This work finds that Managerial and Investor Sentiment are determined by differing sets of economic variables, that share some common factors: inflation, liquidity and the term premium. Decomposing the Sentiment Indices, it is found that the Investor Sentiment Model Component and the Managerial Sentiment Residual Component are primarily responsible for the predictive power of predicting cross-sectional stock returns that is much stronger than previous results. Evidence is presented that part of the predictive power is due to the components predicting priced market factors. Overall, there is strong evidence that the predictive power of Managerial Sentiment is driven primarily by private information, while the predictive power of Investor Sentiment is driven by public information.

This work is concerned with the automatic understanding of evaluative text. We investigate sentence level opinion polarity prediction by assigning lexical polarities and deriving sentence polarity from these with the use of contextual valence shifters. A methodology for iterative failure analysis is developed and used to refine our lexicon and identify new contextual shifters. Algorithms are presented that employ these new shifters to improve sentence polarity prediction accuracy beyond that of a state-of-the-art existing algorithm in the domain of consumer product reviews. We then apply the best configuration of our algorithm to the domain of movie reviews.

This work is concerned with the automatic understanding of evaluative text. We investigate sentence level opinion polarity prediction by assigning lexical polarities and deriving sentence polarity from these with the use of contextual valence shifters. A methodology for iterative failure analysis is developed and used to refine our lexicon and identify new contextual shifters. Algorithms are presented that employ these new shifters to improve sentence polarity prediction accuracy beyond that of a state-of-the-art existing algorithm in the domain of consumer product reviews. We then apply the best configuration of our algorithm to the domain of movie reviews.

Introduction: Comment les aidants familiaux selon qu’ils soient conjoint ou enfant et que leur proche atteint de maladie d’Alzheimer habite à domicile ou en institution, perçoivent-ils le stress qu’ils vivent et le soutien qu’ils reçoivent ? Quels aspects négatifs mais aussi positifs de la relation d’aide sont susceptibles d’influencer leur santé psychique ? L’étude des effets respectifs des dimensions du fardeau et du sentiment de compétence sur la santé psychique des aidants apporte des éléments de réponse.Méthode: 264 aidants familiaux remplissent des questionnaires d’auto-évaluation du soutien social perçu (SSQ-6), de la dépression (BDI‑2), de l’anxiété (STAI), de l’hostilité (STAXI-2) et de la satisfaction de vie (SWLS). Une analyse en composantes principales des échelles de fardeau Zarit (BI) et du sentiment de compétence (SCQ) est effectuée. Un modèle d’équations structurales est testé simultanément dans quatre groupes d’aidants : statut (enfants vs conjoints) x contexte (domicile vs institution). Résultats: L’aide fournie par les conjoints et les enfants a des effets indirects multiples sur leur santé psychique via les composantes du stress perçu : la détérioration de leur vie personnelle, leur culpabilité et leur satisfaction vis-à-vis de sa compétence. Le statut et le contexte ont un effet modérateur sur le poids de ces relations. Chez les conjoints aidants à domicile le soutien social perçu n’a pas d’effet protecteur de la santé psychique.Conclusions: Dans le cadre du modèle transactionnel du stress, cette étude pointe le rôle de l’évaluation que fait les aidants de leurs difficultés et du soutien. Les implications en terme d’aide aux aidants sont discutées. / Introduction: How do spouses and children of older people with dementia living in the community or in an institution perceive the the stress related to caregiving and the support provided by the professionals? Which negative and positive aspects of caregiving influence the caregivers’ mental health? This study examines the specific effects of the burden of caring and their sense of competence on caregivers’ mental health.Methods: Data are gathered from 264 family caregivers, using self-rating scales measuring perceived social support (6 items SSQ6), depression (BDI-II), anxiety (STAI-Y), hostility (STAXI-2), satisfaction with life (SWLS). A principal component analysis of the 22 items of Zarit Burden Inventory (BI) and 27 items of the Sense of Competence Questionnaire (SCQ) is computed. Structural Equation Models (SEMs) are tested simultaneously in four conditions: children vs spouses (status), the care recipient living at home vs institutionnalised (context).Results : Caring indirectly affects the caregiver’s mental health via the perceived stress components : consequences for the caregiver’s personal life (SCQ), caregiver’s feelings of guilt (BI) and satisfaction with one’s performance as caregiver (SCQ). These effects are moderated by the caregiving context and the caregiver status. Perceived social support does not protect the mental heath of spouses caring for their relative at home.Conclusion: According to the stress process model this study highlights the key role of the caregiver’s evaluation of his/her own difficulties and perceived support. Implications for interventions and support for carers are discussed.

This work is concerned with the automatic understanding of evaluative text. We investigate sentence level opinion polarity prediction by assigning lexical polarities and deriving sentence polarity from these with the use of contextual valence shifters. A methodology for iterative failure analysis is developed and used to refine our lexicon and identify new contextual shifters. Algorithms are presented that employ these new shifters to improve sentence polarity prediction accuracy beyond that of a state-of-the-art existing algorithm in the domain of consumer product reviews. We then apply the best configuration of our algorithm to the domain of movie reviews. / Science, Faculty of / Computer Science, Department of / Graduate

This dissertation mainly explores the effect of investor sentiment on stock returns and volatility on Growth Enterprise in China using monthly data from Shenzhen Stock Exchange of China from June 2010 to November 2019. Using five explicit and market-related implicit indicators an investor sentiment has been measured and constructed with the help of principal component analysis. The analysis has been done by employing a vector autoregression(VAR) model and impulse response functions (IRFs) generated from a VAR model to examine the relationship between the unanticipated changes in investor sentiment and stock returns and volatility. We also establish EGARCH model to test the validity of previous results and if the asymmetric impact of positive and negative news on market returns volatility. The results show a significant impact of investor sentiment on stock return and volatility. We also document that there is a positive leverage effect between investor sentiment and the volatility of returns. The findings of this paper can help both individual and institutional investors have a better understanding of GEM market and improve their investment returns by incorporating investor sentiment into their asset forecasting model. This paper also provides policymakers guidance on reducing volatility on stock markets from the perspective of investor sentiment. Additionally, this paper has important contributions to behavioral finance and adds to the limited number of studies on investor sentiment and stock return in not only the Chinese market but emerging markets.

