Volume 12 - Issue 63
/ March 2023
99
http:// www.amazoniainvestiga.info ISSN 2322- 6307
DOI: https://doi.org/10.34069/AI/2023.63.03.9
How to Cite:
Bondarchuk, N., Hrytsiv, N., Bekhta, I., & Melnychuk, O. (2023). Sentiment analysis of weather news in British online
newspapers. Amazonia Investiga, 12(63), 99-108. https://doi.org/10.34069/AI/2023.63.03.9
Sentiment analysis of weather news in British online newspapers
Сентимент-аналіз новин про погоду у британських електронних газетах
Received: January 22, 2023 Accepted: March 30, 2023
Written by:
Nataliya Bondarchuk1
https://orcid.org/0000-0002-5772-8532
Nataliia Hrytsiv2
https://orcid.org/0000-0001-6660-7161
Ivan Bekhta3
https://orcid.org/0000-0002-9848-1505
Oksana Melnychuk4
https://orcid.org/0000-0003-4619-363X
Abstract
The advancement of modern technologies has
influenced the way news is presented and
consumed, particularly online. Weather is an
important topic for the public as it relates to the
human experience and addresses current societal
issues. In this paper, we introduce a systematic
approach to conduct sentiment analysis of weather
news stories, to specify the emotional tone and
examine the role of subjectivity in online news
reporting. This research falls within the scope of a
lexicon-based (unsupervised) approach to
sentiment analysis, which involves finding the
sentiment polarity of words. The analysis is
predominantly based on sentence-level sentiment
analysis. Two popular online web services,
MonkeyLearn and SentiStrength, were applied to
automatically detect human emotions. We
compared the efficiency of each tool and found that
MonkeyLearn provided better final results in
comparison to SentiStrength, which tended to
misclassify negative sentiments into neutral ones.
The final results of frequency calculation showed
the dominance of weather news stories with
negative sentiment polarity over positive and
neutral ones, with neutral sentiments being in the
minority. Based on the empirical findings, we
observed an objectivity-to-subjectivity shift in
online news reporting.
1
Candidate of Philological Sciences (PhD), Associate Professor, Lviv Polytechnic National University, Institute of Computer
Sciences and Information Technologies, Department of Applied Linguistics, Ukraine.
2
Candidate of Philological Sciences (PhD), Doctoral Researcher, Lviv Polytechnic National University, Institute of Computer
Sciences and Information Technologies, Department of Applied Linguistics, Ukraine.
3
Doctor of Philology, Professor, Lviv Polytechnic National University, Institute of Computer Sciences and Information Technologies,
Department of Applied Linguistics, Ivan Franko National University of Lviv, Faculty of Foreign Languages, Department of English
Philology, Ukraine.
4
Candidate of Philological Sciences (PhD), Doctoral Researcher, Lviv Polytechnic National University, Institute of Computer
Sciences and Information Technologies, Department of Applied Linguistics, Ukraine.
100
www.amazoniainvestiga.info ISSN 2322- 6307
Keywords: text, emotion, natural language,
negative/positive sentiment, polarity, sentiment
analysis, subjectivity, weather news.
Introduction
Sentiment analysis, also known as opinion
mining, is one of the most common methods in
computational linguistics. The method is used to
identify and/or extract subjective information in
text data as well as research person’s attitude
towards the topic described. Sentiment analysis
tools determine whether data is positive, negative
or neutral, thus specifying the emotional
tone/load of the text. This study focuses on
sentiment analysis of weather news in British
online newspapers based on a lexicon-based
method.
The presentation of news has changed
considerably over the last decades in the light of
technological advances. The new electronic
environment determined the emergence of a new
electronic form (apart from written and oral
forms) of communication, where the information
is conveyed across time and borders (Bhatia et
al., 2022; Blake, 2019, Yates & Orlikowski,
1992) and is manifested in the combination of
interaction and communication.
Changes brought about by innovative
technologies have also transformed the
consumption and perception of the news content.
Greater accessibility of news for the users, its
increasing importance, and ubiquitous presence
in a new medium (electronic one) induced
linguistic research and predetermined the
material for our scrutiny. As such we have
chosen online British newspapers: two quality
papers (The Times, The Guardian) and two mass
ones (The Sun, The Daily Mail) as these four are
the most powerful newspapers in the UK and can
best demonstrate the use of a language in the
current time.
