quadgram

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

quadgram	frequency
in the united states	57
the number of tweets	46
in the context of	35
as well as the	34
the total number of	30
one of the most	28
the preprint in perpetuity	27
to display the preprint	27
has granted medrxiv a	27
license to display the	27
granted medrxiv a license	27
medrxiv a license to	27
a license to display	27
who has granted medrxiv	27
display the preprint in	27
on the other hand	26
made available under a	25
international license it is	25
can be used to	25
license it is made	25
it is made available	25
is made available under	25
copyright holder for this	24
holder for this preprint	24
were more likely to	24
the copyright holder for	24
is the author funder	23
in the case of	23
are more likely to	23
use of social media	22
available under a is	21
in the age of	21
under a is the	21
a is the author	21
in the field of	21
and the number of	21
preprint this version posted	20
for this preprint this	20
this preprint this version	20
the spread of the	20
of the number of	20
at the time of	20
as shown in fig	19
in the number of	18
is one of the	18
the rest of the	18
a smt immunity link	17
as a result of	16
likelihood of sharing the	16
spread of the virus	15
it is possible that	15
number of tweets in	15
in terms of the	15
it is important to	15
the extent to which	14
this version posted may	14
in support of h	14
the early stages of	14
the measure of concern	14
unified theory of brand	14
in the form of	14
it is possible to	14
the use of twitter	14
at the same time	14
content analysis of tweets	14
more likely to be	14
theory of brand equity	14
more likely to share	14
with higher levels of	13
analysis of tweets during	13
the h n outbreak	13
the world health organization	13
the number of times	13
total number of tweets	12
a high level of	12
right and right wing	12
the use of social	12
the geocov tweets dataset	12
with the help of	12
reported likelihood of sharing	12
can be seen in	12
and local tv viewership	12
during the h n	12
levels of twitter followers	11
and team financial performance	11
per number of population	11
the purpose of this	11
the analysis of the	11
as shown in figure	11
stages of the outbreak	11
to the fact that	11
a large number of	11
is based on the	11
personal versus news classification	11
for each of the	11
the size of the	11
on the basis of	11
presence of altmetric data	11
as one of the	11
deaths per number of	11
x x x x	10
can be found in	10
the time of the	10
the age of twitter	10
the unified theory of	10
should be noted that	10
it should be noted	10
as the number of	10
the distribution of the	10
as a function of	10
social media use in	10
the use of the	10
natural language processing and	10
a case study of	10
in the use of	10
due to the fact	10
the highest number of	10
the case of the	10
for the purpose of	10
different stages of the	10
be more likely to	10
a resilient communication ecosystem	10
that most of the	10
attendance and local tv	10
types of altmetric data	10
of tweets during the	10
the context of the	10
higher levels of twitter	10
the majority of the	10
abuse sent to mps	9
the diffusion patterns of	9
on the relationship between	9
as hot research topics	9
at the university of	9
is worth noting that	9
with the number of	9
to the best of	9
of abuse for the	9
in order to identify	9
tweets during the h	9
it is necessary to	9
the daily number of	9
we are able to	9
can be defined as	9
it is worth noting	9
are likely to be	9
analysis of twitter data	9
all abuse sent to	9
of the tweets in	9
the number of replies	9
to the number of	9
the results of the	9
the democrat and republican	9
the new york times	9
the evolution of the	9
in the early stages	9
pandemics in the age	9
an overview of the	9
social and machine learning	9
of all abuse sent	9
of this study is	9
abuse for the period	9
used in this study	9
the fact that the	9
a l p r	8
likely to be true	8
of likelihood of sharing	8
number of tweets and	8
n a l p	8
state of alarm and	8
are shown in table	8
by the number of	8
on the number of	8
j o u r	8
early stages of the	8
l p r e	8
we are interested in	8
in the social media	8
r n a l	8
validated network of verified	8
o u r n	8
r o o f	8
the number of topics	8
in the same way	8
the united states during	8
u r n a	8
the vast majority of	8
is consistent with the	8
this is consistent with	8
a greater likelihood of	8
the official accounts of	8
was significantly predicted by	8
to changes in the	8
the validated network of	8
the political orientation of	8
the frames used in	8
p r o o	8
for disease control and	8
data from social media	8
we were able to	8
of social media use	8
used in the analysis	8
the public health frame	8
on social media platforms	8
of words in the	8
in addition to the	8
of social media users	8
hot research topics in	8
the performance of the	8
the number of retweets	8
the number of users	8
is labeled as a	8
on the part of	8
the context of covid	8
members of the public	7
of hot research topics	7
to jurisdictional claims in	7
that are related to	7
the united states and	7
the positive relationship between	7
were likely to be	7
of the pandemic in	7
of tweets in each	7
the us presidential election	7
remains neutral with regard	7
note springer nature remains	7
the name of the	7
best of our knowledge	7
the start of the	7
during a health crisis	7
in published maps and	7
state of the art	7
between non seasonal and	7
it is not clear	7
refuting a smt immunity	7
at the end of	7
media coverage of the	7
of altmetric data for	7
claims in published maps	7
centers for disease control	7
a review of the	7
the fuzzy accuracy assessment	7
higher likelihood of sharing	7
of words associated with	7
social sciences and humanities	7
between april st and	7
information cascades on twitter	7
published maps and institutional	7
can be applied to	7
springer nature remains neutral	7
in the fields of	7
april st and th	7
more likely to have	7
to mps between april	7
political orientation of users	7
the frequency with which	7
a crucial role in	7
sent to mps between	7
between in season and	7
the best of our	7
the twitter rest api	7
is organized as follows	7
neutral with regard to	7
mps between april st	7
there is a need	7
of the most popular	7
jurisdictional claims in published	7
regard to jurisdictional claims	7
in times of crisis	7
in each hashtag group	7
the onset of the	7
have the potential to	7
the presence of altmetric	7
the same way as	7
network of verified users	7
nature remains neutral with	7
can also be used	7
we found that the	7
the text of the	7
platforms such as twitter	7
the three disinformation items	7
of the impact of	7
can be used as	7
the beginning of the	7
with regard to jurisdictional	7
the relationship between non	7
the content of the	7
marketing assets and team	7
st and th inclusive	7
we observe that the	7
the cov tweets dataset	7
as a tool for	7
maps and institutional affiliations	7
the relationship between the	7
the early days of	7
are positively related to	6
of social media platforms	6
increase in the number	6
in major league baseball	6
the news update frame	6
to one of the	6
the end of the	6
reported a greater likelihood	6
a higher likelihood of	6
during the recovery period	6
belief that the stories	6
significantly predicted by higher	6
during the early days	6
during the maintenance stage	6
in the analysis was	6
hashtags into six main	6
analysis of the tweets	6
higher levels of attention	6
on the same day	6
disease control and prevention	6
response to the covid	6
of the pandemic on	6
from the twitter api	6
group the hashtags into	6
in natural language processing	6
as a percentage of	6
of the presence of	6
is the number of	6
of tweets in the	6
in order to understand	6
network analysis of twitter	6
social media data for	6
as well as a	6
the most frequently used	6
the political frame was	6
people are more likely	6
