Intersubjectivity and sentiment:from language to knowledge

Intersubjectivity is an important concept in psychology and sociology. It refers to sharing conceptualizations through social interactions in a community and using such shared conceptualization as a resource to interpret things that happen in everyday life. In this work, we make use of intersubjectivity as the basis to model shared stance and subjectivity for sentiment analysis. We construct an intersubjectivity network which links review writers, terms they used, as well as the polarities of the terms. Based on this network model, we propose a method to learn writer embeddings which are subsequently incorporated into a convolutional neural network for sentiment analysis. Evaluations on the IMDB, Yelp 2013 and Yelp 2014 datasets show that the proposed approach has achieved the state-of-the-art performance.

Social-media monitoring for cold-start recommendations

Generating personalized movie recommendations to users is a problem that most commonly relies on user-movie ratings. These ratings are generally used either to understand the user preferences or to recommend movies that users with similar rating patterns have rated highly. However, movie recommenders are often subject to the Cold-Start problem: new movies have not been rated by anyone, so, they will not be recommended to anyone; likewise, the preferences of new users who have not rated any movie cannot be learned. In parallel, Social-Media platforms, such as Twitter, collect great amounts of user feedback on movies, as these are very popular nowadays. This thesis proposes to explore feedback shared on Twitter to predict the popularity of new movies and show how it can be used to tackle the Cold-Start problem. It also proposes, at a finer grain, to explore the reputation of directors and actors on IMDb to tackle the Cold-Start problem. To assess these aspects, a Reputation-enhanced Recommendation Algorithm is implemented and evaluated on a crawled IMDb dataset with previous user ratings of old movies,together with Twitter data crawled from January 2014 to March 2014, to recommend 60 movies affected by the Cold-Start problem. Twitter revealed to be a strong reputation predictor, and the Reputation-enhanced Recommendation Algorithm improved over several baseline methods. Additionally, the algorithm also proved to be useful when recommending movies in an extreme Cold-Start scenario, where both new movies and users are affected by the Cold-Start problem.

Alpha generation with Twitter’s daily momentum

This thesis does not set out to focus on the dynamics relationship between Twitter and stock prices, but instead tries to understand if using relevant information extracted from tweets has the power to increase investors’ stock picking ability, and generate alpha in portfolio’s choice relative to a benchmark. Despite the short period analyzed, it gives promising results that the sentiment analysis performed by Social Market Analytics Inc. applied to an equity portfolio, is able to generate positive abnormal returns, statistically significant in and out of sample.

Text mining e twitter : o poder das redes sociais num mercado competitivo

Actualmente, com a massificação da utilização das redes sociais, as empresas passam a sua mensagem nos seus canais de comunicação, mas os consumidores dão a sua opinião sobre ela. Argumentam, opinam, criticam (Nardi, Schiano, Gumbrecht, & Swartz, 2004). Positiva ou negativamente. Neste contexto o Text Mining surge como uma abordagem interessante para a resposta à necessidade de obter conhecimento a partir dos dados existentes. Neste trabalho utilizámos um algoritmo de Clustering hierárquico com o objectivo de descobrir temas distintos num conjunto de tweets obtidos ao longo de um determinado período de tempo para as empresas Burger King e McDonald’s. Com o intuito de compreender o sentimento associado a estes temas foi feita uma análise de sentimentos a cada tema encontrado, utilizando um algoritmo Bag-of-Words. Concluiu-se que o algoritmo de Clustering foi capaz de encontrar temas através do tweets obtidos, essencialmente ligados a produtos e serviços comercializados pelas empresas. O algoritmo de Sentiment Analysis atribuiu um sentimento a esses temas, permitindo compreender de entre os produtos/serviços identificados quais os que obtiveram uma polaridade positiva ou negativa, e deste modo sinalizar potencias situações problemáticas na estratégia das empresas, e situações positivas passíveis de identificação de decisões operacionais bem-sucedidas.

Text mining na análise de sentimentos em contextos de big data

Dissertação de mestrado integrado em Engenharia e Gestão de Sistemas de Informação

A Classification of Adjectives for Polarity Lexicons Enhancement

Subjective language detection is one of the most important challenges in Sentiment Analysis. Because of the weight and frequency in opinionated texts, adjectives are considered a key piece in the opinion extraction process. These subjective units are more and more frequently collected in polarity lexicons in which they appear annotated with their prior polarity. However, at the moment, any polarity lexicon takes into account prior polarity variations across domains. This paper proves that a majority of adjectives change their prior polarity value depending on the domain. We propose a distinction between domain dependent and romain independent adjectives. Moreover, our analysis led us to propose a further classification related to subjectivity degree: constant, mixed and highly subjective adjectives. Following this classification, polarity values will be a better support for Sentiment Analysis.

Automatic Extraction of Polar Adjectives for the Creation of Polarity Lexicons

Automatic creation of polarity lexicons is a crucial issue to be solved in order to reduce time andefforts in the first steps of Sentiment Analysis. In this paper we present a methodology based onlinguistic cues that allows us to automatically discover, extract and label subjective adjectivesthat should be collected in a domain-based polarity lexicon. For this purpose, we designed abootstrapping algorithm that, from a small set of seed polar adjectives, is capable to iterativelyidentify, extract and annotate positive and negative adjectives. Additionally, the methodautomatically creates lists of highly subjective elements that change their prior polarity evenwithin the same domain. The algorithm proposed reached a precision of 97.5% for positiveadjectives and 71.4% for negative ones in the semantic orientation identification task.

