Aspect-based Sentiment and Correlation-based Emotion Detection on Tweets for Understanding Public Opinion of Covid-19

Background: During the Covid-19 period, the government made policies dealing with it. Policies issued by the government invited public opinion as a form of public reaction to these policies. The easiest way to find out the public's response is through Twitter's social media. However, Twitter data have limitations. There is a mix between facts and personal opinions. It is necessary to distinguish between these. Opinions expressed by the public can be both positive and negative, so correlation is needed to link opinions and their emotions. Objective: This study discusses sentiment and emotion detection to understand public opinion accurately. Sentiment and emotion are analyzed using Pearson correlation to determine the correlation. Methods: The datasets were about public opinion of Covid-19 retrieved from Twitter. The data were annotated into sentiment and emotion using Pearson correlation. After the annotation process, the data were preprocessed. Afterward, single model classification was carried out using machine learning methods (Support Vector Machine, Random Forest, Naïve Bayes) and deep learning method (Bidirectional Encoder Representation from Transformers). The classification process was focused on accuracy and F1-score evaluation. Results: There were three scenarios for determining sentiment and emotion, namely the factor of aspect-based and correlation- based, without those factors, and aspect-based sentiment only. The scenario using the two aforementioned factors obtained an accuracy value of 97%, while an accuracy of 96% was acquired without them. Conclusion: The use of aspect and correlation with Pearson correlation has helped better understand public opinion regarding sentiment and emotion more accurately © 2023 The Authors. Published by Universitas Airlangga.

Fatigue Management: Machine Learning Application for Predicting Mining Worker Fatigue

Mining workers can experience various kinds of physical and psychological impacts that will affect the emergence of fatigue. A lot of working hours and shift work mechanisms will drain many employees' energy. Research related to fatigue gives results that this affects employee performance, and even worse. This may have an impact on the emergence of an incident at work. Even greater impact will affect the company's business activities. Many mining companies have used fatigue monitoring mechanisms, but most of them provide results that are less fast and are less able to follow the pattern produced by each individual employee. This study shows the creation of a fatigue prediction model for employees using machine learning. Machine learning can identify potential whether employees are experiencing fatigue or not, so that it can assist management in making decisions. The collected data has 2 categories, namely fit and unfit. This research also uses the smote technique to balance the model so it doesn't lean towards one classes. Based on this study, it was found that the Random Forest algorithm was able to provide the best results which gives 95.4% accuracy compared to Decision Tree and Logistic Regression. According to the findings, it was found that there were variables that will have a major impact on the prediction results, namely sleep patterns and drug consumption since this data was taken during the Covid-19 pandemic. This research can also be used as a reference for establishing a model for determining fatigue both during Covid-19 and after Covid-19. © 2022 IEEE.

An Interactive Dashboard to Support Policymaking of Reopening School on Covid-19 Pandemic

In mid-April 2020, UNESCO monitored 191 countries and stated that around 1.723 trillion students in the world were affected by the policy of school from home. It is feared that school closures could hamper the provision of education services and could disrupt the education process which will affect the level of quality of education. There is still no creation of a computational model for the spread of covid as the main framework for schools reopening safely during the pandemic situation. Although there is already a framework from WHO and the government, there is no measuring tool that can evaluate the effect of reopening schools while the Covid-19 pandemic. For this reason, this research seeks to produce a model for the spread of Covid-19 as a basis for determining policies for safely reopening schools during the pandemic. In this research, we produced a recommendation to reopen face-to-face learning in the form of a dashboard. Recommendations are given by predicting the number of cases in each subdistrict using a predictive model. The prediction results are also combined with the factors that have been determined by the government to give recommendations. The allotment of recommendations process involves a critical factor analysis process where we identify which factors are dominant as a basis of a controllable pandemic. © 2022 IEEE.