marketing assets are positively	6
in the era of	6
the number of confirmed	6
media briefing on covid	6
during the period of	6
based on the analysis	6
a political news story	6
during the measles outbreak	6
using natural language processing	6
we would like to	6
identify hot research topics	6
the media briefing on	6
the number of covid	6
from march to april	6
play a crucial role	6
the behavior of users	6
social media as a	6
political news story online	6
the directed validated network	6
been shown to be	6
remarks at the media	6
a weaker relationship between	6
related to social distancing	6
on the analysis of	6
as shown in table	6
information on social media	6
news story online that	6
the pandemic in the	6
of the most important	6
into six main categories	6
in the absence of	6
the hashtags into six	6
communities in large networks	6
as more likely to	6
high level of attention	6
at the media briefing	6
the number of mentions	6
there are a number	6
a description of the	6
are a number of	6
stages of the pandemic	6
the impact of the	6
the chain of events	6
the number of increased	6
an analysis of the	6
greater likelihood of sharing	6
opening remarks at the	6
between democrats and republicans	6
through the lens of	6
labeled as a personal	6
an example of a	6
number of confirmed cases	6
during different stages of	6
the time of this	6
early days of the	6
be used as a	6
on the use of	6
the state of the	6
likelihood of having seen	6
the seriousness of the	6
in the face of	6
this version posted june	6
that the majority of	6
assets are positively related	6
of a resilient communication	6
a number of other	6
the tweets in the	6
story attributed to source	6
the use of hashtags	6
in addition to that	6
detecting influenza epidemics using	6
the progression of the	6
the characteristics of the	6
themselves as more likely	6
understand the chain of	6
identifying hot research topics	5
the spread of misinformation	5
semantic network analysis of	5
support of h b	5
international conference on advances	5
used in social media	5
analysis was carried out	5
united states during the	5
the authors declare that	5
the probability of observing	5
the percentage of the	5
were also more likely	5
on the one hand	5
the set of labeled	5
had seen them before	5
is important to note	5
detection on social media	5
presented in this paper	5
deaths in each country	5
the time was made	5
over the course of	5
majority of the tweets	5
publications in the fields	5
during the early stages	5
perception of the pandemic	5
as can be seen	5
would like to thank	5
relationship between in season	5
daily number of deaths	5
of coverage and intensity	5
the first week of	5
of altmetric data in	5
also more likely to	5
the activity of the	5
it is worth mentioning	5
the field of social	5
diffusion patterns of covid	5
was not certified by	5
is the set of	5
certified by peer review	5
the most prominent news	5
intensity of the pandemic	5
the date of the	5
the goal of this	5
that the number of	5
not take into account	5
which was not certified	5
was used as a	5
with at least one	5
of communities in large	5
the surgery of trauma	5
in an effort to	5
as discussed in chapter	5
for the surgery of	5
onset of the covid	5
it is labeled as	5
in the validated network	5
tested positive for covid	5
alpha for this measure	5
a wide number of	5
the steel blue community	5
was found to be	5
is based on a	5
twitter as a tool	5
a considerable amount of	5
support of h a	5
on social media to	5
kinds of altmetric data	5
peaks of moc and	5
as a news tweet	5
spread of the disease	5
a better understanding of	5
this is the first	5
to focus on the	5
the complexity of the	5
social media activities of	5
our goal is to	5
cybercrime and domestic abuse	5
the presence of different	5
in the real world	5
university of south alabama	5
association for the surgery	5
they were likely to	5
in social media and	5
and right wing parties	5
as an early warning	5
season marketing assets and	5
as well as for	5
it is likely that	5
number of population and	5
in the geocov tweets	5
and public health professional	5
despite the fact that	5
during times of crisis	5
the most popular hashtags	5
the spread of disinformation	5
social network analysis of	5
hot research topics are	5
total number of publications	5
and other social media	5
we counted the number	5
the context of a	5
and the peaks of	5
the identification of hot	5
the threshold is set	5
in social big data	5
of the united states	5
reporting a higher likelihood	5
would be more likely	5
the cwts classification system	5
word cloud on the	5
in the middle of	5
they had seen them	5
table shows the results	5
undirected version of the	5
with the highest number	5
the spread of information	5
the party in power	5
to be able to	5
shows the results of	5
the different hashtag groups	5
in online social networks	5
remains to be seen	5
more likely to engage	5
game attendance and local	5
th to may th	5
the most common symptoms	5
the most commonly used	5
subject fields and research	5
not certified by peer	5
the time of writing	5
of different altmetric data	5
more likely to tweet	5
in the label propagation	5
by the fact that	5
it remains to be	5
it is clear that	5
the average number of	5
event detection from twitter	5
rest of the paper	5
based on twitter data	5
in the digital age	5
of social media in	5
the university of south	5
the increase in the	5
the results of this	5
conference on advances in	5
if the tweet contains	5
use of twitter by	5
at the research topic	5
fields and research topics	5
important to note that	5
this study is to	5
social media such as	5
for this measure is	5
they are likely to	5
the research topic level	5
the twitter streaming api	5
paper is organized as	5
the sum of the	5
the same time period	5
at least one of	5
democrat and republican parties	5
known to be untrue	5
indicated that they had	5
social media and the	5
the titles of the	5
due to the covid	5
on twitter during the	5
the word cloud on	5
the intensity of the	5
be found in the	5
severe acute respiratory syndrome	5
notation description see appendix	5
the importance of the	5
that the tweet is	5
is worth mentioning that	5
response to the pandemic	5
the impact of covid	5
assets and team financials	5
there is a significant	5
in the supplementary information	5
the period of data	5
fast unfolding of communities	5
unfolding of communities in	5
description see appendix a	5
is defined as the	5
the middle of march	5
stay home save lives	5
the words that are	5
subject fields of science	5
is likely to be	5
at the time was	5
excluded from further analysis	5
natural language processing techniques	5
measles outbreak in the	5
social media platforms such	5
from the perspective of	5
analysis and opinion mining	5
later found out was	5
are shown in the	5
related to the pandemic	5
and higher levels of	5
counted the number of	5
the potential of twitter	5
that the stories were	5
the number of cases	5
the personal versus news	5
the relationship between in	5
korea missile test surge	5
social media data mining	5
found out was made	5
making it difficult to	5
is the total number	5
identified hot research topics	5
a personal negative tweet	5
identification of hot research	5
in a timely manner	5
a result of the	5
media platforms such as	5
one million population for	4
as well as in	4
pairs of altmetric data	4
sections of the questionnaire	4
is different from the	4
taking into account the	4
of this paper is	4
consistency of the items	4
global public health intelligence	4
the scope of this	4
verified and unverified users	4
news on social media	4
users in the directed	4
of altmetric data presence	4
based on the data	4
size was planned to	4
identified as hot research	4
all scales had acceptable	4
of negative versus non	4
evolution of the number	4
the core of the	4
had shared untrue material	4
context of the pandemic	4
more than of the	4
