as	18
structure and dynamics	18
the authors also	18
input layer with	18
described in the	18
is proportional to	18
with the help	18
networking in the	18
and for the	18
discussed in the	18
the knowledge acquired	18
for a single	18
at the level	18
we have to	18
wetting properties of	18
of a given	18
there are some	18
at a time	18
fedex and ups	18
measure of the	18
the result with	18
evaluation of the	18
with a frequency	18
the human interactome	18
of the entire	18
networks that are	18
has been recently	18
these parameters are	18
which deals with	18
the relationships between	18
a framework for	18
are determined in	18
or more hidden	18
of the internet	18
dependent on the	18
the most popular	18
dependencies between the	18
a model of	18
be noted that	18
of an innovation	18
ability to represent	18
and financial return	18
the efficacy of	18
and social network	18
dynamics and control	18
a key role	18
neural network has	18
the droplets used	17
friction and the	17
until the results	17
a ductile iron	17
training process is	17
for the spread	17
a more applied	17
associated with a	17
prediction for each	17
connect the units	17
represent complex input	17
determined in an	17
typical ann model	17
synaptic weights of	17
that while there	17
in a single	17
about the material	17
or conditions of	17
disease control and	17
coefficient of friction	17
deals with such	17
in the middle	17
hydrophobic and superhydrophobic	17
composite including complex	17
in the synaptic	17
to the units	17
and function of	17
roughness or conditions	17
over several stages	17
of verified users	17
relevant to the	17
is surface science	17
of the proteins	17
of contacting surfaces	17
tribology deals with	17
applied a multilayer	17
have focused on	17
optimized design of	17
small rigid particles	17
with units representing	17
so as to	17
the results obtained	17
given by where	17
in the hidden	17
multilayer perception neural	17
nodal connections leading	17
generating the prediction	17
study the wetting	17
of the second	17
and engineers often	17
them over several	17
predicting new outcomes	17
remains a data	17
scaling relationships such	17
and retrieving acquired	17
hidden layer compute	17
the reference model	17
to apply machine	17
composite materials for	17
example of a	17
of various materials	17
and making adjustments	17
the wear rate	17
the disruption propagation	17
is highly empirical	17
varying connection weights	17
studies is that	17
of small rigid	17
compute their activations	17
finally delivered to	17
is larger than	17
see that the	17
and synapses in	17
about the frictional	17
adhesion is called	17
and output layer	17
which may be	17
are given in	17
repellent properties of	17
contacting surfaces as	17
of metallic composite	17
and water contact	17
the synaptic weights	17
related to colloidal	17
for the optimized	17
material composition and	17
by a single	17
they cannot be	17
different from the	17
storing and retrieving	17
a typical ann	17
context of the	17
that do not	17
colloidal science is	17
important scaling relationships	17
parameters of surfaces	17
and surface roughness	17
in the tribological	17
representing the input	17
and a non	17
it possible to	17
of novel hydrophobic	17
science is surface	17
as the coefficient	17
the units with	17
accurate in predicting	17
the length of	17
model complex neurons	17
is also a	17
of the territory	17
in detail in	17
convert them over	17
graphite composite including	17
is worth noting	17
as the granular	17
representation for colloidal	17
by examining individual	17
linear transfer function	17
incorporate a series	17
the material composition	17
computer models somewhat	17
models learn by	17
the network in	17
topology of the	17
units representing the	17
was set to	17
connected to the	17
to the other	17
that model complex	17
on top of	17
by examples and	17
weights until the	17
surface roughness or	17
nodes that model	17
for link prediction	17
transfer function and	17
spread of a	17
complex neurons and	17
are suited for	17
a way that	17
and they cannot	17
with parameters of	17
details of the	17
process is stored	17
are connected to	17
connections leading to	17
of surfaces such	17
has interconnected nodes	17
network more accurate	17
the proposed protocol	17
will be discussed	17
deal with parameters	17
and on the	17
surfaces as the	17
and transmit it	17
the prediction for	17
smaller than the	17
the period of	17
that the number	17
layer with units	17
process the input	17
for each record	17
connection weights until	17
communities in the	17
examples and training	17
responsible for the	17
has three parts	17
surface scientists and	17
is close to	17
and after peak	17
somewhat resembling neural	17
the optimized design	17
interdisciplinary area is	17
metallic composite materials	17