Stance Detection on Tweets with Multi-task Aspect-based Sentiment: A Case Study of COVID-19 Vaccination

Public opinion analyses on Twitter conducted based on sentiment analysis cannot identify the author’s stance regarding agreement or disagreement with a given target. Stance detection determines whether the author of a text is in favor, against, or neutral towards a target. However, stance detection based on text-only is less representative opinion, especially on a tweet, which is a short text with slightly contextual information. Therefore, more information is needed to represent the author's stance better. In previous research, most research on stance detection was carried out using simple sentiment information to measure the support to target. This study addresses multi-task aspect-based sentiment analysis (ABSA) and social features for stance detection based on deep learning models of BiGRU-BERT on tweets. Our contribution combines aspect-based sentiment information with features based on textual and contextual information that does not emerge directly from Twitter texts. ABSA approach can provide more accurate sentiment information at aspect level on tweets, which is possible contains multiple issues discussed. Aspect information on tweets can reflect the issue that influences the author’s stance toward a target. Multi-task learning was applied to help improve the generalization performance of ABSA with simultaneous processes. We extracted social attributes and online behavioral features for contextual information. Since same community tends to have the same opinion towards a target, we applied a community detection task and combine with the Twitter social attributes. The proposed method has significantly improved evaluation metrics (>10%) than textual features only for stance detection on tweets © 2022. International Journal of Intelligent Engineering and Systems.All Rights Reserved

Fuzzy Unsupervised Approaches to Analyze Covid-19 Spread for School Reopening Decision Making

Virus SARS-Cov-2 causing Covid-19 spreads quickly and brings high risks to transmissions. The government to rule strictly to arrange strategies to minimize interactions through School-From-Home (SFH) policy. Unfortunately, the school closure is the potential to hamper deliveries of education services and may entail destructive impacts to quality education performance. There must be a consideration to school reopen safely during the pandemic. The objective of the research is to produce a model of Covid-19 spreads to analyze the readiness of school to reopen. This study adopts a SEIR model to predict the spread of Covid-19 using dataset from 23 March through 31 December 2020. The best model is selected from the one having the least error and adopted to predict the spread in the next 100 days starting from 01 January 2021 through 10 April 2021. Clustering was then implemented to acquire the character's proximity in each area using K-Means algorithm. While unsupervised fuzzy was picked out to seize the phenomenon of the dynamic as Covid-19 spread as a basis to decision making on school reopen safely during the pandemic. These whole concepts will serve the decision making effectively and intelligently by generating a better estimation. This study resulted in a Covid-19 spread model with an average error of 0.2% based on the RMSLE calculation.

Topic Detection in Sentiment Analysis of Twitter Texts for Understanding The COVID-19 Effect in Local Economic Activities

During the COVID-19 situation, discussions about the effect of COVID-19 increase on Twitter. Not only affecting the health sector, but the COVID-19 pandemic has also affected other fields, such as economic activities. Issues related to the economy become an essential discussion on Twitter because this sector has close links with other sectors in public activities. It makes twitter relevant as a knowledge extraction medium to identify users' opini comparisons. The contribution of this research is to find the effect of the COVID-19 pandemic on the comparison of sentiment and emotion in three different locations in Surabaya. Based on the results of emotion detection, at the beginning of the COVID-19 pandemic, topics related to economic activities and personal activities were dominated by anger emotion in the ITS campus and the TP mall area. Then, despite the gradual decrease in the intensity of tweets, the dominance of anger emotion tends to be stable. On economics topics, 40% of tweets in the ITS campus area and 84% of tweets in the TP mall area were dominated by anger emotion. Then 37% of tweets in the ITS campus area and 32% tweets in the Tunjungan Plaza mall area based on personal activities were dominated by anger. The economics topic is related to buying-selling and shopping activities, while personal activity is related to lifestyle and daily activities. These results indicate that during the COVID-19 pandemic, anger became the most dominant sentiment related to local economic activity from Twitter users in Surabaya.

Identifying Precautionary Measures for High-Risk Disease from Doctor's Answer Text Using LDA

The Online Health Consultation (OHC), which contains a QA collection of various diseases since 2014, has received an increasing number of visits due to the COVID-19. Based on the benefits and increasing health information need for people who seek information in OHC, health information related to precautionary measures to avoid diseases, especially high-risk diseases, become critical because not all seeker and readers of health information are diagnosed with certain diseases. However, It has currently unidentified whether the text of the doctor's answer corpus, especially in high-risk diseases, contains words that imply precautionary. This study aims to find the pattern of doctor's answer for high-risk diseases through the corpus of doctor's answer text on OHC by identifying whether the doctor's answer text contains words that imply precautionary against disease. Thus, it can help health information seekers and readers take precautionary against disease early on. This paper's contribution was to identify precautionary measures from doctor's answer text for high-risk disease in 2014-2021 using the best model of the two models, namely Single LDA (only LDA Method) and Hybrid LDA (a combination of LDA and Collapsed Gibbs Sampling). The results showed that the best model was Hybrid LDA, and medical experts identified groups of words with this model into four domains, namely symptoms/diagnosis, treatments, precautionary measurements, and general text. The pattern that emerges from the identification of precautionary measures shows (1) which precautionary measures are divided based on what disease, (2) Some words that mean precautionary measures also mean treatment or symptom/diagnosis. © 2021 IEEE.