A large proportion of World Wide Web (WWW) users treat it as a social medium, i.e. many of them use the WWW to express and communicate their opinions. Economic value or utility can be created if these utterances, reactions, or feedback are extracted from various social media platforms and their content analysed. Some of these benefits are related to e-commerce, marketing, product improvements, improving machine learning algorithms etc. Moreover, establishing links between different social media platforms, based on shared topics and content, could provide access to the comments of users of different platforms. However, studies to date have generally tackled the area of content extraction from each type of social media in isolation. There is a lack of research of some aspects of social media, namely, linking the references from a blog post, for example, to information related to the same issue on Twitter. In addition, while studies have been carried out on various languages, there has been little investigation into social media in the Arabic language. This thesis tackles opinion mining and sentiment analysis of Arabic language social media, particularly in blogs and Twitter. The thesis focuses on Arabic language technology blogs in order to identify the expressed sentiments and then to link an issue within a blog post to relevant tweets in Twitter. This was done by assessing the similarity of content and measuring the sentiments scores. In order to extract the required data, text-mining techniques were used to build up corpora of the raw blog data in Modern Standard Arabic (MSA) and to build tools and lexicons required for this research. The results obtained through this research contribute to the field of computer science by furthering the employment of text-mining techniques, thus improving the process of information retrieval and knowledge accumulation. Moreover, the study developed new approaches to working with Arabic opinion mining and the domain of sentiment analysis.

<div>I examine movements in the closing price that are different than the movements of the high and low prices on a given day. I construct the measure HLDiff which accumulates the differences between the high-low midpoint return and the closing price return over a month. Instances in which the closing price deviates from the movements in the midpoint between the high and low are a strong predictor of future abnormal returns. The predictive power of the HLDiff measure holds across size groups and sub-periods and holds in the presence of other common determinants of stock returns. The predictive power of HLDiff appears to be driven by the existence of market frictions. Specifically, I find that the premium associated with a factor constructed based on HLDiff is consistent with short-selling constraints inhibiting the correction of overpricing. The factor also appears to improve the pricing ability of the single-factor and five-factor models.</div>

There exist a lot of empirical researches, that examine what factors effect the stock market volatility. The concept of investor sentiment is quite popular and is frequently discussed. However, there does not exist any research which would study the relation between the change in election preferences during the presidential campaigns and stock market volatility. The present thesis explores the effect of political sentiment on United States and French models. Here, we construct the model, which examines the effect of change in election preferences on the volatility. The results suggest, that change in election preferences does not affect the stock market volatility during the presidential campaign. Thus, its inclusion to the model does not increase the prediction power.

Cette thèse, rédigée sous la forme de trois essais formant un ensemble, a pour objectif d'expliciter le fondement humaniste et existentiel du phénomène de l'appartenance. Le premier essai, intitulé Habiter, Demeurer, Appartenir, développe la dialectique entre l'acte d'appartenir et celui de demeurer, au sein de la notion plus vaste de l'habiter. L'essai explore l'intime interrelation entre l'acte d'habiter une maison, qui renvoie au fait de demeurer, et l'acte d'habiter avec la maisonnée qui relève proprement de l'appartenance. Intitulé Chez-soi et Chez-nous. Figures habitées de l'appartenance, le second essai, exploite le thème de l'appartenance par le mode d'habitation des lieux. Proposant une distinction entre l'expérience de se sentir chez-soi et celle de se sentir chez-nous, l'essai en développe les aspects spatiaux, relationnels, mythiques et identitaires, pour finalement présenter l'appartenance comme étant à l'articulation de la verticalité du chez-soi et de l'horizontalité du chez-nous. Cette nouvelle perspective conduira à dégager certaines pathologies de l'appartenance. Le troisième essai, intitulé Deux modes de l'appartenance, commence par une critique du point de vue naturaliste dominant sur l'appartenance. La critique porte sur le risque d'aliénation existentielle et sur la négligence de l'altérité propre à cette conception qui limite l'appartenance à une attribution catégorielle fondée sur le partage d'attributs similaires. L'essai propose une définition et une caractérisation existentielle de l'appartenance, conçue comme une condition ontologique de l'être humain et en tant qu'une relation participative entre une personne et une unité englobante. La conception proposée se fonde sur une éthique du vivre-ensemble qui veut contrer la tendance actuelle marquée par une vision individualiste d'appropriation et de catégorisation. Pour conclure, le phénomène de l'appartenance est également interrogé sous ces deux formes discursives: « j'appartiens à ... » et « cela m'appartient ». L'appartenance sera finalement définie comme étant le rapprochement équilibré des horizons de l'expérience personnelle et de l'expérience collective au sein d'une communauté, qui se réalise par une double articulation, premièrement, celle de la rencontre de l'horizon de notre espace vécu avec l'horizon des lieux collectifs, puis, deuxièmement, celle de la rencontre de l'horizon de nos récits personnels avec l'horizon mythique du récit collectif identitaire qui fonde les communautés auxquelles nous appartenons.