from social media	17
the members of	17
retrieving acquired knowledge	17
stages into the	17
of the new	17
characteristics of contacting	17
series of functions	17
with varying connection	17
activations based on	17
be predicted from	17
strong community structure	17
default mode network	17
branch of surface	17
change in the	17
in wetting experiments	17
an experimental manner	17
each node in	17
the first physical	17
the neural networks	17
examining individual records	17
seen in the	17
in this area	17
function and transmit	17
perception neural network	17
often deal with	17
levitating droplet clusters	17
the innovation system	17
the larger the	17
tribology remains a	17
comparing the result	17
conclude that using	17
anns incorporate a	17
such characteristics of	17
amount of experimental	17
area closely related	17
and levitating droplet	17
results are finally	17
in the brain	17
parameters are determined	17
of friction and	17
surfaces such as	17
the tribological studies	17
droplets used in	17
as compared to	17
belonging to the	17
and systemic risk	17
of important scaling	17
novel hydrophobic and	17
of a ductile	17
several stages into	17
water contact angle	17
synapses in the	17
properties of various	17
network representation for	17
rate at which	17
this interdisciplinary area	17
consists of a	17
the critical point	17
understanding these correlations	17
the data is	17
into the desired	17
to predict surface	17
for colloidal systems	17
to capture the	17
of traditional medicine	17
composition and surface	17
of functions to	17
in predicting new	17
would like to	17
deals with friction	17
and engineering components	17
input layer and	17
models somewhat resembling	17
suggested to apply	17
when compared to	17
is used as	17
in an experimental	17
applied branch of	17
to ensure that	17
of the information	17
different levels of	17
they are suited	17
more applied branch	17
colloidal crystals made	17
area is highly	17
are finally delivered	17
complex dependencies between	17
to colloidal science	17
except for the	17
is called tribology	17
whether or not	17
model has interconnected	17
of nodes with	17
suited for storing	17
properties of metallic	17
and the wear	17
their activations based	17
first physical principles	17
and adhesion is	17
correlations allows predicting	17
surface wetting properties	17
been recently suggested	17
allows predicting water	17
surface properties of	17
of the pandemic	17
with such characteristics	17
transcripts and proteins	17
are involved in	17
layers connect the	17
and superhydrophobic materials	17
and surface properties	17
anns learn by	17
as surface roughness	17
table shows the	17
exposure to liquid	17
units with varying	17
of a large	17
more accurate in	17
three parts or	17
ann models learn	17
of the protein	17
with the number	17
of the droplets	17
functions to process	17
ann model has	17
predict surface wetting	17
tribological studies is	17
be explained by	17
the system is	17
including complex dependencies	17
surface free energy	17
an infected individual	17
the network has	17
center of the	17
and convert them	17
data and convert	17
recently suggested to	17
the contact angle	17
these correlations allows	17
interconnected nodes that	17
resembling neural networks	17
behavior and in	17
free behavior and	17
edges of the	17
rather than the	17
driven inductive science	17
the limit of	17
acquired during the	17
learn by examining	17
crystals made of	17
layer compute their	17
they found that	17
techniques in order	17
design of novel	17
be used in	17
parts or layers	17
models can be	17
are computer models	17
for storing and	17
hidden layers connect	17
suggest that the	17
learn by examples	17
a multilayer perception	17
since ann models	17
the previous section	17
the magnitude of	17
the wetting properties	17
a survey on	17
knowledge acquired during	17
used in wetting	17
materials for the	17
another area closely	17
training makes the	17
engineers often deal	17
other types of	16
value at risk	16
the spectral radius	16
the increase in	16
the remainder of	16
are represented by	16
to a specific	16
our understanding of	16
the details of	16
any of the	16
and n is	16
neural networks and	16
detail in the	16
the results for	16
attention to the	16
to overcome the	16
proved to be	16
the features of	16
can then be	16
the same as	16
neural networks are	16
p and q	16
contribution to the	16
between the nodes	16
the guarantee network	16
assume that the	16
relation to the	16
infectious disease outbreaks	16
to compare the	16
the network with	16
the best of	16
we aim to	16
of individuals in	16
conducting polymer nanocomposite	16
this section we	16
the user can	16
zhang et al	16
human social networks	16
the immune system	16
to obtain the	16
the degree sequence	16
