A análise de sentimentos é um campo de estudo com recente popularização devido ao crescimento da Internet e do conteúdo que é gerado por seus usuários, principalmente nas redes sociais, nas quais as pessoas publicam suas opiniões em uma linguagem coloquial e em muitos casos utilizando de artifícios gráficos para tornar ainda mais sucintos seus diálogos. Esse cenário é observado no Twitter, uma ferramenta de comunicação que pode facilmente ser usada como fonte de informação para várias ferramentas automáticas de inferência de sentimentos. Esforços de pesquisas têm sido direcionados para tratar o problema de análise de sentimentos em redes sociais sob o ponto de vista de um problema de classificação, com pouco consenso sobre qual é o classificador com melhor poder preditivo, bem como qual é a configuração fornecida pela engenharia de atributos que melhor representa os textos. Outro problema é que em um cenário supervisionado, para a etapa de treinamento do modelo de classificação, é imprescindível se dispor de exemplos rotulados, uma tarefa árdua e que demanda esforço humano em grande parte das aplicações. Esta tese tem por objetivo investigar o uso de agregadores de classificadores (classifier ensembles), explorando a diversidade e a potencialidade de várias abordagens supervisionadas quando estas atuam em conjunto, além de um estudo detalhado da fase que antecede a escolha do classificador, a qual é conhecida como engenharia de atributos. Além destes aspectos, um estudo mostrando que o aprendizado não supervisionado pode fornecer restrições complementares úteis para melhorar a capacidade de generalização de classificadores de sentimento é realizado, fornecendo evidências de que ganhos já observados em outras áreas do conhecimento também podem ser obtidos no domínio em questão. A partir dos promissores resultados experimentais obtidos no cenário de aprendizado supervisionado, alavancados pelo uso de técnicas não supervisionadas, um algoritmo existente, denominado de C3E (Consensus between Classification and Clustering Ensembles) foi adaptado e estendido para o cenário semissupervisionado. Este algoritmo refina a classificação de sentimentos a partir de informações adicionais providas pelo agrupamento em um procedimento de autotreinamento (self-training). Tal abordagem apresenta resultados promissores e competitivos com abordagens que representam o estado da arte em outros domínios. / Sentiment analysis is a field of study that shows recent popularization due to the growth of Internet and the content that is generated by its users. More recently, social networks have emerged, where people post their opinions in colloquial and compact language. This is what happens in Twitter, a communication tool that can easily be used as a source of information for various automatic tools of sentiment inference. Research efforts have been directed to deal with the problem of sentiment analysis in social networks from the point of view of a classification problem, where there is no consensus about what is the best classifier, and what is the best configuration provided by the feature engineering process. Another problem is that in a supervised setting, for the training stage of the classification model, we need labeled examples, which are hard to get in the most of applications. The objective of this thesis is to investigate the use of classifier ensembles, exploring the diversity and the potential of various supervised approaches when these work together, as well as to provide a study about the phase that precedes the choice of the classifier, which is known as feature engineering. In addition to these aspects, a study showing that unsupervised learning techniques can provide useful and additional constraints to improve the ability of generalization of the classifiers is also carried out. Based on the promising results got in supervised learning settings, an existing algorithm called C3E (Consensus between Classification and Clustering Ensembles) was adapted and extended for the semi-supervised setting. This algorithm refines the sentiment classification from additional information provided by clusters of data, in a self-training procedure. This approach shows promising results when compared with state of the art algorithms.

This netnography depicts the notions of India’s subaltern voices entering the public via social media. The study puts an emphasis on feminists and caste critics, divided into two case studies. The study witnessed dynamics of Twitter use between sociality and activism as well as the notions of performance and identity of these two intersecting, yet polarised groups.Privilege remains a governing factor, which regulates access, accessibility and the use of the subaltern sphere and makes it exclusive for a privileged group of the subaltern. The main benefits of Twitter in the subaltern sphere, as the study suggests, is the factor of sociality and networking around causes, which leads to peer dialogue in the public sphere and increases visibility. This eventually leads to more attention for certain causes in the public discourse and to the countering of mainstream media narratives, for example in the case study of the Dalit Lives Matter Movement and its ad hoc fame, which evolved after the suicide of the Dalit PhD scholar Rohith Vemula.Further, while online activism is present, its impact remains hard to measure. The main benefits of the space are the plurality of voices that inhabit it. Also, the unleashing of the counter­narratives towards the mainstream media that are even more controlled by the state than the new media landscape, is an important benefit.

