| 1 |
Explainable sentiment analysis application for social media crisis management in retail
|
|
|
|
In: Cirqueira, Douglas orcid:0000-0002-1283-0453 , Almeida, Fernando, Cakir, Gültekin orcid:0000-0001-9715-7167 , Jacob, Antonio orcid:0000-0002-9415-7265 , Lobato, Fabio orcid:0000-0002-6282-0368 , Bezbradica, Marija orcid:0000-0001-9366-5113 and Helfert, Markus orcid:0000-0001-6546-6408 (2020) Explainable sentiment analysis application for social media crisis management in retail. In: 4th International Conference on Computer-Human Interaction Research and Applications - Volume 1: WUDESHI-DR, 5-6 Nov 2020, Budapest, Hungry (Online). ISBN 978-989-758-480-0 (2020)
|
|
BASE
|
|
Show details
|
|
|
| 2 |
Using Twitter Streams for Opinion Mining: a case study on Airport Noise
|
|
|
|
In: ISSN: 1865-0929 ; Communications in Computer and Information Science ; https://hal.archives-ouvertes.fr/hal-03018998 ; Communications in Computer and Information Science, Springer Verlag, 2020, ⟨10.1007/978-3-030-44900-1_10⟩ (2020)
|
|
BASE
|
|
Show details
|
|
|
| 3 |
An Algerian Corpus and an Annotation Platform for Opinion and Emotion Analysis
|
|
|
|
In: Proceedings of the 12th Language Resources and Evaluation Conference ; 12th Language Resources and Evaluation Conference, LREC 2020 ; https://hal.archives-ouvertes.fr/hal-03102495 ; 12th Language Resources and Evaluation Conference, LREC 2020, May 2020, Marseille, France. pp.1202-1210 ; https://www.aclweb.org/anthology/2020.lrec-1.151/ (2020)
|
|
BASE
|
|
Show details
|
|
|
| 4 |
Affective behavior modeling on social networks ; Modélisation des sentiments sur les réseaux sociaux
|
|
|
|
In: https://tel.archives-ouvertes.fr/tel-03339755 ; Social and Information Networks [cs.SI]. Université Montpellier, 2020. English. ⟨NNT : 2020MONTS073⟩ (2020)
|
|
Abstract:
Affective Computing (AC) is an emerging area of research that aims to develop intelligent computer systems that can recognize, synthesize, and respond to the various concepts of human affect. With the vast increase of textual user-generated content on social media networks, the detection of human affect from text became an imperative need. Many tasks in Natural Language Processing (NLP) are directly related to affect recognition such as sentiment analysis, opinion mining, abusive language, at-risk user detection, and also those concerning human-computer interactions such as conversational frameworks and chatbots. Subjective and affective concepts in NLP research including feelings, intentions, emotions, moods, and sentiments are used interchangeably. However, bearing in mind the differences of these affect-related terms helps for more reliable and efficient detection systems. Many traditional systems and their modern extensions employ extensive feature engineering steps for text representation including hand-crafted, lexical features, or classical static word embedding. However these models may focus on the important parts of the input text, they disregard other parts and aspects which may harm model generalization for different affective states of the different affective concepts.In order to mitigate these limitations, we introduce different models that use/extend advanced NLP deep learning models for more reliable text representation. These models use the transfer learning capabilities and are empowered with the attention mechanisms to consider all the contextual information with varied emphasis on different parts with a higher influence on the decisions. Moreover, the proposed models accord special attention to the characteristic differences of the different affective concepts. We consider the affective characteristics of the most important conscious affect-driven subjectivity concepts, precisely, the sentiment, emotion, and mood:Sentiment: We addressed the problem of sentiment analysis and proposed a deep learning model that applies transfer learning and multi-levels of self-attention layers to focus on the most important parts of the text that have a high influence on sentiments. The model is evaluated on several datasets and shows very competitive results. Furthermore, we evaluate the impact of attention mechanisms on the model's interpretability and user perceptions.Emotion: We tackle the problem of detection and classification of basic emotions in textual dialogues. We extend the basic model used for sentiment