the nature of the	4
of positive and negative	4
seasonal ma and team	4
number of times a	4
step sentiment classification method	4
to refer to the	4
a strong floor effect	4
candidate at the university	4
the result of the	4
available under a author	4
uniform manifold approximation and	4
the variance in self	4
they believed they were	4
frames used in social	4
in this study we	4
they later found out	4
of moc and the	4
ma and game attendance	4
no conflict of interest	4
focus on boris johnson	4
one of the main	4
the authoritativeness of the	4
were the same as	4
we find that the	4
as it is the	4
are based on the	4
both sets of circumstances	4
of tweets related to	4
demographics are shown in	4
in the aftermath of	4
tracking social media discourse	4
we have the percentage	4
deaths per one million	4
of social media to	4
number of replies that	4
american society of clinical	4
authoritativeness of the story	4
indicated they had shared	4
the rate at which	4
the focus of this	4
harmful to the president	4
during these unprecedented times	4
media platforms can be	4
and support vector machine	4
web of science publications	4
societal impact of covid	4
number of mentions of	4
to social media platforms	4
number of increased cases	4
is associated with a	4
are summarized in table	4
decreased significantly from to	4
scales had acceptable reliability	4
as the pandemic intensifies	4
a brief summary of	4
asked about their historical	4
the information ecology framework	4
support of h c	4
progression of the pandemic	4
research topics with higher	4
all over the world	4
the aim of this	4
with the exception of	4
study was completed online	4
are provided in table	4
the frequency of the	4
while out of indicated	4
at a spatiotemporal scale	4
it is likely to	4
inquiry and word count	4
chatter dataset for open	4
it is interesting to	4
sets of circumstances were	4
of using social media	4
gender as either male	4
their historical sharing of	4
that contain abusive language	4
historical sharing of untrue	4
an analysis of tweets	4
levels of facebook use	4
sentiment analysis and opinion	4
cloud in fig shows	4
spread of true and	4
analysis to understand the	4
are summarised in table	4
of uk mps during	4
material under both sets	4
from the titles of	4
number of tweets from	4
analysis in this study	4
for sentiment analysis and	4
in regard to the	4
data sources are shown	4
analysis of the presence	4
of twitter followers on	4
more likely to discuss	4
counts for the top	4
center right and right	4
having seen them before	4
to establish whether the	4
did not report their	4
of the mueller investigation	4
conspiracy narratives on twitter	4
have been found to	4
in order to detect	4
collect a total of	4
number of replies received	4
the model explained of	4
from january to april	4
stages of the crisis	4
news stop word list	4
percentage of replies that	4
of personal negative tweets	4
of sharing the items	4
not people had shared	4
of those concerned with	4
for open scientific research	4
media activities of uk	4
under both sets of	4
and if they thought	4
existence of the virus	4
females were more likely	4
of the paper is	4
studies conducted in clinical	4
that they later found	4
more likely to propagate	4
that there is currently	4
most prominent news frames	4
coverage of the mueller	4
from the beginning of	4
that it does not	4
such as the number	4
are related to the	4
the tweets corresponding to	4
peak at date c	4
out of indicated they	4
of social media for	4
worth mentioning that the	4
a very skewed distribution	4
levels of coverage and	4
the authors found that	4
is a need to	4
same sets of participant	4
information associated with the	4
in this section we	4
power to detect r	4
the planned analysis was	4
ratings of likelihood of	4
a decision support system	4
substantive sections of the	4
around the coronavirus topic	4
we take care of	4
gender as a predictor	4
is possible that the	4
of policy document citations	4
new media literacy were	4
story online that they	4
from march to march	4
has the potential to	4
analysis and topic modeling	4
for notation description see	4
during the us presidential	4
is the percentage of	4
relationship between non seasonal	4
within the context of	4
internet and social media	4
the logarithm of the	4
media discourse about the	4
the week before the	4
quotas were used to	4
prior to any data	4
likely to tweet about	4
and behavior analysis methods	4
biomedical and health sciences	4
of twitter to track	4
or not people had	4
are included in the	4
in order to make	4
the pew research center	4
social media mining toolkit	4
the tweets collected between	4
thought they had seen	4
likelihood of sharing disinformation	4
the number of new	4
in online social media	4
a stronger relationship between	4
a semantic network analysis	4
each of the three	4
news detection on social	4
as a predictor variable	4
quantitatively and qualitatively understand	4
conducted in clinical settings	4
the same sets of	4
and at the same	4
most of the tweets	4
divided by the number	4
of indicated they had	4
we have developed a	4
black lives matter protests	4
untrue material under both	4
not report their gender	4
about their historical sharing	4
of this research is	4
shared material known to	4
the target sample size	4
in the first half	4
the virus is a	4
time of the tweet	4
of twitter followers reflect	4
were used in the	4
sources are shown in	4
was a significant predictor	4
with the same sets	4
tweets containing the common	4
is shown in figure	4
a linguistic analysis of	4
we begin with a	4
on the real network	4
having shared material known	4
abuse towards uk politicians	4
the tweet contains a	4
is represented by the	4
topics with higher levels	4
the pandemic and the	4
tool for health research	4
terms of demographic characteristics	4
the content of tweets	4
the tweets and the	4
health agency of canada	4
of whether or not	4
the panic buying group	4
which was not peer	4
a small number of	4
proportion of tweets that	4
across subject fields and	4
having unknowingly shared untrue	4
been asked about their	4
the properties of the	4
phrase i in t	4
in online abuse towards	4
public health intelligence network	4
shared one that they	4
a content analysis of	4
abuse toward uk mps	4
of untrue political stories	4
period of data collection	4
development of a public	4
of social media during	4
as either male or	4
in social media use	4
around the general election	4
predict the probability of	4
and the news timeline	4
shows the distribution of	4
given inclusion of gender	4
indicated that the model	4
the first half of	4
in the next section	4
public coronavirus twitter dataset	4
the output of the	4
a multiple regression analysis	4
the spread of true	4
this work was supported	4
that machine learning systems	4
linguistic inquiry and word	4
report their gender as	4
the data sources are	4
on advances in social	4
the centers for disease	4
number of tweets for	4
the authors would like	4
as having been posted	4
the context in which	4
of the daily number	4
twitter rest api to	4
increase the impact of	4
the main dependent variable	4
path between in season	4
items with participant attitudes	4
mediates the positive relationship	4
were examined using logistic	4
censorship of social media	4
the peaks of news	4
during the first week	4
were excluded from further	4
participants had also been	4
out prior to any	4
number of tweets as	4
social media data to	4
and the use of	4
qualitatively understand the chain	4
items if they believed	4
is much lower than	4
of the network and	4
and state of alarm	4
as opposed to the	4
see if the account	4
hate speech on twitter	4
at the publication level	4
circumstances were examined using	4
the activity history of	4
we can see that	4
as far as we	4
the influence of the	4
diffusion patterns of information	4
we make use of	4
are associated with a	4
authors would like to	4
