Understanding Emojis for Sentiment Analysis

Yoo, Byung Kyu; Rayz, Julia Taylor

doi:10.32473/flairs.v34i1.128562

Cited by 9 publications

(5 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The positive category included love and joy while the negative was anger, surprise, fear, and sadness. The approach was observed to be different from previous studies that only focused on the accuracy of the emotion [13,15,27] or sentiment using positive and negative labels [28,29].…”

Section: Introductioncontrasting

confidence: 69%

Development and Comparison of Multiple Emotion Classification Models in Indonesia Text Using Machine Learning

Zamsuri,

Defit,

Nurcahyo

2024

JAIT

View full text Add to dashboard Cite

Emotion is an individual's response to an event or situation. Research related to emotions in the field of data science falls under sentiment analysis, where sentiment analysis mostly focuses on determining the positive or negative emotional tone. This study attempts to classify emotions into labels such as Happy, Love, Surprise, Anger, Fear, and Sadness. Then, these six emotion labels are further categorized into positive and negative groups. The dataset for this research comprises tweets from Twitter related to the 2024 presidential election in Indonesia. Several machine learning algorithms are employed in this study, including Naï ve Bayes (Multinomial Bayes, Bernoulli Bayes, Complement Bayes), K-Nearest Neighbors (KNN), and Support Vector Machines (SVM), comparing two feature extraction methods: Term Frequency-Inverse Document Frequency (TF-IDF) and Bag of Words (BoW). The resultsshow that the use of SVM does not yield better accuracy, whether using TF-IDF or BoW. Therefore, this study improves the accuracy of the SVM algorithm by combining kernels available in SVM (Polynomial, Radial Basis Function (RBF), and Linear). From the conducted experiments, it is evident that SVM with a combined kernel, referred to as SVM Polynomial, RBF, Linier (PoRLi) in this study, achieves better accuracy compared to a single kernel. This is reflected in the classification accuracy with 6 labels, reaching 62%, indicating a 2% increase from SVM Linear, which obtained the highest accuracy using a single kernel. Furthermore, in the classification of 2 labels, there is also a 1% increase when using SVM PoRLi. It can be concluded that SVM PoRLi can enhance accuracy across various labels.

show abstract

Section: Introductioncontrasting

confidence: 69%

Development and Comparison of Multiple Emotion Classification Models in Indonesia Text Using Machine Learning

Zamsuri,

Defit,

Nurcahyo

2024

JAIT

View full text Add to dashboard Cite

show abstract

“…The questionnaire includes both straight and reversely worded items in both dimensions. According to the literature, total scores are divided into 3 risk categories: low (16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35); moderate (36)(37)(38)(39)(40)(41)(42)(43); and high (44+). OLBI has been validated and used in low-and middle-income countries (LMICs) [31].…”

Section: Discussionmentioning

confidence: 99%

“…The goal of the conversations is to help users reflect on experiences and learn techniques to manage stress, mood, and anxiety. A mood tracking tool aids reflection, “emojis” mimic a natural human interaction, and gamification increases user engagement [ 24 ]. Vitalk users choose between different thematic conversations (“carelines”) and use standardized instruments to assess their mental health and resilience status before and after completing a careline at monthly or more frequent intervals.…”

Section: Methodsmentioning

confidence: 99%

“…Users of the Vitalk app had the option of assessing their mood using emojis representing 5 mood levels (0–4): very bad; bad; okay; good; and very good. Emojis are successfully used in sentiment analysis and show a stronger emotion compared to words [ 24 ]. The timing of mood assessments was matched to mental health assessments prior to data analysis since participants took them independently.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Effectiveness of a chatbot in improving the mental wellbeing of health workers in Malawi during the COVID-19 pandemic: A randomized, controlled trial