to be considered	16
increase of the	16
adjacency matrix a	16
also be used	16
approach to the	16
k is the	16
of social network	16
of information and	16
resulting from the	16
the absorbing state	16
such that the	16
per unit time	16
the base station	16
the private sphere	16
to calculate the	16
that they are	16
the percentage of	16
to find a	16
be represented as	16
social network ties	16
all of these	16
to account for	16
that social media	16
are characterized by	16
and of the	16
of the device	16
of the research	16
the higher the	16
this work is	16
electrical conducting polymer	16
networks in a	16
time step t	16
in many cases	16
is a measure	16
along with the	16
of an outbreak	16
we describe the	16
insight into the	16
can be identified	16
such as those	16
interpersonal social network	16
are related to	16
implementation of the	16
to the original	16
of disease transmission	16
the shortest path	16
removed from the	16
proposed a new	16
is important for	16
the random walk	16
for the first	16
measured by the	16
degree distribution of	16
as illustrated in	16
account for the	16
is different from	16
of the art	16
neural networks for	16
the internet of	16
the data packet	16
is essential to	16
methods have been	16
od airport pairs	16
the disease dynamics	16
the idea of	16
ratio of the	16
the network are	16
network analysis and	16
this type of	16
the network as	16
we used a	16
to explain the	16
in the analysis	16
to the best	16
in a way	16
could also be	16
social networks are	16
presented in this	16
of susceptible individuals	16
entrepreneurial orientation and	16
where n is	16
ties with a	16
sum of the	16
to illustrate the	16
contributes to the	16
screening and referral	16
the ripple effect	15
the targeted immunization	15
of the underlying	15
the prediction of	15
the world health	15
li et al	15
is supported by	15
mathematical theory of	15
used as the	15
of systemic risk	15
to examine the	15
old at cells	15
construction of the	15
research on the	15
of coupling nodes	15
the rate at	15
stability of the	15
types of networks	15
observe that the	15
half of the	15
the time to	15
could not be	15
the hm model	15
in which each	15
better understanding of	15
correspond to the	15
the loss function	15
the university of	15
in one of	15
the critical value	15
of this article	15
of the individual	15
of disease spreading	15
offline social networks	15
spread of epidemic	15
the default mode	15
the community detection	15
is an open	15
various types of	15
of the average	15
for disease control	15
we have used	15
machine learning algorithms	15
in dynamic networks	15
in each of	15
outbreak and after	15
in the training	15
with the largest	15
make use of	15
should be noted	15
as the network	15
results are shown	15
is in the	15
networks with the	15
of this approach	15
network structure of	15
each node i	15
of the dynamics	15
of multilayer networks	15
the string database	15
to represent the	15
expected number of	15
of the power	15
they can be	15
cognitive distance between	15
described in section	15
be the case	15
interact with each	15
loan guarantee network	15
the scope of	15
results for the	15
the mathematical theory	15
to build a	15
a random network	15
the special path	15
number of layers	15
an introduction to	15
of big data	15
a discussion of	15
information of the	15
we see that	15
the literature on	15
of our knowledge	15
and the other	15
network structure is	15
to address this	15
a pair of	15
number of interactions	15
as we have	15
which indicates that	15
have a higher	15
epidemic threshold in	15
to generate a	15
is known to	15
the success of	15
be seen that	15
interactions between the	15
linked to the	15
the reconstruction of	15
is connected to	15
a node with	15
the supplementary material	15
random graph models	15
refer to the	15
in wireless sensor	15
in this research	15
nodes refer to	15
a node in	15
the next section	15
in this chapter	15
position of the	15
the topological structure	15
of the transmission	15
it is still	15
can be derived	15
of scaling in	15
obtained from the	15
by the authors	15
for the latter	15
of an individual	15
explained by the	15
neural network is	15
of the spanish	15
visual analytics for	15
two kinds of	15
to the covid	15
have also been	15
static and dynamic	15
for which the	15
the target data	15
focusing on the	15
the threshold of	15
of the adjacency	15
listed in table	15
are the most	15
influential spreaders in	15
the small world	14
average shortest path	14
development of a	14
in the main	14
the southern ocean	14
the training of	14
we analyze the	14
a change in	14
the expected number	14
power law distribution	14
traditional knowledge systems	14
of the n	14
which is defined	14
to characterize the	14
epidemic threshold is	14
an infectious disease	14
to use the	14
after peak stage	14