model explained of the	4
to analyze the behavior	4
has been shown to	4
carried out prior to	4
five main subject fields	4
seasonal marketing assets and	4
a subset of the	4
the same as used	4
words in each hashtag	4
until may th inclusive	4
the same set of	4
for the top ten	4
was planned to exceed	4
it is unclear whether	4
of sharing the three	4
likely to be a	4
this was followed by	4
million tweets related to	4
of words in a	4
exclusions were carried out	4
this allows us to	4
abuse was found in	4
identify the action words	4
the items with participant	4
the size of our	4
and false news online	4
data twitter as a	4
total number of posts	4
in more detail below	4
seasonal marketing assets are	4
per one million population	4
to mps in the	4
shows the number of	4
machine learning algorithms in	4
in the volume of	4
significantly higher than the	4
of social and machine	4
the rise of the	4
using social media to	4
people reported a greater	4
the dynamics of hate	4
material known to be	4
in table and fig	4
used by public health	4
before the start of	4
the frames used by	4
on the impact of	4
each tw i in	4
this research is to	4
of having seen them	4
of the story source	4
the first years of	4
the perception of the	4
twitter data as social	4
the measles outbreak in	4
also been asked about	4
to see if the	4
positive relationship between non	4
rise of social bots	4
words related to death	4
twitter mentions and facebook	4
higher levels of facebook	4
of the virus and	4
hundred thousand population for	4
find words such as	4
will be able to	4
a large amount of	4
the peaks of the	4
be due to the	4
examined using logistic regressions	4
in the rest of	4
the stories were true	4
in terms of demographic	4
of the items with	4
the set of tweets	4
understand the impact of	4
the percentage of replies	4
very skewed distribution with	4
it is the case	4
target sample size was	4
of gender as a	4
detailed information of the	4
of personal protective equipment	4
multiple regression analysis was	4
and qualitatively understand the	4
institutional and news media	4
abuse of uk mps	4
research was done on	4
used to predict the	4
true and false news	4
personal negative tweets and	4
that a number of	4
data screening and processing	4
that our model outperforms	4
performance of the model	4
of tweets that were	4
untrue at the time	4
shortage of critical equipment	4
in the second step	4
the most frequent words	4
explained of the variance	4
in this case the	4
to predict the probability	4
sharing of untrue political	4
the natural language toolkit	4
who did not report	4
associated with the stories	4
use social media to	4
same as used in	4
news frames during the	4
in the name of	4
likely to be shared	4
the number of publications	4
and q a mentions	4
and exclusions were carried	4
be untrue at the	4
to be related to	4
of the main text	4
the hashtags mentioned in	4
in social media data	4
the number of posts	4
the mediation results of	4
in the first step	4
of the variance in	4
reflect a path of	4
with a strong floor	4
unknowingly shared untrue material	4
sharing the items if	4
numbers of twitter followers	4
had also been asked	4
what are the diffusion	4
lower levels of abuse	4
such as facebook and	4
link the virus to	4
social and mainstream media	4
high level of abuse	4
to account for the	4
common symptoms of covid	4
it is difficult to	4
planned to exceed n	4
of the tweets and	4
of the users who	4
in the directed validated	4
before and after the	4
a wide range of	4
twitter followers on the	4
society of clinical oncology	4
we find words such	4
consensus information associated with	4
shown in the appendix	4
media use in crisis	4
social media can be	4
had shared one that	4
people had shared untrue	4
we also observe that	4
these checks and exclusions	4
as described in section	4
state of alarm is	4
the actual case is	4
the proportion of tweets	4
to better understand the	4
the number of followers	4
false information on social	4
the detailed information of	4
if the number of	4
infected by the virus	4
that the model explained	4
and graph machine learning	4
the spread of covid	4
be seen in fig	4
to identify hot research	4
are the diffusion patterns	4
to quantitatively and qualitatively	4
the age of misinformation	4
the abuse of uk	4
a certain level of	4
of true and false	4
altmetric data show different	4
most prominent news frame	4
seasonal ma and game	4
th until may th	4
partially mediates the positive	4
american association for the	4
is defined as a	4
stress symptoms at a	4
the items if they	4
the role of social	4
rated likelihood of sharing	4
believed they were likely	4
their likelihood of sharing	4
tweets in our dataset	4
deaths per one hundred	4
whether or not people	4
information extracted from the	4
that there is a	4
fake news detection on	4
such as data mining	4
the majority tweet phq	4
manifold approximation and projection	4
activities of uk mps	4
their gender as either	4
if they thought they	4
the study was completed	4
we can find the	4
sharing the three disinformation	4
to detect depressed users	4
of sharing the stimuli	4
of the relationship between	4
theories that link the	4
the results show that	4
inclusion of gender as	4
the hashtag cloud in	4
were carried out prior	4
to the data collection	4
of altmetric data across	4
can be represented by	4
both unknowing and deliberate	4
people were more likely	4
is mainly composed by	4
dataset for open scientific	4
is available at https	4
discourse about the covid	4
as measured by the	4
platform for young adults	4
political polarization on twitter	4
being one of the	4
per one hundred thousand	4
online abuse towards uk	4
they had shared one	4
in the time of	4
have been applied to	4
online that they later	4
the path between in	4
around the world and	4
between the death nls	4
to predict the personality	4
not included in the	4
under a author funder	4
the first public coronavirus	4
the rise of social	4
first public coronavirus twitter	4
distribution with a strong	4
prominent news frames during	4
sample size was planned	4
the rest of this	4
it was shown that	4
field of social media	4
a comparison of the	4
skewed distribution with a	4
number of deaths in	4
in the first place	4
have been shown to	4
states during the early	4
in the supplementary material	4
for altmetric data with	4
if they believed they	4
use of twitter to	4
and found that the	4
to be untrue at	4
social media discourse about	4
of uk mps via	4
with respect to the	4
the one of the	4
on social media and	4
time of this writing	4
text of the tweets	4
mentions and facebook mentions	4
the source of the	4
it is assumed that	4
february th until may	4
rest of the models	4
spread of hate speech	4
participant demographics are shown	4
the contents of the	4
social media data sources	4
to any data analysis	4
either male or female	4
twitter chatter dataset for	4
tweets from the twitter	4
to use social media	4
promoting and refuting tweets	4
one order of magnitude	4
have the percentage of	4
the number of nodes	4
in which they are	4
uk mps via twitter	4
the global public health	4
high levels of abuse	4
of hate speech in	4
thought this was semi	4
a result of a	4
altmetric data at the	4
of social media by	4
shared untrue material under	4
as a proxy for	4
can be regarded as	4
to the lack of	4
regression analysis was carried	4
and a set of	4
as a personal tweet	4
of mlb brand equity	4
had a very skewed	4
we see that the	4
words such as family	4
the label propagation of	4
to assess how the	4
of the spread of	4
public health agency of	4
checks and exclusions were	4
social media has become	4
this paper presents a	4
a tool for health	4
on the social media	4
the increasing number of	4
topics of interest of	4
to social distancing and	4
levels of social media	4
be seen in figure	4
uk mps during covid	4
of circumstances were examined	4
to simultaneously test hypotheses	4