Kleinau,

Lamba,

Jaskiewicz

et al. 2024

PLoS ONE

View full text Add to dashboard Cite

We conducted a randomized, controlled trial (RCT) to investigate our hypothesis that the interactive chatbot, Vitalk, is more effective in improving mental wellbeing and resilience outcomes of health workers in Malawi than the passive use of Internet resources. For our 2-arm, 8-week, parallel RCT (ISRCTN Registry: trial ID ISRCTN16378480), we recruited participants from 8 professional cadres from public and private healthcare facilities. The treatment arm used Vitalk; the control arm received links to Internet resources. The research team was blinded to the assignment. Of 1,584 participants randomly assigned to the treatment and control arms, 215 participants in the treatment and 296 in the control group completed baseline and endline anxiety assessments. Six assessments provided outcome measures for: anxiety (GAD-7); depression (PHQ-9); burnout (OLBI); loneliness (ULCA); resilience (RS-14); and resilience-building activities. We analyzed effectiveness using mixed-effects linear models, effect size estimates, and reliable change in risk levels. Results support our hypothesis. Difference-in-differences estimators showed that Vitalk reduced: depression (-0.68 [95% CI -1.15 to -0.21]); anxiety (-0.44 [95% CI -0.88 to 0.01]); and burnout (-0.58 [95% CI -1.32 to 0.15]). Changes in resilience (1.47 [95% CI 0.05 to 2.88]) and resilience-building activities (1.22 [95% CI 0.56 to 1.87]) were significantly greater in the treatment group. Our RCT produced a medium effect size for the treatment and a small effect size for the control group. This is the first RCT of a mental health app for healthcare workers during the COVID-19 pandemic in Southern Africa combining multiple mental wellbeing outcomes and measuring resilience and resilience-building activities. A substantial number of participants could have benefited from mental health support (1 in 8 reported anxiety and depression; 3 in 4 suffered burnout; and 1 in 4 had low resilience). Such help is not readily available in Malawi. Vitalk has the potential to fill this gap.

show abstract

“…The modernday emojis can be traced back to chatrooms in the 1990s. They were used in conversations to signal a smile, anger or to portray a joke or sarcastic statement (kwan Yoo and Rayz, 2021). According to Emoji Statistics 1 , there were 3,633 emojis in total in the Unicode Standard as of September 2021.…”

Section: Introductionmentioning

confidence: 99%

Investigating Sentiment-Bearing Words- and Emoji-based Distant Supervision Approaches for Sentiment Analysis

Mabokela,

Roborife,

Celik

2023

Proceedings of the Fourth Workshop on Resources for African Indigenous Languages (RAIL 2023)

View full text Add to dashboard Cite

Sentiment analysis focuses on the automatic detection and classification of opinions expressed in texts. Emojis can be used to determine the sentiment polarities of the texts (i.e. positive, negative, or neutral). Several studies demonstrated how sentiment analysis is accurate when emojis are used (Kaity and Balakrishnan, 2020). While they have used emojis as features to improve the performance of sentiment analysis systems, in this paper we analyse the use of emojis to reduce the manual effort in labelling text for training those systems. Furthermore, we investigate the manual effort reduction in the sentiment labelling process with the help of sentiment-bearing words as well as the combination of sentiment-bearing words and emojis. In addition to English, we evaluated the approaches with the low-resource African languages Sepedi, Setswana, and Sesotho. The combination of emojis and words sentiment lexicon shows better performance compared to emojis-only lexicons and words-based lexicons. Our results show that our emoji sentiment lexicon approach is effective, with an accuracy of 75% more than other sentiment lexicon approaches, which have an average accuracy of 69.1%. Furthermore, our distant supervision method obtained an accuracy of 77.0%. We anticipate that only 23% of the tweets will need to be changed as a result of our annotation strategies.

show abstract

Understanding Emojis for Sentiment Analysis

Cited by 9 publications

References 7 publications

Development and Comparison of Multiple Emotion Classification Models in Indonesia Text Using Machine Learning

Development and Comparison of Multiple Emotion Classification Models in Indonesia Text Using Machine Learning

Effectiveness of a chatbot in improving the mental wellbeing of health workers in Malawi during the COVID-19 pandemic: A randomized, controlled trial

Investigating Sentiment-Bearing Words- and Emoji-based Distant Supervision Approaches for Sentiment Analysis

Contact Info

Product

Resources

About