English profanities (i.e. potentially offensive words, including swear words) have been in use for decades in the Nordic languages – Icelandic, Norwegian, Danish, Swedish and Finnish – and offer a multitude of opportunities for linguistic expression, along with the domestic, heritage profanities in each language. The Nordic countries present an interesting context for studying the impact of English on languages in remote-contact settings, where many, especially young people, are bilingual but English has no official status. While previous studies have mostly focused on the function of such words and investigated their appearance in each Nordic language in isolation, this study utilizes social media data from the Nordic Tweet Stream (Laitinen et al., 2018) to compare the forms and frequencies of the English profanities fuck, shit, ass, damn, bitch and hell across the Nordic languages, shedding light on the factors which are conducive to their use. Surprisingly, the English profanities were many times more frequent in the Icelandic material compared to the other languages, although Iceland has a strong tradition of linguistic purism and frequencies were expected to be lower than in the other languages. Contrastingly, the profanities were found to be morphologically and orthographically adapted to a higher degree in Icelandic, reflecting the purist tradition in other ways. Frequencies in the other four languages did not quite match the findings of previous studies on loanwords in the Nordic languages, while degrees of adaptation were more similar to previous results. Comparing the frequencies of the English profanities in this study with the frequencies of heritage profanities on Twitter found by Coats (2021) showed that, although especially fuck and shit are on par with and sometimes more frequent than the most frequent heritage profanities, they do not seem to be replacing domestic equivalents. Finally, through exploiting the geo-location tags that accompany each tweet in the Nordic Tweet Stream, the frequencies of English profanities were found to be higher among users tweeting primarily from large cities in Denmark, Sweden and Finland, while in the Norwegian data no significant difference was found. Nevertheless, this supports Vaattovaara & Peterson’s (2019) claim that English borrowings carry social indices of globalism and urbanicity that promote their use among people in certain social groups. / Engelska fula ord (dvs. potentiellt stötande ord, inklusive svärord) har varit i bruk i årtionden i de nordiska språken – isländska, norska, danska, svenska och finska – och tillhandahåller en stor mängd språkliga uttrycksmöjligheter tillsammans med de inhemska fula orden i varje språk. De nordiska länderna utgör en intressant kontext för studier av engelskans inflytande över andra språk i distans-kontaktsituationer, där många, speciellt unga, är tvåspråkiga emedan engelska saknar officiell status. Då flertalet tidigare studier har fokuserat på denna typ av ords funktion, och undersökt deras förekomst i de nordiska språken var för sig, använder denna studie data från sociala medier, nämligen Nordic Tweet Stream (Laitinen et al., 2018) i en jämförelse av form och frekvens för de engelska fula orden fuck, shit, ass, damn, bitch och hell mellan de nordiska språken, för att därigenom synliggöra faktorer som gynnar deras bruk. Överraskande nog var de engelska fula orden långt mer frekventa i det isländska materialet jämfört med de övriga språken, trots att Island har en stark språkpuristisk tradition och frekvenserna därför förväntades vara lägre än i de andra språken. Däremot uppträdde de fula orden i högre grad i morfologiskt och ortografiskt anpassad form i isländska, vilket påvisar den puristiska traditionen på annat vis. Frekvenserna i de andra fyra språken skiljde sig något från resultaten av tidigare studier av lånord i de nordiska språken, medan anpassningsgraden var mer jämförbar med tidigare studier. En jämförelse av de engelska fula ordens frekvenser i denna studie med inhemska fula ords frekvenser på Twitter i en studie av Coats (2021) visade att medan fuck och shit mäter sig i frekvens med de vanligaste inhemska fula orden, och överträffar dem ibland, verkar de inte ersätta lokala motsvarigheter. Till sist påvisades, genom att utnyttja den geografiska platsdata som åtföljer varje tweet i Nordic Tweet Stream, att de engelska fula ordens frekvenser var högre bland användare som twittrar främst från stora städer i Danmark, Sverige och Finland, medan ingen signifikant skillnad kunde ses i den norska datan. Icke desto mindre styrker detta Vattovaara & Petersons (2019) tes gällande att lån från engelska bär med sig globala och urbana sociala indikationer som främjar deras bruk bland vissa sociala grupper.

Social media has become an important means of everyday communication. It is a mechanism for “sharing” and “resharing” of information. While social network platforms provide the means to users for resharing/reblogging (aka retweeting), it remains unclear what motivates users to share. Predicting the spread of content is quite important for several purposes such as viral marketing, popular news detection, personalized message recommendation and on-line advertisement. Social content systems store all the information produced in the interactions between users. However, to turn this data into information that allows us to extract patterns, it is important to consider the different phenomena involved in these interactions. In this work, two phenomena that influence the evolution of networks are studied for Twitter: diffusion of information and communication among users. Previous studies have shown that history of interaction among users and properties of the message are good attributes to understand the retweet behavior of users. Factors like content of message and time are less investigated. We propose a prediction model for retweet actions of users. It formulates a function which ranks the users according to how receptive they are to a particular message. The function generates a confidence score for the edges joining the initiator of the message and the followers. Two different pieces of information propagate through different users in the network. We divide the task of calculating confidence score into two parts. The first part is independent of the test tweet. It models transmission rate of the tie between the initiator and the follower. We call this as ‘Pairwise Influence Estimation’. The second part incorporates the tweet properties and user activeness as per time in the ranking function. The proposed model exploits all the dimensions of information dif-fusion process-influence, content and temporal properties. We have captured local aspects of diffusion. It has been observed that users do not read all the messages on their site. This results in shortcomings in the above models. Considering this, we first study the temporal behavior of users’ activities, which directly reflects their availability pertaining to the upcoming post. Also, as it is a continuous task of predicting retweet behavior, we design a user-centric, and temporally localized incremental classification model by considering the fact that users do not read all their tweets. We have tested the effectiveness of this model by using real data from Twitter. We demonstrate that the new proposed model is more accurate in describing the information propagation in microblog compared to the existing methods. Our model works well when we consider different classes of users depending on their activity patterns. In addition, we also investigate the parameters of the model for different classes of users. We report some interesting distinguishing patterns in retweeting behavior of users.

No / Marketers desire to utilise electronic word of mouth (eWOM) marketing on social media sites. However, not all online content generated by marketers has the same effect on consumers; some of them are effective while others are not. This paper aims to examine different characteristics of marketer-generated content (MGC) that of which one lead users to eWOM. Twitter was chosen as one of the leading social media sites and a content analysis approach was employed to identify the common characteristics of retweeted and favourited tweets. 2,780 tweets from six companies (Booking, Hostelworld, Hotels, Lastminute, Laterooms and Priceline) operating in the tourism sector are analysed. Results indicate that the posts which contain pictures, hyperlinks, product or service information, direct answers to customers and brand centrality are more likely to be retweeted and favourited by users. The findings present the main eWOM drivers for MGC in social media.