that link the virus	4
twitter has been used	4
they thought they had	4
gives an overview of	4
we observe that hashtags	4
in the literature that	4
of machine learning and	4
framework for social media	4
also be used to	4
we group the hashtags	4
predictors of whether or	4
the most weighted edges	4
one hundred thousand population	4
it is essential to	4
neither the authoritativeness of	4
in figure and figure	3
we develop a scalable	3
the coverage of five	3
the social media mining	3
predictor of likelihood of	3
to limit the spread	3
materials were the same	3
or urgency of endoscopic	3
research response during the	3
using the expanded predictor	3
during its early days	3
likelihood of sharing it	3
analysis was revised to	3
crisis and emergency risk	3
the details of the	3
the set of publications	3
opportunity to learn about	3
and public concern in	3
influenza epidemics using twitter	3
figure shows the distribution	3
of having seen the	3
goal of this study	3
twitter use at the	3
the potential to help	3
and conditions of the	3
has an r square	3
commonly encountered clinical scenarios	3
to compare the results	3
the authors propose a	3
have a degree of	3
spatial uncertainty of the	3
a higher level of	3
help elucidate these definitions	3
the results from the	3
for the sake of	3
preferences for research topics	3
whether there are similarities	3
that emerged from the	3
the majority of tweets	3
compare to the general	3
to test whether the	3
a label propagation a	3
is the extent to	3
to reassemble the machine	3
is to analyze the	3
linguistic features present in	3
social media platforms can	3
the use of frames	3
case of details noted	3
tweet is a personal	3
carried out using the	3
represents the number of	3
assess disease outbreaks from	3
to the public for	3
technically urgent or emergent	3
the different political parties	3
concerned with domestic abuse	3
when it comes to	3
to answer the following	3
mentions and policy document	3
were restricted to be	3
mechanisms for early detection	3
of the behavior of	3
and emergency risk communication	3
the social media activities	3
accounts related to the	3
in most of the	3
the following research questions	3
can be described as	3
the paper is organized	3
might be due to	3
appear to be the	3
for identifying hot research	3
to ensure that the	3
acm international conference on	3
pattern of diversion and	3
have been proposed for	3
a set of phrases	3
machine learning approach to	3
model is able to	3
in the second stage	3
impact of the pandemic	3
times higher than the	3
utilizing social media data	3
the resulting network featured	3
and information diffusion analysis	3
of the progression of	3
data using the twitter	3
in the present paper	3
to track levels of	3
f i g u	3
and the healthmap project	3
strongly subjective clues and	3
network featured nodes and	3
hot research topics with	3
are sizable and rich	3
the black lives matter	3
purpose of this research	3
surveillance using twitter data	3
study set out to	3
public health professional in	3
will report themselves as	3
attention on boris johnson	3
of sharing or liking	3
can play a crucial	3
logarithm of the daily	3
on a list of	3
were used to identify	3
our first research question	3
is useful for comparative	3
not technically urgent or	3
positive and statistically significant	3
social media during the	3
of observing a link	3
sense of an eid	3
the potential reach of	3
gastroenterologists agreed on procedure	3
the moderating effect of	3
data for wos publications	3
during an eid outbreak	3
according to the following	3
disease outbreaks and other	3
accounts in the mention	3
those infected by the	3
for the understanding of	3
the social media field	3
of smt misinformation and	3
context in which they	3
during our other efforts	3
the perception of risk	3
common clinical scenarios that	3
the right wing community	3
we have proposed a	3
one of the earliest	3
both static and dynamic	3
out to repeat study	3
the number of the	3
mendeley readers and twitter	3
are presented in table	3
feb th to may	3
the arabic content of	3
performance of our model	3
corpus for sentiment analysis	3
expanded set of predictors	3
we used deaths per	3
whether the user will	3
of march and then	3
the death nls and	3
reputable and non reputable	3
the different types of	3
social media during covid	3
users compare to the	3
social media is a	3
the literature that are	3
followers are associated with	3
of biomedical and health	3
middle of march and	3
the existence of the	3
the hypothesis that president	3
suite of tools aimed	3
strategies to mitigate its	3
based on the number	3
this section we will	3
of types of altmetric	3
survey of current work	3
to spread research with	3
the real network with	3
during the covid crisis	3
in terms of a	3
symptoms in social media	3
the best performing one	3
of the supplementary material	3
personal protective equipment for	3
and lower levels of	3
statistics of the data	3
to express their emotions	3
and others understand the	3
to identify shifting individuals	3
and because of the	3
each of the following	3
with a uk sample	3
analyze the behavior of	3
monitoring mechanisms for early	3
table presents the detailed	3
we analyzed more than	3
predictor set from study	3
work was supported by	3
is supported by the	3
diffusion patterns between sirsim	3
rapid research response during	3
effect of twitter followers	3
using social media data	3
sizable and rich enough	3
presence of different altmetric	3
of the h n	3
available to the public	3
fields of biomedical and	3
start of the pandemic	3
twitter catches the flu	3
the performance of our	3
gives counts for the	3
contour lines in each	3
is also a platform	3
used the number of	3
for most of the	3
misinformation on social media	3
the case of twitter	3
sentiments in the tweets	3
for predicting hate generation	3
the spread of hate	3
studies have shown that	3
disease surveillance using twitter	3
and its geo version	3
analysis of online social	3
the tweets were restricted	3
related to public healthcare	3
prominent news frame of	3
analysis and pattern recognition	3
information in order to	3
the rapid research response	3
at least symptom or	3
the concept of social	3
to the frequency of	3
and policy document citations	3
the daily distribution of	3
based on social sensors	3
to the nature of	3
the great east japan	3
number of tweets containing	3
data collection and behavior	3
used to ensure the	3
develop a scalable seeded	3
the united states the	3
can be very useful	3
urgent or emergent endoscopic	3
are shown in fig	3
topic modeling in twitter	3
if you want to	3
a contribution to the	3
of the pandemic and	3
three symptoms they experienced	3
presence of types of	3
twitter can play a	3
gastroenterologists regarding timing of	3
the information extracted from	3
users in the validated	3
gray contour lines in	3
in a number of	3
research topics on twitter	3
health professional in the	3
journal of cultural studies	3
by public health agency	3
so that they can	3
a louvain community detection	3
was presented on a	3
contribution to conspiracy narratives	3
the school closures group	3
during the influenza a	3
replicated that of study	3
summarizes the critical public	3
of alarm and covid	3
widely used by public	3
for twitter sentiment analysis	3
of writing this paper	3
how many of the	3
la raghavan et al	3
users on social media	3
table table table table	3
a call to action	3
the end of january	3
the distribution of infected	3
number of strongly subjective	3
is a measure of	3
related to the covid	3
the game on television	3
descriptive statistics are summarised	3
metrics are used to	3
shows the daily distribution	3
sentiment analysis of tweets	3
from january to march	3
usa states ranged from	3
number of times the	3
word can refer to	3
physical magnitude s of	3
linguistic analysis of the	3
were found to be	3
state of the pandemic	3
does not contain any	3