the edges of	14
there have been	14
was supported by	14
a convolutional neural	14
presented in the	14
the flow of	14
neural networks with	14
in the private	14
the duration of	14
it comes to	14
the network can	14
for the training	14
the theory of	14
of the process	14
we can see	14
network filtering methods	14
as mentioned above	14
been proposed to	14
the assumption that	14
based on this	14
the social brain	14
is the average	14
which corresponds to	14
of community structure	14
can be calculated	14
the potential of	14
the case for	14
on the structure	14
validity of the	14
with a higher	14
the ashwagandha network	14
dynamical processes on	14
better than the	14
with a high	14
application of the	14
provided by the	14
nodes i and	14
can be understood	14
the adoption of	14
if it is	14
is to be	14
a contact network	14
ml based devices	14
to focus on	14
the changes in	14
data on the	14
has the potential	14
has led to	14
immune response to	14
the evaluation of	14
by using a	14
on the left	14
each pair of	14
important to note	14
the time scale	14
approximation of the	14
was based on	14
of shortest paths	14
network of the	14
the mechanism of	14
the total population	14
the estimation of	14
old and young	14
the nodes with	14
for a long	14
infectious diseases and	14
is the same	14
would be a	14
with probability p	14
on how to	14
have not been	14
the activation of	14
with the increase	14
the location of	14
the probability distribution	14
to the spread	14
in the time	14
network as a	14
community detection in	14
in the absence	14
the two layers	14
exponential random graph	14
efficiency of the	14
the connections between	14
no more than	14
sterblom and bodin	14
parameters of the	14
of temporal networks	14
each other and	14
there are a	14
of social distancing	14
back to the	14
only in the	14
the systemic risk	14
with a small	14
wavelet based denoising	14
order to improve	14
differences between the	14
on the contrary	14
to make a	14
in the figure	14
of the immune	14
a period of	14
is referred to	14
to minimize the	14
on virus propagation	14
a long time	14
that this is	14
is crucial to	14
considered as a	14
the case in	14
can be very	14
key role in	14
in a similar	14
business network ties	14
input to the	14
between pairs of	14
best of our	14
of a population	14
b and c	14
use of a	14
of weak ties	14
the power law	14
network analysis in	14
network and its	14
of the sis	14
to be an	14
correlation between the	14
outcome of the	14
selected as the	14
in the real	14
that does not	14
than in the	14
the volume of	14
should not be	14
resulting in a	14
have been made	14
the extent of	13
shows the results	13
in the transmission	13
the final deconvolutional	13
raw movement data	13
of our approach	13
be represented by	13
different values of	13
emergence of scaling	13
od airport pair	13
materials and methods	13
the data in	13
network for the	13
can be achieved	13
it can also	13
to show that	13
can be obtained	13
we argue that	13
that cannot be	13
it is difficult	13
less likely to	13
human ppi network	13
of the global	13
j o u	13
of one health	13
the virtual robot	13
the forwarder list	13
the scalability of	13
of the roads	13
the coronavirus disease	13
on the degree	13
that they have	13
to the epidemic	13
of a single	13
be seen from	13
to better understand	13
to the study	13
in this model	13
gene ly e	13
which could be	13
network size and	13
structure in networks	13
social media on	13
in the original	13
component of the	13
on entrepreneurial entry	13
of disease outbreaks	13
c is the	13
small world network	13
u r n	13
quantized neural networks	13
of the full	13
we have a	13
produced by the	13
and there is	13
and the same	13
we consider a	13
of degree d	13
machine learning methods	13
expect that the	13
model and the	13
is called the	13
generative adversarial networks	13
a susceptible node	13
does not necessarily	13
estimates of the	13
be related to	13
the transmission tree	13
levels of clustering	13
the training data	13
the validated network	13
the label propagation	13
during the period	13
is the only	13
is done by	13
independent of the	13
r n a	13
of nodes that	13
while for the	13
l p r	13
is the degree	13
other aspects of	13
final deconvolutional layer	13
if the network	13
are presented in	13
when it comes	13
o u r	13
out to be	13
p r o	13
to support the	13
to develop a	13
of these two	13
at which the	13
on the relationship	13
widely used in	13
the next step	13
orders of magnitude	13
r o o	13
this work was	13
of the structure	13
of the patient	13
p i q	13
scaling in random	13
and its impact	13
extension of the	13
of biological networks	13
in the world	13
was carried out	13
east respiratory syndrome	13
the point of	13
is considered to	13
three types of	13
of such a	13
in the southern	13
