RESUMO
Problem: Sentiment Analysis (SA) automates the classification of the sentiment of people's attitudes, feelings or reviews employing natural language processing (NLP) and computational approaches. Deep learning has recently demonstrated remarkable success in the field of SA in many languages including Arabic. Arabic sentiment analysis, however, still has to be improved, due to the complexity of the Arabic language's structure, the variety of dialects, and the lack of lexicons. Moreover, in Arabic, anxiety as a psychological sentiment has not been the target of much research. Aim: This paper aims to provide solutions to one of the challenges of Arabic Sentiment Analysis (ASA) using a deep learning model focused on predicting the anxiety level during COVID-19 in Saudi Arabia. Methods: A psychological scale to determine the level of anxiety was built and validated. It was then used to create the Arabic Psychological Lexicon (AraPh) containing 138 different dialectical Arabic words that express anxiety, which was used to annotate our corpus (Aranxiety). Aranxiety comprises 955 Arabic tweets representing the level of user anxiety during COVID-19. Bi-GRU model with word embedding was then applied to analyze the sentiment of the tweets and to determine the anxiety level. Results: For SA, the applied model achieved 88% on accuracy, 89% on precision, 88% on recall, and 87% for F1. A majority of 77% of tweets presented no anxiety, whereas 17% represented mild anxiety and a mere 6% represented high anxiety. Conclusion: The proposed model can be used by the Saudi Ministry of Health and members of the research community to formulate solutions to increase psychological resiliency among the Saudi population.
RESUMO
Personalized gamification aims to address shortcomings of the one-size-fits-all (OSFA) approach in improving students' motivations throughout the learning process. However, studies still focus on personalizing to a single user dimension, ignoring multiple individual and contextual factors that affect user motivation. Unlike prior research, we address this issue by exploring multidimensional personalization compared to OSFA based on a multi-institution sample. Thus, we conducted a controlled experiment in three institutions, comparing gamification designs (OSFA and Personalized to the learning task and users' gaming habits/preferences and demographics) in terms of 58 students' motivations to complete assessments for learning. Our results suggest no significant differences among OSFA and Personalized designs, despite suggesting user motivation depended on fewer user characteristics when using personalization. Additionally, exploratory analyses suggest personalization was positive for females and those holding a technical degree, but negative for those who prefer adventure games and those who prefer single-playing. Our contribution benefits designers, suggesting how personalization works; practitioners, demonstrating to whom the personalization strategy was more or less suitable; and researchers, providing future research directions. Supplementary Information: The online version contains supplementary material available at 10.1007/s40593-022-00326-x.
RESUMO
Massive Open Online Courses (MOOCs) have become universal learning resources, and the COVID-19 pandemic is rendering these platforms even more necessary. In this paper, we seek to improve Learner Profiling (LP), i.e. estimating the demographic characteristics of learners in MOOC platforms. We have focused on examining models which show promise elsewhere, but were never examined in the LP area (deep learning models) based on effective textual representations. As LP characteristics, we predict here the employment status of learners. We compare sequential and parallel ensemble deep learning architectures based on Convolutional Neural Networks and Recurrent Neural Networks, obtaining an average high accuracy of 96.3% for our best method. Next, we predict the gender of learners based on syntactic knowledge from the text. We compare different tree-structured Long-Short-Term Memory models (as state-of-the-art candidates) and provide our novel version of a Bi-directional composition function for existing architectures. In addition, we evaluate 18 different combinations of word-level encoding and sentence-level encoding functions. Based on these results, we show that our Bi-directional model outperforms all other models and the highest accuracy result among our models is the one based on the combination of FeedForward Neural Network and the Stack-augmented Parser-Interpreter Neural Network (82.60% prediction accuracy). We argue that our prediction models recommended for both demographics characteristics examined in this study can achieve high accuracy. This is additionally also the first time a sound methodological approach toward improving accuracy for learner demographics classification on MOOCs was proposed.
RESUMO
The COVID-19 pandemic has affected the whole world in various ways. One type of impact is that communication, work, interaction, a great part of our lives has moved online on various platforms, with some of the most popular being the social media ones. Another, arguably less visible impact, is the emotional impact. Detecting and understanding emotions is important, to better discern the emotional health and well-being of the global population. Thus, in this work, we use a social media platform (Twitter) to analyse emotions in detail. Our contribution is twofold: (1) we propose EmoBERT, a new emotion-based variant of the BERT transformer model, able to learn emotion representations and outperform the state-of-the-art; (2) we provide a fine-grained analysis of the pandemic's effect in a major location, London, comparing specific emotions (annoyed, anxious, empathetic, sad) before and during the epidemic.
Assuntos
COVID-19 , Mídias Sociais , COVID-19/epidemiologia , Surtos de Doenças , Emoções , Humanos , PandemiasRESUMO
Social media platforms changed from being socialization platforms to serve businesses through advertisements. This research aims at investigating active young users' experience with social media ads by studying the personalization and the usefulness of the ads, and the role of the host architecture of the used platform. The results prove that users' experience was affected by the designated variables: personalization, perceived usefulness, and the host architecture. Specifically, It was found that social media users find social media ads useful, and personalized, and that the perceived usefulness and personalization significantly affect the usage of host architecture which significantly affects users' experience. Additionally, a significant difference is found between clusters of student answers in terms of personalization and perceived usefulness effect on user experience.
RESUMO
Sentiment lexicons and word embeddings constitute well-established sources of information for sentiment analysis in online social media. Although their effectiveness has been demonstrated in state-of-the-art sentiment analysis and related tasks in the English language, such publicly available resources are much less developed and evaluated for the Greek language. In this paper, we tackle the problems arising when analyzing text in such an under-resourced language. We present and make publicly available a rich set of such resources, ranging from a manually annotated lexicon, to semi-supervised word embedding vectors and annotated datasets for different tasks. Our experiments using different algorithms and parameters on our resources show promising results over standard baselines; on average, we achieve a 24.9% relative improvement in F-score on the cross-domain sentiment analysis task when training the same algorithms with our resources, compared to training them on more traditional feature sources, such as n-grams. Importantly, while our resources were built with the primary focus on the cross-domain sentiment analysis task, they also show promising results in related tasks, such as emotion analysis and sarcasm detection.