classification to model textual conversations and track the emotion over turns. We participate in the SemEval-2019 shared task on contextual emotion detection in text. The model shows very competitive results and ranked 9th out of more than 150 participants.Mood-I: However user mood can be classified into two main types - positive and negative mood, mood disturbances inflict various mental illnesses/disorders. We consider the problem of early detection of depression, anorexia, and self-harm using users' writings on Reddit. We proposed a new multi-stage architecture that models users' temporal mood variations. We participated in eRisk-2018 and eRisk-2019 tasks. The proposed models perform comparably to other contributions and ranked the 2nd out of 13 teams in eRisk-2019.Mood-II: We foster the study of the mood consequences to include the problem of suicide thoughts detection. Therefore, we propose a novel backbone-independent model that uses state-of-the-art Transformer-based models through Negative Correlation Learning (NCL) configuration. We evaluate the model on different tasks for at-risk users detection. The models achieve significant improvements over the existing state-of-the-art results reported for five out of six tasks for the different risk sources. ; L’informatique affective (AC) est un domaine de recherche émergent qui vise à développer des systèmes informatiques intelligents capables de reconnaître, de synthétiser et de répondre aux différents concepts de l’affect humain. C’est l’une des pistes les plus importantes de la recherche sur l’intelligence artificielle (IA). L’AC est considérée comme le point de basculement permettant de passer de la définition cognitive étroite de l’IA à une IA plus générale, sentimentale et émotionnelle. Comme les émotions jouent un rôle essentiel dans les communications interhumains, les machines doivent également avoir la fluidité ou la souplesse nécessaires pour réagir aux situations en fonction des émotions. Les humains utilisent de multiples moyens pour communiquer leurs affects, notamment les expressions faciales, les gestes, le langage corporel, le ton de la voix, le langage et les indices verbaux. Avec l’augmentation considérable du contenu textuel généré par les utilisateurs sur les réseaux de médias sociaux, la détection de l’affect humain à partir du texte est devenue un besoin impératif. De nombreuses tâches du traitement du langage naturel (TALN) sont directement liées à la reconnaissance de l’affect, comme l’analyse des sentiments, l’exploration des opinions, la détection du langage abusif, la détection des utilisateurs à risque, et aussi celles concernant les interactions homme-machine, comme les cadres de conversation et les chatbots. Les concepts subjectifs et affectifs dans la recherche en TALN, y compris les sentiments, les intentions, les émotions, les humeurs et les émotions sont utilisés de manière interchangeable. Cependant, garder à l’esprit les différences de ces termes liés à l’affectivité permet d’obtenir des systèmes de détection plus fiables et plus efficaces. De nombreux systèmes traditionnels et leurs extensions modernes utilisent des étapes d’ingénierie de caractéristiques étendues pour la représentation de textes, y compris des caractéristiques lexicales artisanales ou l’intégration de mots statiques classiques. Cependant, ces modèles peuvent se concentrer sur les parties importantes du texte d’entrée, ils ignorent d’autres parties et aspects qui peuvent nuire à la généralisation du modèle pour différents états affectifs des différents concepts affectifs.Afin d’atténuer ces limitations, nous introduisons différents modèles qui utilisent ou étendent des modèles avancés d’apprentissage profonds utilisés en TALN pour une représentation plus fiable du texte. Ces modèles utilisent les capacités d’apprentissage par transfert et sont dotés de mécanismes d’attention permettant de prendre en compte toutes les informations contextuelles en mettant l’accent sur différentes parties ayant une plus grande influence sur les décisions. En outre, les modèles proposés accordent une attention particulière aux différences caractéristiques des différents concepts affectifs. Nous considérons les caractéristiques affectives des concepts les plus importants de la subjectivité affective consciente, précisément, le sentiment, l’émotion et l’humeur.