The increasing number of weather disasters over
the last years in most parts of the world cannot be
ignored. Frequent hurricanes, floods, droughts,
and long heatwaves causing bushfires, started the
list of news on extreme weather events over the
last years. The idea of weather extremes has
become widely synonymous with that of global
warming and anthropogenic climate change.
Climate controversy is becoming a key issue, as
well as a growing consensus for scientists,
political bodies, media, and general public.
Moreover, the weather topic has an
anthropocentric focus as everything happening in
the weather domain has a direct impact on
people. Abnormal weather events, weather
catastrophes and climate changes leading to the
destruction of the environment, mutilation, and
death of people immediately become hard news.
On the one hand, weather news is a piece of
writing the structure, topic, and language use of
which is strictly governed by the
requirements/conventions for news reporting. It
should be objective, include facts, and be concise
with stable clichés typical for newspaper style.
The information is to be communicated with a
standard set of lexis which makes the text
recognizable and identifiable as a member of a
specific genre and enables the reader to better
comprehend, process, and interact with the text.
On the other hand, it is a product/creation of the
journalist with his own stock and choice of
vocabulary, attitude towards the events/facts he
describes, opinions, a peculiar style of writing
(Bondarchuk & Bekhta, 2021), the use of
emotional appeal and/or persuasion techniques to
gain credibility of the reader (Al-Omari et al.,
2019; Bekhta & Hrytsiv, 2021; Wiebe & Riloff,
2005), which altogether affect the reception of
such news.
The lexical choices, selection of events and their
formulation in the news have a persuasive intent,
and display idiosyncratic features of writer’s
style and culturally-specific conventions
governing the behavior and attitude of people. In
this context, any news text can be approached
from three different perspectives ‒ the author,
reader and text itself. In this paper we aim to
identify whether weather news stories are
predominantly objective or subjective, and
whether subjective type includes more positive
or negative segments, thus focusing on text
perspective.
According to the rules and norms of news
writing, news stories should report factual news
Bondarchuk, N., Hrytsiv, N., Bekhta, I., Melnychuk, O. / Volume 12 - Issue 63: 99-108 / March, 2023
Volume 12 - Issue 63
/ March 2023
101
http:// www.amazoniainvestiga.info ISSN 2322- 6307
or events without clearly stating the opinion or
attitudes to facts, i.e. in an unbiased and
depersonalized manner. However, in the daily
news practice it can be observed that traditional
boundaries of journalism (especially in online
news environment) are crossed and the
separation of facts and personal opinion is
blurred. In addition, the journalists write from a
first person perspective mainly and perceive the
information through the prism of their
“subjective sensory-emotional consciousness”
(Zhou, 2021). This explains the rationale behind
the difficulties in differentiating the public
experience from the private (journalist’s) one.
Theoretical Framework
While the issue of binary opposition fact/opinion
in the news has been analysed in a number of
studies (Boesman & Costera, 2018; Höller, 2021;
Alhindi, Muresan, & Preotiuc-Pietro, 2020) the
application of sentiment analysis to news texts
needs commensurate attention.
Sentiment analysis is a field of Natural Language
Processing that focuses on discovering
techniques to decipher the sentiments hidden in
text comments from reviews or opinions posted
online (D’Aniello et al., 2022). Thus, the goal of
sentiment analysis is to identify, study or
categorize the emotions/opinions of people that
are expressed in a written text. There are two
generally accepted ways to approach sentiment
analysis: lexicon-based (unsupervised) and
machine learning (supervised) methods. The
lexicon-based method lies in the use of specific
lexicons containing the words which have been
tagged as being positive, negative, or neutral to
automatically detect the sentiment polarity and
further classify the sentiments. Machine-learning
approach deals with the automated calculation of
the sentiment scores within a particular text
based on trained data and test data.
Normally, sentiment analysis can be performed
on three different classification levels which are
presented in Figure 1. Sentence-level
classification is based on sentiment detection of
individual sentences, in turn, document-level
classification means to recognize the sentiment
of the whole document (Wiebe et al., 2002;
Mahmood et al., 2020). Aspect-level
classification is related to the detection of
sentiments on entities and their features.
Figure 1. Classification levels of sentiment analysis
Much work on news sentiment analysis has been
carried out using both lexicon-based (Hota at al.,
2021, Dharmale et al. 2023, Gupta & Urvashi.,
2023, Beigi & Moattar, 2020, Bonta et al., 2019,
Mutinda et al., 2020) and machine learning
approaches (Garay et al., 2019, Dagar et al.,
2021, Jurafsky & Martin, 2021, Mahmood et al.