in a variety of	3
the sentiment of the	3
asked to rate their	3
of hate speech on	3
has also been found	3
the correlation between the	3
pattern recognition and trend	3
which can be used	3
from january to february	3
both of the two	3
cholera and kivu ebola	3
number of new cases	3
the two most important	3
by the end of	3
information and software technology	3
answer the following research	3
public response to the	3
sampling quotas were used	3
response during the covid	3
community structure in networks	3
in the previous subsection	3
symptoms at a spatiotemporal	3
are able to assign	3
an alternative to toilet	3
information network of twitter	3
measures and materials were	3
set of predictors eventually	3
expanded predictor set from	3
that the results presented	3
we introduce the social	3
application of altmetric data	3
during the last years	3
is consistent with our	3
more than million tweets	3
a discussion of the	3
positive relationship between in	3
understand the behavior of	3
were associated with higher	3
but is classified as	3
encountered clinical scenarios during	3
cloud on the right	3
the focus of attention	3
was revised to include	3
number of publications constant	3
clinical depressive symptoms in	3
used by researchers to	3
of the analysis results	3
far as we are	3
associated with a stronger	3
related to local tv	3
machine learning methods to	3
but not technically urgent	3
most common symptoms of	3
tweets during this time	3
asks about the diffusion	3
the age of covid	3
in our qualitative sample	3
recognition and trend analysis	3
infoveillance based on social	3
there is an assumption	3
in the english language	3
more likely to mention	3
season marketing assets are	3
save that in this	3
agencies such as the	3
same scores to substantive	3
with a list of	3
higher numbers of twitter	3
to detect the outbreak	3
impact of smt misinformation	3
that the total number	3
highlight the importance of	3
without the need for	3
moderately correlated with citations	3
techniques and graph machine	3
the undirected version of	3
local tv viewership is	3
presence across subject fields	3
the performance of retina	3
the power law model	3
there were only three	3
individuals will report themselves	3
to be the most	3
the hadoop distributed file	3
number of abusive replies	3
enough to support such	3
receiving a high level	3
eventually used in study	3
with the peaks of	3
at covid monitor austria	3
as a personal negative	3
has more than million	3
times to support the	3
of the most weighted	3
public health specialists and	3
of the three disinformation	3
number of times it	3
to survey gastroenterologists worldwide	3
the months before march	3
have been designed to	3
the last two weeks	3
focused the most on	3
with a high number	3
with the intention of	3
how twitter users compare	3
using commonly encountered clinical	3
to the control group	3
shows that there is	3
that it is not	3
the field of ssh	3
again had a very	3
the verified users of	3
with higher likelihood of	3
role in the rapid	3
the usage of the	3
to increase the impact	3
the methodology exactly replicated	3
the duration of the	3
want to be retweeted	3
compliance with government guidelines	3
colored based on the	3
of the three stimuli	3
of sharing the disinformation	3
the probability of the	3
to rate their likelihood	3
words based on their	3
potential to help policy	3
scale analytics on factors	3
based on the mapreduce	3
difference from study was	3
of verified and unverified	3
social media in the	3
as a set of	3
possible explanation is that	3
the creative commons attribution	3
graph machine learning algorithms	3
could be applied to	3
of the unified theory	3
the label for each	3
the fields of biomedical	3
and understanding of the	3
uncertainty of the analysis	3
in the lockdown group	3
influenza a h n	3
may be due to	3
fechner law and power	3
in more than of	3
increases the complexity of	3
to the levels of	3
as we are aware	3
the first three symptoms	3
denote the number of	3
respondents who did not	3
aspects of retweeting on	3
outbreak in the netherlands	3
of some of the	3
the ongoing development of	3
of retweeting on twitter	3
was based on the	3
has been applied to	3
framing of the h	3
this paper are based	3
social media to communicate	3
the news stop word	3
for most altmetric data	3
antagonistic responses to uk	3
strongest predictor of likelihood	3
of the tweets that	3
it is critical to	3
web and social media	3
the findings of the	3
is in charge of	3
most important sources of	3
emerging unified theory of	3
specific topics of interest	3
the population during the	3
with the notion of	3
the measure provides indices	3
the number of occurrences	3
classified as hot research	3
of team brand equity	3
the initial tweet had	3
opinion mining and sentiment	3
such as death and	3
twitter followers are associated	3
onset and spread of	3
s of the stimulus	3
to answer this question	3
the false discovery rate	3
with research interests in	3
altmetric data across subject	3
are less likely to	3
quarantine measures and effects	3
as used in study	3
to support the scientific	3
on all three platforms	3
provide a sample of	3
correlated with the peaks	3
detect public health events	3
a corpus for sentiment	3
be explained by the	3
financial conflicts of interest	3
twitter to track levels	3
was designed by the	3
because the same word	3
we developed a two	3
as well as those	3
the critical public responses	3
a data mining perspective	3
implementations of machine learning	3
with an average of	3
limit the spread of	3
consistent with the notion	3
the appearance of the	3
historical sharing in study	3
type of error is	3
of data was based	3
so that it can	3
if and only if	3
to deal with the	3
on social media as	3
great east japan earthquake	3
johns hopkins coronavirus resource	3
those that concerned covid	3
new media literacy scale	3
the number of words	3
tweets that contain the	3
is the date of	3
the number of unique	3
media tweet dataset for	3
the first step is	3
shows that for the	3
resulting network featured nodes	3
reported to the police	3
for each type of	3
the pattern of results	3
of a public coronavirus	3
events related to public	3
data used in this	3
been in contact with	3
on the mapreduce paradigm	3
we focus on the	3
to collect a total	3
followers on the relationship	3
the dark violet red	3
in a way that	3
and the g conspiracy	3
using search engine query	3
developer agreement and policy	3
label propagation a la	3
contemporary news articles published	3
in a total of	3
of disease activity and	3
networks analysis and mining	3
have clear guidelines regarding	3
data at the micro	3
in section of the	3
male or female were	3
the pandemic in different	3
as well as other	3
in the following section	3
was used in the	3
track levels of disease	3
of tweets in various	3
of tweets posted by	3
effect shows that higher	3
of this work is	3
aimed to survey gastroenterologists	3
the framing of the	3
influenza in the uae	3
were included in the	3
of influenza in the	3
case the nmls was	3
can be seen as	3
news frames on twitter	3
demonstration on flu and	3
different sources such as	3
the age of social	3
for early detection of	3
unprecedented times to support	3
a large quantity of	3
public coronavirus twitter data	3
an effect on the	3
multiple regression with predictors	3
on the topic of	3
data across subject fields	3
this shows that the	3
were used for the	3
represented by the number	3
of predictors eventually used	3
were added to the	3
the three time segments	3
with higher numbers of	3
in the cov tweets	3
is more than times	3
twitter during the covid	3
during the peak of	3
behaviour in social media	3
standardizing social media data	3
females were also more	3
the possibility of using	3
the notion of psychophysical	3
of a set of	3
related words in the	3
altmetric data for wos	3
related to team financial	3