Optimization Learning Approaches in Predicting Facebook Metrics from User Posts Behavior

The current situation of the Covid-19 pandemic has an impact on increasing the use of social media. In various aspects, social media has a role in human activities, especially in working-age groups. Breaking the stigma that social media interferes with someones' performance, we argue that using social media actually supports someones' work activities. In this preliminary study, we explore post behavior on Facebook social media networks for understanding user productivity. The dataset used in this study is gained from an online survey with the respondent of social media users over age 15 years old. Later on, based on surveys' responses, web scraping of Facebook post were set to complete the data needed. From the dataset, demographic features, metadata-based features, and behavior-based features are examined with some regression algorithms such Support Vector Regression (SVR) and Particle Swarm Optimization Extreme Learning Machine (PSO-ELM). The result from this study is only one feature that positively correlated to almost all other features during the pandemic. © 2021 IEEE.

Analysis the Opinion of School-from-Home during the COVID-19 Pandemic using LSTM Approach

The purpose of opinion analysis in this research is to perceive public responses concerning School-From-Home (SFH) policy during the pandemic in attempt to curb virus spread and worry about new cluster emergences. The policy entails diverse reactions from the societies, including the citizens in virtual world through their chirps in social media, such as Twitter. Analysis on the social media has proved that it has remarkable potentials to apprehend public opinions on various issues. The opinion analysis was performed to get insights about public perception towards SFH policy. As initially predicted, the result of our analysis would show that the public perceptions towards SFH would be mainly negative. The researcher adopted LSTM model as a deep learning approach. Moreover, implementing the N-Gram extraction technique was able to improve the model's performance. Model performance accuracy reached 83.30%. It is concluded that the increasing of model accuracy is about 0.018%. While the running time efficiency of LSTM has improved 19.4%. The results of the analysis of SFH's opinion were 77.90% negative and 22.10% positive. © 2021 IEEE.

Identifying degree-of-concern on covid-19 topics with text classification of twitters

The COVID-19 pandemic has various impacts on changing people’s behavior socially and individually. This study identifies the Degreeof-Concern topic of COVID-19 through citizen conversations on Twitter. It aims to help related parties make policies for developing appropriate emergency response strategies in dealing with changes in people’s behavior due to the pandemic. The object of research is 12,000 data from verified Twitter accounts in Surabaya. The varied nature of Twitter needs to be classified to address specific COVID-19 topics. The first stage of classification is to separate Twitter data into COVID-19 and non-COVID-19. The second stage is to classify the COVID-19 data into seven classes: warnings and suggestions, notification of information, donations, emotional support, seeking help, criticism, and hoaxes. Classification is carried out using a combination of word embedding (Word2Vec and fastText) and deep learning methods (CNN, RNN, and LSTM). The trial was carried out with three scenarios with different numbers of train data for each scenario. The classification results show the highest accuracy is 97.3% and 99.4% for the first and second stage classification obtained from the combination of fastText and LSTM. The results show that the classification of the COVID-19 topic can be used to identify Degreeof-Concern properly. The results of the Degree-of-Concern identification based on the classification can be used as a basis for related parties in making policies to formulate appropriate emergency response strategies in dealing with changes in public behavior due to a pandemic. © 2021, the author(s).

Time Series Analysis for Understanding Local Policy Impact of COVID-19 Cases in East Java

Daily new cases of COVID-19 as a series of data points ordered in time is one representation of time series data. Our works expect to understand the data characteristics that is related to existing government policies. This paper aims to highlight the impact report and analyses of some COVID-19 policies in East Java province districts. The study is focused on the policy execution before and during the new normal situation of COVID-19 using time series data of new cases as possible and easily observable results. Aside from time series analysis, some visual analysis is performed as well. The experiments focused on some questions related to the policy effectiveness: finding patterns of daily new cases to instigate the need for other local policies and understanding any precedence on occurred cases for nearby districts. Although the second question is not confirmable, the first question verifies more tightened social distancing is still necessary for the new normal. © 2020 IEEE.