2020, Amin et al. 2021, D’Aniello et al., 2022).
In recent research on news subjectivity (Wilson
et al., 2019) sentiment computation was
conducted using a Naive Bayes classifier (basic
classifier in machine learning). The focus of
another survey carried out by Chaturvedi et al.,
(2018) was subjectivity detection based on word
embeddings. The issue of climate change
sentiment analysis on social media platforms has
been analyzed by performing a comparative
evaluation of different sentiment analysis
techniques, namely lexicon-based, machine
learning, and hybrid approaches (Mohamad
Sham & Mohamed, 2022). This research showed
that hybrid approaches outperformed both
lexicon and machine learning approaches.
Methodology
Manual identification of either positive or
negative sentiments would be a very challenging
and time-consuming process. Our analysis of
subjectivity and opinion is automatic and utilizes
Sentiment analysis
Sentence level
Document level
Aspect level
102
www.amazoniainvestiga.info ISSN 2322- 6307
the lexicon-based (unsupervised) method, which
consists in finding sentiment polarity of the
words.
The automatic processing of texts to identify
positive, negative and neutral sentiments was
conducted using online web services
MonkeyLearn (https://monkeylearn.com/) and
SentiStrength (http://sentistrength.wlv.ac.uk/).
The results obtained by applying both the
SentiStrength algorithm and MonkeyLearn have
been compared to evaluate the performance of
each tool in this study. We hypothesize that
sentence-level sentiment analysis best fits the
aim of the survey, as it would be problematic to
detect subjectivity (subjective sentences) on a
document level since weather news stories can
include not only topics related to weather but also
other ones. Document-level classification may be
applied when analyzing, for instance, reviews or
comments, i.e. texts on one topic.
The dataset contains 125 news stories (which
should be enough for granting the
representativeness of the results) related to the
topic of weather collected on the dates between
2014 and 2018 from four official websites of
British quality and mass newspapers
(www.thetimes.co.uk,
www.theguardian.com/uk,
www.www.thesun.co.uk, www.thedailymail.co
.uk). The news stories have been queried by
using the keyword “weather”.
To compare the performances of MonkeyLearn
web service and SentiStrength program for
sentiment analysis, different pre-processing steps
have been used. For sentiment analysis, pre-
processing methods are of crucial importance
and their proper use increases the accuracy of the
results. Therefore, two datasets, which are
composed of 6000 sentences each, have been
formed.
We have registered an account in MonkeyLearn.
For MonkeyLearn dataset weather news stories
have been captured in a CSV file (Figure 2).
Figure 2. MonkeyLearn dataset in CSV format
Pre-Processing stage
Every weather news story underwent the
following pre-processing procedures to lessen
the noise of the text:
1) All words were changed into lower case
using python string lower() method.
Example: Why do cars slide into each other when
in Sweden you are required by law to change to
your snow tyres on national snow-tyre day.
Outcome: why do cars slide into each other when
in sweden you are required by law to change to
your snow tyres on national snow-tyre day.
2) All stop words were obtained from Natural
Language Tool Kit (hereinafter ‒NLTK) and
further removed to provide more accurate
results.
3) Lemmatization and stemming were
performed using NLTK.
4) All punctuation marks, numeric values and
unnecessary spaces have been removed.
Volume 12 - Issue 63
/ March 2023
103
http:// www.amazoniainvestiga.info ISSN 2322- 6307
Example: Ash rain – such a common sight now it
feels strange to think it needs an explanation – is
caused as the rain picks up the smoke, the filth,
the charred debris floating over cities.
Outcome: ash rain such a common sight now it
feels strange to think it needs an explanation is
caused as the rain picks up the smoke the filth the
charred debris floating over cities.
For SentiStrength dataset, there are 4 separate
input text files necessary for the algorithm to
work properly which are;
• Emoticon LookUp Table – contains a list of
emoticons with a strength 1 to 5 or -1 to -5
• Idiom Lookup Table - includes idiomatic
phrases and sentiment strengths.
• Negating Word List - a list of negation
words which reserve the polarity
• Booster Word List - a list of sentiment
intensity modifiers
In SentiStrength dataset we have split a text
document into sentences and made it into Excel
file. We have mainly been interested in capturing
the polarity of the sentences in which the sum of
polarities of individual words comprise the
polarity of a sentence.