be used as social	3
given the size of	3
the second most important	3
monitoring clinical depressive symptoms	3
that have been used	3
retweet in twitter network	3
using the mapreduce paradigm	3
italian embassies around the	3
the first and second	3
tweeted and retweeted by	3
results of this study	3
each tweet into a	3
the web page that	3
the uae based on	3
in the current context	3
media and social media	3
they were asked to	3
in the uae based	3
dataset and its geo	3
with a combination of	3
by a significant margin	3
on twitter and correlation	3
from study was in	3
can be extended to	3
path between non seasonal	3
in crisis and risk	3
the sample size conferred	3
to solve this problem	3
clinical scenarios that do	3
a significant proportion of	3
males were more likely	3
material in the past	3
not have clear guidelines	3
be difficult because the	3
for the united states	3
were not associated with	3
a sense of locality	3
revised to include the	3
the number of foreign	3
sharing or liking the	3
this study aims to	3
section of the supplementary	3
predicted ratings of likelihood	3
is positively related to	3
as discussed in sect	3
receiving a high percentage	3
between state of alarm	3
a copy of this	3
and financial reflect a	3
of the two newspapers	3
plan to flood america	3
a unique opportunity to	3
of twitter followers are	3
use of hashtags at	3
and its impact on	3
in the mention network	3
of altmetric data at	3
consistent with their pre	3
distribution of altmetric data	3
the results presented in	3
by taking into account	3
and new media literacy	3
the accounts in the	3
in the presentation of	3
of tweets per minute	3
hadoop distributed file system	3
next we want to	3
thousand population for countries	3
one of the largest	3
used to classify the	3
and content analysis of	3
scalable seeded topic modeling	3
the methods used by	3
the word cloud in	3
for public health specialists	3
across the different hashtag	3
south alabama with research	3
we present the most	3
of the tweets were	3
during the first two	3
level of attention on	3
spread of conspiracy narratives	3
for the analysis of	3
tweets in each category	3
figure we visualize the	3
gastroenterologists worldwide using twitter	3
whereas lower levels of	3
a higher degree of	3
of each of the	3
hundreds of thousands of	3
be used to control	3
social media messages in	3
leader of the opposition	3
topics in the field	3
extent to which the	3
scenarios during the covid	3
pause censorship of social	3
our results also show	3
world or meaning accretion	3
disease prevention and control	3
become part of the	3
conclusions can be drawn	3
analysis of the behavior	3
emerging infectious disease intelligence	3
altmetric data and citations	3
patterns of information cascades	3
can be difficult because	3
in the liwc dictionary	3
a high number of	3
disease outbreaks from tweets	3
is the case of	3
of overcrowding at hospitals	3
about the diffusion patterns	3
the network and the	3
alabama with research interests	3
has been used to	3
the second and third	3
to mitigate its impact	3
the impact of smt	3
results based on our	3
of the tweets to	3
our seeded lda model	3
the fields of ssh	3
the social consequences of	3
conduct a linguistic analysis	3
collected all the tweets	3
smt immunity link were	3
the cumbersome details of	3
based method using profanity	3
the number of their	3
ma and local tv	3
be used to detect	3
the types of tweets	3
would like to acknowledge	3
were asked to rate	3
reported at least symptom	3
certain communities such as	3
lack of availability of	3
misinformation epidemic on twitter	3
from this work have	3
levels of disease activity	3
a specific disease can	3
there appears to be	3
two independent ols models	3
circulation of conspiracy narratives	3
than the general population	3
positively related to team	3
the brand equity model	3
it may be that	3
in order to use	3
the development of a	3
for personal versus news	3
use of twitter in	3
or female were excluded	3
deleted by the user	3
twitter rather than facebook	3
expressed by twitter users	3
for countries ranged from	3
of data presence across	3
related to the coronavirus	3
twitter communications to understand	3
predict the retweet dynamics	3
some of the most	3
number of replies to	3
men were more likely	3
twitter users compare to	3
in a multiple regression	3
useful for comparative findings	3
measles during the measles	3
from a number of	3
moderating effect of twitter	3
that the amount of	3
to identify the action	3
question survey was designed	3
model is used to	3
from the end of	3
it is desirable to	3
and other events related	3
in figure we see	3
the former italian prime	3
that are associated with	3
to track the public	3
in comparison to other	3
the internet and social	3
terms and conditions of	3
it is known that	3
for usa states ranged	3
this work have the	3
team financial performance is	3
other social media platforms	3
the effectiveness of our	3
retweets and quote tweets	3
language use during covid	3
epub ahead of print	3
this is because tweets	3
role labeling approach to	3
authors declare that they	3
in the formation of	3
table gives counts for	3
this study is that	3
planned analysis was revised	3
the role of digital	3
to pause censorship of	3
detect depressed users on	3
twitter as a corpus	3
sentiment analysis to understand	3
makers and others understand	3
with a stronger relationship	3
an early warning system	3
likely to be retweeted	3
data from twitter and	3
using twitter to help	3
a function of the	3
due to the pandemic	3
to monitoring clinical depressive	3
of twitter followers will	3
pairs of words in	3
that they have no	3
such as how to	3
measure provides indices of	3
seen the stories before	3
the extension of the	3
guidelines regarding the timing	3
the authors declare no	3
about disease outbreaks and	3
sensitive but not technically	3
collection and behavior analysis	3
by thelwall et al	3
of tweets that are	3
the topic on twitter	3
the help of the	3
coronavirus twitter data set	3
the need for a	3
tweet dataset for covid	3
for a period of	3
notion of psychophysical numbing	3
access the social media	3
days of the pandemic	3
be defined as the	3
on social media is	3
the uk general election	3
tweets related to influenza	3
propagation a la raghavan	3
as facebook and twitter	3
approved the final manuscript	3
we propose a novel	3
limitations of this study	3
the accuracy of the	3
are consistent with the	3
replies that contain abusive	3
million tweets per day	3
and social media use	3
alarm and covid information	3
of respondents thought that	3
data for research purposes	3
topic detection and tracking	3
the physical magnitude s	3
for each of these	3
of the total number	3
similar conclusions can be	3
of the web pages	3
to help policy makers	3
very little disease related	3
the novel coronavirus outbreak	3
a percentage of replies	3
and total number of	3
for the purposes of	3
a wide variety of	3
one of the two	3
a multiple regression with	3
the number of twitter	3
learning and machine learning	3
due to the recent	3
a social media network	3
the first is that	3
in social networks analysis	3
prime minister matteo renzi	3
data mining and analysis	3
is an open access	3
to slow the spread	3
likely to engage with	3
the machine learning model	3
the expanded set of	3
of information cascades on	3
depicted graphically in fig	3
financial reflect a path	3
social media data from	3
the definition of the	3
to substantive sections of	3
ma and team financial	3
large scale analytics on	3
depression via social media	3
the fact that a	3
labeled as a news	3
on an exploratory basis	3
received higher levels of	3
of the new york	3
survey gastroenterologists worldwide using	3
altmetric data with relatively	3
as a corpus for	3
of the existing literature	3
semantic role labeling approach	3
there has been a	3
introduce the social media	3
to the general public	3