|
|
Keyword:
[INFO.INFO-SI]Computer Science [cs]/Social and Information Networks [cs.SI]; Deep learning; Early risk detection; Emotion detection; Nlp; Réseaux sociaux; Sentiment analysis; Social Network
|
|
URL: https://tel.archives-ouvertes.fr/tel-03339755
https://tel.archives-ouvertes.fr/tel-03339755/document
https://tel.archives-ouvertes.fr/tel-03339755/file/RAGHEB_2020_archivage.pdf
|
|
BASE
|
|
Hide details
|
|
|
| 5 |
An Enhanced Corpus for Arabic Newspapers Comments
|
|
|
|
In: ISSN: 1683-3198 ; International Arab Journal of Information Technology ; https://hal.archives-ouvertes.fr/hal-03124728 ; International Arab Journal of Information Technology, Colleges of Computing and Information Society (CCIS), 2020, 17 (5), pp.789-798. ⟨10.34028/iajit/17/5/12⟩ (2020)
|
|
BASE
|
|
Show details
|
|
|
| 6 |
Clickbait detection using multimodel fusion and transfer learning ; Détection de clickbait utilisant fusion multimodale et apprentissage par transfert
|
|
|
|
In: https://tel.archives-ouvertes.fr/tel-03139880 ; Social and Information Networks [cs.SI]. Institut Polytechnique de Paris, 2020. English. ⟨NNT : 2020IPPAS025⟩ (2020)
|
|
BASE
|
|
Show details
|
|
|
| 7 |
Neural approach for Arabic sentiment analysis ; Une approche neuronale pour l’analyse d’opinions en arabe
|
|
|
|
In: https://tel.archives-ouvertes.fr/tel-03084468 ; Informatique et langage [cs.CL]. Université du Maine; Université de Sfax (Tunisie), 2020. Français. ⟨NNT : 2020LEMA1022⟩ (2020)
|
|
BASE
|
|
Show details
|
|
|
| 8 |
ПРИМЕНЕНИЕ ТЕХНОЛОГИИ WORD2VEC В ЗАДАЧЕ ВЫДЕЛЕНИЯ ИНВЕРТОРОВ ТОНАЛЬНОСТИ ... : APPLYING WORD2VEC TECHNOLOGY TO SHIFTER EXTRACTION TASK ...
|
|
|
|
BASE
|
|
Show details
|
|
|
| 9 |
BanglaEmotion: A Benchmark Dataset for Bangla Textual Emotion Analysis ...
|
|
|
|
BASE
|
|
Show details
|
|
|
| 10 |
BanglaEmotion: A Benchmark Dataset for Bangla Textual Emotion Analysis ...
|
|
|
|
BASE
|
|
Show details
|
|
|
| 11 |
Hotel Review Sentiment Analysis using Natural Language Processing ...
|
|
|
|
BASE
|
|
Show details
|
|
|
| 12 |
Czech image captioning, machine translation, sentiment analysis and summarization (Neural Monkey models)
|
|
|
|
BASE
|
|
Show details
|
|
|
| 13 |
A sentiment analysis dataset for code-mixed Malayalam-English
|
|
|
|
BASE
|
|
Show details
|
|
|
| 14 |
How Combining Terrorism, Muslim, and Refugee Topics Drives Emotional Tone in Online News: A Six-Country Cross-Cultural Sentiment Analysis
|
|
|
|
In: International Journal of Communication; Vol 14 (2020); 26 ; 1932-8036 (2020)
|
|
BASE
|
|
Show details
|
|
|
| 15 |
A Sentiment Analysis Dataset for Code-Mixed Malayalam-English ...
|
|
|
|
BASE
|
|
Show details
|
|
|
| 16 |
Evaluation of literature by professional and layperson critics. A digital and literary sociological analysis of evaluative talk of literature through the prism of literary prizes (2007-2017) ...
|
|
|
|
BASE
|
|
Show details
|
|
|
| 17 |
A Sentiment Analysis Dataset for Code-Mixed Malayalam-English ...
|
|
|
|
BASE
|
|
Show details
|
|
|
| 18 |
Evaluation of literature by professional and layperson critics. A digital and literary sociological analysis of evaluative talk of literature through the prism of literary prizes (2007-2017) ...
|
|
|
|
BASE
|
|
Show details
|
|
|
| 19 |
Distant Spectators: Distant Reading for periodicals of the Enlightenment ...
|
|
|
|
BASE
|
|
Show details
|
|
|
| 20 |
Distant Spectators: Distant Reading for periodicals of the Enlightenment ...
|
|
|
|
BASE
|
|
Show details
|
|
|
|
|