Frequency analysis was used to calculate the
frequency of positive, negative and neutral
sentiments to identify a general tendency
characteristic of weather news stories. Some
failures of automatic detection of sentiments
have also been observed and described with
illustrative examples.
Results and Discussion
Having used Monkeylearn web service, the
sentences have been further classified into 3
types: positive, negative and neutral. Table 1
shows some of the examples of such
classification and Figure 2 exemplifies the code
used to perform such classification.
Table 1.
Classification results using MonkeyLearn
Sentiment score
Total
why can’t our trains cope with snow when Japanese trains can and by the way while we’re here
they are also much faster And this chorus of self-immolation is taken up countrywide: why non-Londoners ask, is the
capital brought to a standstill by a little snow? The weekend will start with a bang today as Scotland is engulfed by violent storms The Met Office has said winter had begun to bite after an unseasonably warm December, with
large parts of the UK facing snow, ice and frost In fact, our miserable, grey weather is all the fault of the jet stream
negative
positive positive
negative negative
So next time some pub bore tells you that this cold month is caused by the extensive melting of
arctic sea ice last summer ask him if the same thing happened in The grey mass appears to show the outline of mainland Great Britain as the sun begins to break
through the dark clouds on the horizon On the canal bridge just behind Kings Cross, a policeman took a huge snowball full in the face
and – I couldn't quite believe this was happening – giggled delightedly (it must have really
hurt) With a little luck, the freeze will come just in time to deliver a near universal white Christmas The sky is streaked with the world's highest type of cloud, visible only in summer after the sun
has set My God, I told myself as I walked through a heavenly avenue with snow-laden branches
bejewelling my steps, this is the most beautiful city in the world My soul was swooning (there, I admit it) yesterday as I stood and saw the snow falling, not on
Joyce's Ireland, but on dirty old London, reborn as a thing of beauty” With a little luck, the freeze will come just in time to deliver a near universal white Christmas On Tuesday afternoon the air seemed to grow colder, the sky turned dark, and then came a
surprise – big beefy snowflakes came tumbling down like large white butterfly In November the skies can be slate-grey and sullen, high and blue, or illumined by low, lemony
winter light.
It felt as though it was never going to end, but the „Beast from the East” will seem a distant
memory this week Residents braced themselves for further flooding as the wettest December since records began
continued to deluge the region
Warned to expect at least two more days without electricity following the weekend’s floods,
Lancaster’s hardy residents faced the darkness with sunny stoicism.
negative
neutral negative
positive negative positive
positive
positive
positive negative
negative negative
negative
104
www.amazoniainvestiga.info ISSN 2322- 6307
Figure 2. Example code used for classification
The program SentiStrength gives sentiment score
to the words with a range -5, -4, -3, -2, 2, 3, 4, 5
showing not only the polarity of words but
strength of the opinion. Table 2 presents some
results of such classification.
Table 2.
Classification results using SentiStrength
Sentiment score
Positive
Negative
Total
why can’t our trains cope with snow when Japanese [proper noun] trains
can and by the way while were here they are also much faster
1
-1
0
and this chorus of self-immolation is taken up countrywide why non-
londoners ask is the capital brought to a standstill by a little snow the weekend will start with a bang today as Scotland is engulfed by
violent storms the met office has said winter had begun to bite after an unseasonably
warm december, with large parts of the uk facing snow ice and frost In fact, our miserable grey weather is all the fault of the jet stream so next time some pub bore tells you that this cold month is caused by the
extensive melting of arctic sea ice last summer ask him if the same thing
happened in the grey mass appears to show the outline of mainland Great Britain as the
sun begins to break through the dark clouds on the horizon on the canal bridge just behind Kings Cross, a policeman took a huge
snowball full in the face and – I couldn't quite believe this was happening
– giggled delightedly (it must have really hurt). With a little luck, the freeze will come just in time to deliver a near
universal white Christmas The sky is streaked with the world's highest type of cloud, visible only in
summer after the sun has set My God, I told myself as I walked through a heavenly avenue with snow-
laden branches bejeweling my steps, this is the most beautiful city in the
world My soul was swooning (there, I admit it) yesterday as I stood and saw the
snow falling, not on Joyce's Ireland, but on dirty old London, reborn as a
thing of beauty” On Tuesday afternoon the air seemed to grow colder, the sky turned dark,
and then came a surprise – big beefy snowflakes came tumbling down like
large white butterfly In November the skies can be slate-grey and sullen, high and blue, or
illumined by low, lemony winter light
It felt as though it was never going to end, but the „Beast from the East”
will seem a distant memory this week Residents braced themselves for further flooding as the wettest December
since records began continued to deluge the region Warned to expect at least two more days without electricity following the
weekend’s floods, Lancaster’s hardy residents faced the darkness with
sunny stoicism
1
1
1
1
1
3
4
3 1
3
3
2
1
1 1
2
-1
-3
-1
-2
-2
-1
-4
-1 -1
-1
-3
-2
-2
-1 -2
-1
0
-2
0
-1
-1
2
0
2 0
2
0
0
-1
0 -1
1
Volume 12 - Issue 63
/ March 2023
105
http:// www.amazoniainvestiga.info ISSN 2322- 6307
We have also computed and annotated a list of
most common adjectives in terms of positive and negative polarity which is presented in
Table 3.