with a view to	3
supervised approach to monitoring	3
the impact of epidemics	3
in time window t	3
there is a strong	3
processing techniques and graph	3
we believe that the	3
due to the data	3
verbs in each group	3
of altmetric data sources	3
research on the use	3
symptoms and the number	3
of the unverified users	3
proceedings of the th	3
crucial role in the	3
we will focus on	3
impacting retweet in twitter	3
around the middle of	3
or emergent endoscopic interventions	3
public opinion and values	3
map to show the	3
data presence across research	3
that the model is	3
likely to have shared	3
show that our model	3
in order to reduce	3
search engine query data	3
the structure of the	3
female were excluded from	3
case of twitter in	3
age of social media	3
definitions using commonly encountered	3
the most popular social	3
this is not the	3
our other efforts on	3
the stages of the	3
is likely to have	3
which increases the complexity	3
results presented in this	3
the proposed corexq algorithm	3
it was possible to	3
correlations between citations and	3
the pandemic and its	3
correlation between sirsim and	3
a measure of the	3
of availability of essential	3
focal analysis in this	3
be defined as a	3
first three symptoms they	3
the focal analysis in	3
hybrid deep learning model	3
and materials were the	3
to the presence of	3
social media platforms to	3
in the twitter dataset	3
the social media data	3
specialists and government decision	3
the death and affect	3
of gastroenterologists agreed on	3
for the focal analysis	3
mps during the covid	3
of the covid crisis	3
support the need for	3
to engage with the	3
monitoring online media data	3
infectious disease intelligence and	3
tweets from march to	3
louvain community detection on	3
classify tweets related to	3
regarding a smt immunity	3
in the rapid research	3
main subject fields of	3
except in the case	3
the timing or urgency	3
the potential impact of	3
can be classified into	3
changes in the national	3
big data from social	3
table gives an overview	3
intelligence and the healthmap	3
we conduct a linguistic	3
twitter during this period	3
number of retweets received	3
for the task of	3
data were collected from	3
the state of alarm	3
applied to the raw	3
information about disease outbreaks	3
the gray contour lines	3
of followers an account	3
is an assumption that	3
and the politics of	3
lines in each panel	3
with the same scores	3
do not have clear	3
news frames of el	3
core of the mention	3
cannot be used to	3
work have the potential	3
approach to identify the	3
as part of the	3
number of cases in	3
in this paper are	3
we select the top	3
for major league baseball	3
advances in social networks	3
former italian prime minister	3
the distributions of the	3
this means that the	3
we collected the data	3
on the usage of	3
that were used to	3
a systematic literature review	3
from the twitter stream	3
regarding the timing or	3
these unprecedented times to	3
levels were associated with	3
as an alternative to	3
social networks have become	3
the statistics of the	3
a h n pandemic	3
from twitter and reddit	3
to include the expanded	3
news and fake news	3
according to the levels	3
the spread of conspiracy	3
is a personal tweet	3
identify information about disease	3
we observed that the	3
the terms and conditions	3
use in crisis and	3
through social media analysis	3
designed by the authors	3
for accuracy assessment of	3
is the concept of	3
scores to substantive sections	3
tweets and reply tweets	3
have been in contact	3
of retweets and favorites	3
since the beginning of	3
death nls and the	3
in each of the	3
is provided in figure	3
depressive symptoms in social	3
the personality of social	3
quote tweets and reply	3
in the last step	3
number of publications and	3
statistically significantly higher than	3
by the total number	3
machine learning framework for	3
created a database called	3
contain the spread of	3
personal versus news classifier	3
slow the spread of	3
least one of the	3
the new york city	3
the dimensionality of the	3
signs of fraudulent accounts	3
assess how the public	3
an r square of	3
cumbersome details of acquiring	3
related stress symptoms at	3
is shown in the	3
from across the world	3
the united states with	3
of diversion and suppression	3
in the present subsection	3
to track and predict	3
during the sampling period	3
of the mention network	3
each pair of data	3
data presence across subject	3
the pandemic gripped the	3
the american society of	3
analytics on factors impacting	3
social media data and	3
a la raghavan et	3
unique opportunity to learn	3
publications in the field	3
the contribution to conspiracy	3
devise strategies to mitigate	3
elucidate these definitions using	3
of this study are	3
complexity of the problem	3
during the onset of	3
magnitude s of the	3
study was in the	3
the number of deaths	3
and social media in	3
for disease surveillance and	3
access institutional and news	3
findings have implications for	3
methodology exactly replicated that	3
we focused on the	3
readers and twitter mentions	3
were only three of	3
of this was to	3
in relation to the	3
social networks analysis and	3
a random sample of	3
in the johns hopkins	3
historical sharing of false	3
the expanded predictor set	3
of all abuse to	3
the public response to	3
is related to a	3
in order to contain	3
take into account the	3
million population for usa	3
is the sum of	3
for hate speech detection	3
the findings of this	3
admitted to intensive care	3
probably due to the	3
the coverage of altmetric	3
of tweets posted in	3
on factors impacting retweet	3
of clinical oncology annual	3
lower levels of twitter	3
tools aimed to encapsulate	3
users to express their	3
able to detect the	3
community detection on the	3
by public health organizations	3
cannot be explained by	3
a subset of tweets	3
these definitions using commonly	3
related to big data	3
could be that the	3
altmetric data presence across	3
influenza epidemics using search	3
from a minority background	3
pointed to by the	3
in times of crises	3
data was based on	3
of altmetric data and	3
risk and crisis communication	3
worldwide using twitter to	3
of ways in which	3
to contain the spread	3
very large number of	3
most likely to be	3
that do not have	3
united states and the	3
the internet research agency	3
by world health organization	3
i g u r	3
others understand the impact	3
epidemic sentiment monitoring system	3
a percentage of all	3
clear guidelines regarding the	3
rich enough to support	3
it has also been	3
dynamics of information diffusion	3
is shown in fig	3
than of gastroenterologists agreed	3
from different sources such	3
the spatial distribution of	3
centrality measures influence by	3
the national daily death	3
mlb sales and marketing	3
a scalable seeded topic	3
the present subsection we	3
exactly replicated that of	3
of the th acm	3
other events related to	3
and devise strategies to	3
the g conspiracy theory	3
urgency of endoscopic evaluation	3
been used to study	3
report themselves as more	3
by the word cloud	3
percentage of all replies	3
likely to check the	3
natural language processing toolkit	3
during january and february	3
sharing of false political	3
of south alabama with	3
also known as the	3
the use of a	3
to encapsulate the cumbersome	3
all abuse to mps	3
outbreaks and other events	3
cascades on twitter about	3
featured nodes and edges	3
the results of sentiment	3
in the presence of	3
make use of the	3
at the start of	3
is the probability of	3
on flu and cancer	3
but the number of	3
used to analyze the	3
alternative to toilet paper	3
to the emergence of	3
words in the death	3
we aimed to survey	3
of users on twitter	3
life and earth sciences	3
the number of contagions	3
aimed to encapsulate the	3
the case of details	3
communities identified in the	3
declare that they have	3
we aim to investigate	3
and a number of	3