THE ROLE OF CULTURAL AWARENESS IN BUILDING CHINESE HEALTH FOOD BRANDS Using Netnography in Searching Cultural Arguments to Be Used for Marketing Health Foods in China

Despite the complexity of the Chinese culture consumer product businesses should apply them in building brands for Chinese markets. There are reasons to believe that cultural values and attitudes affect the buying behavior of Chinese consumers. Companies that wish to create brands in China should therefore be aware of the prevailing cultural values and consumer attitudes. This thesis will examine which values and attitudes mostly affect Chinese consumers of health food products. The study will be done by conducting a netnography. Because netnography is actually a collection methods rather than a single method, other auxiliary methods will also be applied. These methods are emotion, language and sentiment analysis (ELS analysis). Emotion analysis will be conducted because cultural values are mostly built on emotional basis. Sentiment analysis will assist in recognizing the key factors that help to locate values and attitudes. Because the netnography will be conducted in Chinese web forums by a non-native researcher, linguistic aspects should also be analyzed in parallel with emotions and sentiment analysis. The study shows that the Chinese consumers of health food products put much importance on functional, analytical and collectivistic attitudes as well as social and psychological values. Of all the twelve cultural values defined, the role of family rose above all. Also perseverance, frugality, guanxi and harmony were highly presented. The attitudes were found by recognizing certain attitude factors. Of all the factors, health foods’ functional benefits and aesthetic content together with consumers’ value consciousness surpassed other factors. Besides these results that can be applied by foreign health food companies willing to enter Chinese consumer markets, also academia can benefit this new approach for conducting ethnographies online.

Apprentissage des réseaux de neurones profonds et applications en traitement automatique de la langue naturelle

En apprentissage automatique, domaine qui consiste à utiliser des données pour apprendre une solution aux problèmes que nous voulons confier à la machine, le modèle des Réseaux de Neurones Artificiels (ANN) est un outil précieux. Il a été inventé voilà maintenant près de soixante ans, et pourtant, il est encore de nos jours le sujet d'une recherche active. Récemment, avec l'apprentissage profond, il a en effet permis d'améliorer l'état de l'art dans de nombreux champs d'applications comme la vision par ordinateur, le traitement de la parole et le traitement des langues naturelles. La quantité toujours grandissante de données disponibles et les améliorations du matériel informatique ont permis de faciliter l'apprentissage de modèles à haute capacité comme les ANNs profonds. Cependant, des difficultés inhérentes à l'entraînement de tels modèles, comme les minima locaux, ont encore un impact important. L'apprentissage profond vise donc à trouver des solutions, en régularisant ou en facilitant l'optimisation. Le pré-entraînnement non-supervisé, ou la technique du ``Dropout'', en sont des exemples. Les deux premiers travaux présentés dans cette thèse suivent cette ligne de recherche. Le premier étudie les problèmes de gradients diminuants/explosants dans les architectures profondes. Il montre que des choix simples, comme la fonction d'activation ou l'initialisation des poids du réseaux, ont une grande influence. Nous proposons l'initialisation normalisée pour faciliter l'apprentissage. Le second se focalise sur le choix de la fonction d'activation et présente le rectifieur, ou unité rectificatrice linéaire. Cette étude a été la première à mettre l'accent sur les fonctions d'activations linéaires par morceaux pour les réseaux de neurones profonds en apprentissage supervisé. Aujourd'hui, ce type de fonction d'activation est une composante essentielle des réseaux de neurones profonds. Les deux derniers travaux présentés se concentrent sur les applications des ANNs en traitement des langues naturelles. Le premier aborde le sujet de l'adaptation de domaine pour l'analyse de sentiment, en utilisant des Auto-Encodeurs Débruitants. Celui-ci est encore l'état de l'art de nos jours. Le second traite de l'apprentissage de données multi-relationnelles avec un modèle à base d'énergie, pouvant être utilisé pour la tâche de désambiguation de sens.

Web Science and Data Analytics Course

Resources from the Singapore Summer School 2014 hosted by NUS. ws-summerschool.comp.nus.edu.sg

What can Flickr photographs tell us about the world?

In their second year, our undergraduate web scientists undertake a group project module (WEBS2002, led by Jonathon Hare & co-taught by Su White) in which they get to apply what they learnt in the first year to a practical web-science problem, and also learn about team-working. For the project this semester, the students were provided with a large dataset of geolocated images and associated metadata collected from the Flickr website. Using this data, they were tasked with exploring what this data could tell us about the world. In this seminar the two groups will present the outcomes of their work. Team Alpha (Ellie Hamilton, Clayton Jones & Alok Acharya) will present their work on "The relationship between Group Photos, Social Integration and Suicide". This work aims to explore whether levels of social integration (which Durkheim posited as a factor in "Egoistic Suicide" rates) can be predicted by measuring the proportion of photos of groups of people to photos of individuals within a geographical region. Team Bravo (Agnieszka Grzesiuk-Szolucha, Thomas Leese & Ammaar Tawil) will present their work on "Sentiment Analysis on Flickr Photo Tags to Classify a Photo as Positive or Negative, In Order to Determine the Happiness of a Country or Region". This work explores whether estimates of sentiment made by applying SentiWordNet to Flickr tags correlate with indices of world happiness and socio-economic well-being.