Table 3.
Llist of adjectives in terms of positive and negative polarity
Positive evaluation
Negative evaluation
cool, balmy, natural, decent, bright, calm, delightful, excellent, exceptional, fair, favourable, fine, glorious, good, great, ideal, light, lovely, mild, nice, clear, pleasant, promising, superb, perfect, clear, suitable, improved, gentle, fantastic, delightful, amazing)
extreme, severe, chilly, dirty, ugly, unfavourable, bad, failed, vicious, apocalypti unpleasant, adverse, shocking, awkward, odd, unseasonable, miserable, relentless, appalling, awful, freak,rotten, beastly, brutal, frightful, terrible, difficult, depressing, disgusting, disturbed, dull, foul, deteriorating, poor, gloomy, grey, grim, hard, harsh, fierce nasty, overcast, rough, filthy,
Our last step was to calculate the frequency of
positive, negative, and neutral sentiments in the
corpus under research. The results are presented
in Figure 3 where negative sentiments comprise
3276 (55 ⁒), positive ‒ 2345 (39⁒) and neutral ‒
379 (6 ⁒). We consider that neutral sentiments
denote objectivity in weather news reporting.
Figure 2. The results of sentiment frequency calculation
From the above analysis, it can be observed that
the majority of weather news stories are with
positive or negative sentiment polarity with
negative sentiments prevailing over the positive
and neutral ones, thus testifying to the general
tendency towards negativity and subjectivity of
weather news stories. The results of the
automated sentiment analysis have also been
compared with the manual linguistic analysis of
texts. Therefore, some limits of the study must be
emphasized. Remarkably, the most challenging
for both web services was to detect the sentiment
scores in the studied corpus when the opinion
was integrated on events and conveyed
implicitly.
An examination of the data revealed some
inaccuracies and failures in the automatic
detection of sentiments when metaphorical
expressions, irony, and sarcasm were employed.
For example, the following two examples with
irony created ambiguity and were incorrectly
2345; 39%
3276; 55%
379; 6%
Sentiment frequency
Positive Negative Neutral
106
www.amazoniainvestiga.info ISSN 2322- 6307
identified by SentiStrength (unlike
MonkeyLearn) as being positive (3 ‒ positive
sentiments, -1 ‒ negative; and 2 ‒ positive
sentiments, -1 ‒ negative respectively).
However, the second example was correctly
classified by MonkeyLearn as being negative.
“Why do schools close, when in Finland they use
the inclement weather to teach children about the
unending miracle of nature’s beauty – and that,
give or take a bit of screen time, is why their
children are so much happier than ours?”.
„GET those thermals out — the country will feel
like a fridge, if not the freezer, this month”.
Let us consider some other errors/inaccuracies:
• SentiStrength incorrectly classified the
sentence as being neutral (2 ‒ positive
sentiments, -2 ‒ negative):
“The heaviest snow in March for 50 years – felt
like a crushing iteration of the coalition
government, its endless austerity Narnia, always
winter, never Christmas, the feeling of sun on
your skin a distant memory from a better age,
like free tertiary education or a humane social
security system”.