p r e	13
light on the	13
o o f	13
in multiplex networks	13
it is clear	13
the disease transmission	13
from a network	13
will not be	13
the start of	13
the proposed approach	13
in both the	13
malighetti et al	13
topological properties of	13
this is also	13
the activity of	13
would be the	13
because of their	13
is the first	13
the simulation results	13
may also be	13
the shape of	13
the capacity of	13
n a l	13
the probabilities of	13
is not possible	13
system can be	13
applied to the	13
the copyright holder	13
process on a	13
for future research	13
are the same	13
the position of	13
a machine learning	13
together with the	13
this study is	13
the test set	13
the underlying network	13
a proxy for	13
for the spreading	13
a high level	13
the same degree	13
investors and researchers	13
in the research	13
the average time	13
number of samples	13
the nineteenth century	13
in the supplementary	13
outside the lungs	13
of network inference	13
is as follows	13
facial detection and	13
model of the	13
a deep learning	13
the network size	13
an infected node	13
in a graph	13
the experimental results	13
r is the	13
as they are	13
in the other	13
we introduce a	13
the weight of	13
its impact on	13
one can also	13
divided by the	13
different immunization strategies	13
can be written	13
reduction in the	13
middle east respiratory	13
processes in complex	13
worth noting that	13
of the port	13
the input image	13
links in the	13
can be modeled	13
can be easily	13
reflected in the	13
in the above	13
the disease is	13
with a single	13
the results show	13
all the other	13
as a proxy	13
according to a	13
the distance between	13
shape of the	13
analysis on the	13
in the s	13
the basic reproductive	13
the human protein	13
a l p	13
the process is	13
was able to	13
to the set	13
to the one	13
connections in the	13
the proposed model	13
this is an	13
a i j	13
into the network	12
implies that the	12
are described in	12
there were no	12
the testing set	12
can be observed	12
of the probability	12
the epidemic spreading	12
a role in	12
results indicate that	12
that most of	12
the node with	12
the question of	12
the benefits of	12
lead apa network	12
in real networks	12
large amount of	12
networks epidemic spreading	12
when it is	12
used to study	12
positive and negative	12
port of antwerp	12
in all cases	12
structures can be	12
network topology and	12
are in the	12
the real network	12
been widely used	12
is the largest	12
to the corresponding	12
a conceptual framework	12
structure of a	12
by comparing the	12
with the other	12
the population is	12
network based on	12
strength of the	12
to reach a	12
on a single	12
during the pandemic	12
on social networks	12
infectious disease dynamics	12
model based on	12
and right wing	12
is easy to	12
control and prevention	12
is required to	12
the validity of	12
the sense that	12
number of steps	12
in most of	12
the significance of	12
our goal is	12
note that this	12
one of these	12
will focus on	12
is greater than	12
were able to	12
the security of	12
up to the	12
of the whole	12
organizations and networks	12
the variation of	12
used for various	12
even though the	12
is with the	12
comes from the	12
of the target	12
approach can be	12
the following two	12
higher level of	12
states of the	12
the acquaintance method	12
reuse allowed without	12
impact of social	12
influential nodes in	12
the same functional	12
a description of	12
of iot devices	12
extent to which	12
of electrical conducting	12
is expected to	12
components of the	12
only a few	12
general facial recognition	12
right and right	12
hepatitis c virus	12
research in the	12
is also used	12
the main text	12
for a network	12
stage of the	12
number of doses	12
for the sis	12
from a single	12
the aim of	12
parameters such as	12
as the average	12
that is to	12
allow us to	12
we study the	12
of the final	12
the correlation length	12
are responsible for	12
from the data	12
a high degree	12
after peak stages	12
to be connected	12
higher pagerank in	12
we did not	12
to have an	12
the advantages of	12
the large number	12
in our model	12
the thermodynamic limit	12
has the highest	12
of the random	12
on systemic risk	12
reported in the	12
ssd mobilenet v	12
values in the	12
the usage of	12
have been widely	12
eigenvalue of the	12
or equal to	12
in the final	12
not only the	12
of contact networks	12
average of the	12
a tool for	12
of the existing	12
deal with the	12
power law degree	12
it is interesting	12
the contagion process	12
not affect the	12
on the dynamics	12
the same type	12
time to detection	12
these types of	12
from each other	12
natural language processing	12
the tendency of	12
immunization strategy is	12
has been widely	12
wt and ko	12
and entrepreneurial entry	12
be affected by	12
the population size	12