• Failures in differentiating the polarity when
ironic juxtaposition of contexts or metaphor
was used. The following examples illustrate
such failures when sentences were
incorrectly typified as being: 1) positive
(MonkeyLearn) and neutral (SentiStrenth);
2) negative (MonkeyLearn) and positive
(SentiStrenth);
1) “And the cry is international too: as I walk
through the St Pancras Eurostar terminal, a
French couple consulting the warnings
about the tube, roll their eyes as one”
2) “In Britain, a flurry of the white stuff makes
everything pretty for about five minutes, but
then we’re smothered by a blanket of
national humility”
• Errors happened when allegedly positive
statements have a negative meaning, such as
in the following examples:
„We are overdue a genuine heatwave and you
know that when one late bus arrives it’s often
followed by two or three”
“On the canal bridge just behind Kings Cross, a
policeman took a huge snowball full in the face
and – I couldn't quite believe this was
happening – giggled delightedly (it must have
really hurt)”.
Conclusions
The technological innovations brought about
transformations in the structure and content of
newspapers. This article details the methodology
and results of sentiment analysis of weather news
stories carried out with the help of two web
services (MonkeyLearn and SentiStrenth) to
detect/classify positive, negative and neutral
sentiments, thus specifying emotional tone of the
texts and examine the role of subjectivity in
online news reporting. Such automatic detection
turned out to be quite efficient, apart from some
inaccuracies when the opinion was not clearly
marked (implicit) or figurative language was
used. It can also be concluded that SentiStrength
has the tendency to mistypify negative
sentiments into neutral. In this sense,
MonkeyLearn gave better classification results in
comparison with SentiStrenth. Moreover, we can
conclude that there is a general tendency towards
negativity the news, even though the frequency
of positive sentiments is quite high.
The application of computer software to the
analysis of opinion and subjectivity of weather
news stories falls within the scope of
computational linguistics and may contribute to
the understanding of the role of emotions in
newspaper discourse, furthermore, findings
obtained from this research outline the tendency
towards the objectivity to subjectivity shift in
news reporting. The method presented in this
research provides more options for further
content analysis of weather news stories in the
British press. Future scrutiny may also include
the comparative analysis of both lexicon-based
and machine learning approaches. We also intend
to investigate additional web services/programs
and their efficiency in detecting the sentiments of
weather news stories. Such a survey may shed
light on the issue of customizing the news by the
readers.
Bibliographic references
Alhindi, T., Muresan, S., & Preotiuc-Pietro, D.
(2020). Fact vs. Opinion: the Role of
Argumentation Features in News
Classification. In Proceedings of the 28th
International Conference on Computational
Linguistics, pages 6139–6149, Barcelona,
Spain (Online). International Committee on
Computational Linguistics.
Al-Omari, H., Abdullah, M., & Bassam, N.
(2019). EmoDet at SemEval-2019 Task 3:
Volume 12 - Issue 63
/ March 2023
107
http:// www.amazoniainvestiga.info ISSN 2322- 6307
Emotion Detection in Text using Deep
Learning. International Workshop on
Semantic Evaluation.
Amin, S., Uddin, M.I., Al-Baity, H.H.,
Zeb, M.A., & Khan, M.A. (2021). Machine
Learning Approach for COVID-19 Detection
on Twitter. Comput. Mater. Contin., 68,
2231–2247.
Beigi, O.M., & Moattar, M.H. (2020). Automatic
construction of domain-specific sentiment
lexicon for unsupervised domain adaptation
and sentiment classification. Knowl. Based
Syst., 213, 106423.
Bekhta, I., & Hrytsiv, N. (2021). Computational
Linguistics Tools in Mapping Emotional
Dislocation of Translated Fiction.
International Conference on Computational
Linguistics and Intelligent Systems.
Bhatia, T. K., Chauhan, K., & Suden, R. (2022).
“A Novel Technique to Detect the Fake News
by Using the Machine Learning
Approaches.” 2022 10th International
Conference on Reliability, Infocom
Technologies and Optimization (Trends and
Future Directions) (ICRITO), 1-6
Blake J. S. (2019). News in a Digital Age:
Comparing the Presentation of News
Information over Time and Across Media
Platforms. Rand Corporation.
Boesman, J., & Costera-Meijer, I.C. (2018).
“Don’t read me the news, tell me the story”:
How news makers and storytellers negotiate
journalism’s boundaries when preparing and
presenting news stories. pp. 13‒32.
Bondarchuk, N., & Bekhta, I. (2021).
Quantitative Characteristics of Lexical-
Semantic Groups Representing Weather in
Weather News Stories (Based on British
Online Press). International Conference on
Computational Linguistics and Intelligent
Systems
Bonta, V., Kumaresh, N., & Janardhan, N.
(2019). A Comprehensive Study on Lexicon
Based Approaches for Sentiment Analysis.
Asian Journal of Computer Science and
Technology, 8, pp. 1–6.
Chaturvedi, I., Cambria, E., Welsch, R.E., &
Herrera, F. (2018). Distinguishing between
facts and opinions for sentiment analysis:
Survey and challenges. Inf. Fusion, 44,
65-77.
https://doi.org/10.1016/j.inffus.2017.12.006.
Dagar, V., Verma, A., & Govardhan, K. (2021).
Sentiment analysis and sarcasm detection
(using emoticons). In: Applications of
artificial intelligence for smart technology.
IGI Global, pp 164–176
D’Aniello, G., Gaeta, M., & La Rocca, I. (2022).
KnowMIS-ABSA: An overview and a
reference model for applications of sentiment
analysis and aspect-based sentiment analysis.
Artif. Intell. Rev. 1–32.
Dharmale, G., Karjagi, S., Kotalwar, H., &
Khobragade, S. A. (2023) Detailed Survey on
Sentimental Analysis on Social Media. In:
Gunjan, V.K., Zurada, J.M. (eds)
Proceedings of 3rd International Conference
on Recent Trends in Machine Learning, IoT,
Smart Cities and Applications. Lecture Notes
in Networks and Systems, vol 540. Springer,
Singapore. https://doi.org/10.1007/978-981-
19-6088-8_22
Garay, J., Yap, R., & Sabellano, M. J. (2019). An
analysis on the insights of the anti-vaccine
movement from social media posts using k-
means clustering algorithm and VADER
sentiment analyser IOP Conference Series:
Materials Science and Engineering, 482(1),
012043
Gupta, S., & Urvashi, S. A. (2023). A
Vocabulary-Based Framework for Sentiment
Analysis. In: Shukla, A., Murthy, B.K.,
Hasteer, N., Van Belle, JP. (eds)
Computational Intelligence. Lecture Notes in
Electrical Engineering, vol 968. Springer,
Singapore. https://doi.org/10.1007/978-981-
19-7346-8_43
Höller, M. (2021). The human component in
social media and fake news: the performance
of UK opinion leaders on Twitter during the
Brexit campaign. European Journal of
English Studies, 25, 80 - 95.
Hota, H.S., Sharma, D.K., & Verma, N. (2021).
Lexicon-based sentiment analysis using
Twitter data. Data Science for COVID-19,
275-295. doi: 10.1016/B978-0-12-824536-
1.00015-0
Jurafsky, D., & H Martin, J. (2021). Naive Bayes
and sentiment classification. In: Speech and
language processing. Stanford University.
Mahmood, A., Kamaruddin, S., Naser, R., &
Nadzir, M. (2020). A combination of lexicon
and machine learning approaches for
sentiment analysis on Facebook. J. Syst.
Manag. Sci., 10, 140–150.
Mohamad Sham, N., & Mohamed, A.H. (2022).
Climate Change Sentiment Analysis Using
Lexicon, Machine Learning and Hybrid
Approaches. Sustainability.
https://doi.org/10.3390/su14084723
Mutinda, J., Mwangi, W., & Okeyo, G.O. (2020).
Lexicon‐pointed hybrid N‐gram Features
Extraction Model (LeNFEM) for sentence
level sentiment analysis. Engineering
Reports, 3.
Wiebe, J., & Riloff, E. (2005). Creating
Subjective and Objective Sentence
Classifiers from Unannotated Texts.
108
www.amazoniainvestiga.info ISSN 2322- 6307
Conference on Intelligent Text Processing
and Computational Linguistics.
Wilson, T., Hoffmann, P., Somasundaran, S.,
Kessler, J., Wiebe, J., Choi, Y., Cardie, C.,
Riloff, E., & Patwardhan, S. (2019). "Opinion
Finder: A system for subjectivity analysis". In
Proceedings of hlt/emnlp on interactive
demonstrations, pp. 34-35.
Yates, J., & Orlikowski, W. J. (1992). Genres of
organizational communication: A
structurational approach. Academy of
Management Review, 17, pp. 299–326.
Zhou, Z. (2021). Emotional thinking as the
foundation of consciousness in artificial
intelligence. Cultures of Science, 4, 112-123.
https://doi.org/10.1177/20966083211052651