peak of the	12
choice of the	12
the stochastic system	12
we model the	12
in the dual	12
of disruption propagation	12
allowed without permission	12
also used for	12
difference in the	12
of the corresponding	12
as described above	12
can be in	12
to the development	12
networks have a	12
it could be	12
part of a	12
variations in the	12
the same hospital	12
the inclusion of	12
shown in the	12
is equivalent to	12
the extent to	12
on their own	12
at any given	12
than that of	12
no reuse allowed	12
traditional chinese medicine	12
rate of the	12
sterblom and sumaila	12
follows a power	12
as the input	12
highly connected nodes	12
the average of	12
be regarded as	12
actors in the	12
position in the	12
trained neural networks	12
aspects of the	12
we are able	12
number of studies	12
the architecture of	12
the public sphere	12
a network can	12
to test the	12
basic reproductive number	12
and vice versa	12
combined with the	12
it is found	12
the other methods	12
networks such as	12
to ensure the	12
outside of the	12
agreement with the	12
is to say	12
due to a	12
the work of	12
maximum number of	12
regarded as a	12
estimated using the	12
infectious diseases in	12
of human proteins	12
the entrepreneurial process	12
of the complex	12
defined as a	12
improve the performance	12
is a set	12
the connectivity of	12
in the period	12
using the same	12
a way to	12
are defined as	12
when there is	12
the central node	12
larger than the	12
multiple types of	12
from the previous	12
design of the	12
result in a	12
the capability of	12
defined as follows	12
can have a	12
subset of the	12
the contact networks	12
newly paved roads	12
between innovation and	12
density of infected	12
the lifetime of	12
the war of	12
captured by the	12
advantage of the	12
in a very	12
studies have shown	12
a very high	12
of action of	12
recurrent neural networks	12
is the author	12
level of the	12
virus propagation in	12
complex networks with	12
it will be	12
of protein interactions	12
can be computed	12
growth of the	12
and the second	12
in the system	12
on the analysis	12
the rough net	12
node v i	12
may be more	12
to get the	12
be attributed to	12
users in the	12
social contact networks	12
creative commons license	12
in materials science	12
these networks are	12
the first two	12
definition of the	12
network of verified	12
insights into the	12
is found that	12
and lead to	12
supervised learning task	12
information regarding the	12
to define the	12
for the epidemic	12
we propose to	12
a multiplex network	12
and social networks	12
of network science	12
of a system	12
by the complex	12
in the thermodynamic	12
the first step	12
of the top	12
located on the	12
the implications of	12
the opinion model	11
of the alters	11
as we will	11
a license to	11
in the development	11
with a total	11
long as the	11
shed light on	11
a fraction of	11
over the last	11
in the degree	11
well as a	11
include granular materials	11
mainly expressed in	11
of the acoustic	11
to all other	11
a network structure	11
can see that	11
an organizational field	11
combination of the	11
of the united	11
on the impact	11
a dynamical process	11
size of an	11
in the recent	11
existence of a	11
the epidemic spread	11
result in the	11
elements of the	11
high degree of	11
used to measure	11
the temporal network	11
consists of the	11
x for peer	11
a wide variety	11
certified by peer	11
in most cases	11
given that the	11
the node degree	11
it is assumed	11
on networks with	11
using the data	11
the two networks	11
t x and	11
the sake of	11
a list of	11
granted medrxiv a	11
ml based techniques	11
added to the	11
the top of	11
relevant for the	11
systemic risk spillovers	11
due to ageing	11
was not certified	11
supplementary figure s	11
the disease to	11
ly e in	11
sovereign bond yields	11
be seen as	11
of interest to	11
in the control	11
proceedings of the	11
the ideas of	11
used in materials	11
and the results	11
such as in	11
to each of	11
is required for	11
medrxiv a license	11
to generate the	11
resulted in a	11
information from the	11
the assumption of	11
performance of our	11
as seen in	11
distance between the	11
suggests that the	11
the markov chain	11
that the structure	11
of the contagion	11
to get a	11
number of states	11
content of a	11
of the susceptible	11
with a given	11
to the following	11
of epidemic disease	11
support vector machines	11
to determine whether	11
the vast majority	11
the authors declare	11
by peer review	11
the nodes and	11
of the blockchain	11
not possible to	11
a class of	11
to quantify the	11
is calculated as	11
map of the	11
transcriptomic and proteomic	11
of the connection	11
display the preprint	11
be written as	11
effects on the	11
be estimated using	11
epidemics on networks	11
of the th	11
of this work	11