Forskning inom corporate social responsibility (CSR) har lyft fram de ekonomiska fördelarna med socialt ansvarstagande och hur investerare reagerar på positiva och negativa nyheter gällande olika aspekter av bolags CSR-arbete. Tredjepartsaktörer har även visat sig kunna överbrygga informationsasymmetri mellan bolag och investerare gällande CSR-arbetet och därmed påverka börsvärdet för berörda bolag. Denna studie undersöker hur tredjepartssignaler i form av ranking inom CSR-området jämställdhet påverkar berörda bolag på Stockholmsbörsen. Detta görs genom en eventstudie där abnormal avkastning beräknas för bolag på dagen för offentliggörandet av stiftelsen Allbrights årliga jämställdhetsrapport där svenska börsbolag kategoriseras och rankas efter graden av jämställdhet. Eventstudien omfattar åren 2012-2018. Studiens resultat visar att negativ kategorisering leder till negativ abnormal avkastning som är koncentrerad till rapportdagen. Vidare indikerar studiens resultat att Allbrights ranking fått ökad kurspåverkan över tid.

The aim of this study was to explore the ranking model TrueSkill™ developed by Microsoft, applying it on various sports and constructing extensions to the model. Two different inference methods for TrueSkill was constructed using Gibbs sampling and message passing. Additionally, the sequential method using Gibbs sampling was successfully extended into a batch method, in order to eliminate game order dependency and creating a fairer, although computationally heavier, ranking system. All methods were further implemented with extensions for taking home team advantage, score difference and finally a combination of the two into consideration. The methods were applied on football (Premier League), ice hockey (NHL), and tennis (ATP Tour) and evaluated on the accuracy of their predictions before each game. On football, the extensions improved the prediction accuracy from 55.79% to 58.95% for the sequential methods, while the vanilla Gibbs batch method reached the accuracy of 57.37%. Altogether, the extensions improved the performance of the vanilla methods when applied on all data sets. The home team advantage performed better than the score difference on both football and ice hockey, while the combination of the two reached the highest accuracy. The Gibbs batch method had the highest prediction accuracy on the vanilla model for all sports. The results of this study imply that TrueSkill could be considered a useful ranking model for other sports as well, especially if tuned and implemented with extensions suitable for the particular sport.

Em 2009, a Federação Internacional de Judô (FIJ) instituiu um sistema de ranqueamento (SR) para classificar os atletas em suas categorias de disputa e para os Jogos Olímpicos de Londres 2012. Porém, o SR não considera o fenômeno conhecido como home advantage, relacionado à vantagem de vencer quando se compete em casa. Esse fenômeno ainda não foi estudado no judô e se comprovado traz implicações para o SR. Assim, o objetivo deste estudo foi verificar se há ocorrência de home advantage nas competições que integram o SR no judô, para a conquista de medalha e número de lutas vencidas. Para tal, foram analisadas todas as competições internacionais de judô do SR realizadas em 2009. Para verificar se os atletas que competiram em casa tiveram mais chance de conquistar medalha foi utilizada a análise de regressão logística e para verificar se houve associação do número de lutas vencidas e competir em casa foi utilizado o modelo linear generalizado de Poisson. Os modelos contaram com uma variável de ajuste referente à qualidade relativa dos atletas em duas possibilidades: posição no SR da FIJ e neutro (construído com lutas disputadas em território neutro). A amostra foi dividida em: Grupo I - todas as participações; Grupo II - somente os atletas que haviam competido dentro e fora de casa. Dado que alguns atletas não puderam lutar em território neutro e, portanto, não tiveram codificação para esse SR, foram criados duas subamostras do Grupo I e do Grupo II, excluindo os atletas sem codificação nesta variável. Todas as análises foram realizadas considerando os gêneros conjunta e separadamente. A variável qualidade relativa foi significante em todas as análises e também modificou os modelos (p<0,001). Ao considerarmos a subamostra do Grupo II, a razão de chance para conquista de medalhas foi superior para os atletas que competiram em casa para o grupo masculino (2,33; p<0,001) e feminino (1,58; p = 0,019). A associação entre número de lutas vencidas e competir em casa foi significante para os atletas do masculino (p<0,001), mas não significante para o feminino (p = 0,088). A ausência da vantagem no feminino para número de lutas vencidas pode ter sido limitada e confundida com o número reduzido de atletas participantes em muitas das competições. As atletas que conquistaram medalha em competições mais vazias, não precisaram vencer mais lutas, dado que um número reduzido de vitórias garantia a medalha. É possível que a torcida seja a principal responsável pelos achados, tanto por influenciar o julgamento dos árbitros a favorecer os atletas da casa como pela motivação para os atletas. Outra possibilidade é que aspectos associados à dominância territorial tenham influenciado as disputas masculinas, porém não as femininas. Assim, foi comprovada a existência do home advantage nas competições que integram o SR de judô para conquista de medalha no masculino e feminino, e para número de lutas vencidas apenas para o masculino. Desta forma, é provável que os atletas de países que sediam competições do SR tenham vantagem para a obtenção da classificação olímpica / In 2009 the International Judo Federation (IJF) established a ranking system (RS) to classify athletes in the competition categories and in the distribution of vacancies for the London 2012 Olympic Games. Nevertheless, the RS does not consider the phenomenon known as home advantage, i.e. the advantage of winning when competing in your home country. This phenomenon has not been studied in judo and its implications for the RS have not been proven. The objective of this study was to determine whether there was occurrence of home advantage in competitions that use the RS in judo, in winning a medal, and in the number of matches won. Therefore, we analyzed all international judo competitions using the RS in 2009. Logistic regression analysis was used to verify if the athletes who competed in their home countries were more likely to win a medal, and we used the Poisson generalized linear model to check whether there were associations between the number of matches won and competing at home. The models have an adjustment variable in the relative quality of athletes in two possibilities: position in the IJF RS and matches played on neutral territory. The sample was divided into two groups: Group I all participants and Group II only athletes who had competed inside and outside their home country. Given that some athletes could not fight on neutral territory and therefore did not have this RS coding, a sub-sample of Group I and a sub-sample of Group II were created, excluding athletes without the coding from this variable. All tests were performed considering the groups together and separately. The relative quality variable was significant in all tests and also modified the models (p <0.001). For the sub-sample of Group II the odds ratio for winning medals was higher for the athletes who competed at home for the male group (2.33, p <0.001) and the female group (1.58, p = 0.019). The association between the number of matches won and competing at home was significant for the male athletes (p <0.001), but not for the females (p = 0.088). The absence of the female advantage in the number of matches won may have been limited and confused by the small number of athletes participating in many of the competitions. The athletes who won medals in competitions with fewer competitors did not need to win more matches since a small number of wins guaranteed the medal. It is possible that the crowd was mainly responsible for the findings, both in influencing the judgement of the referees to favour the home athletes and in motivating the athletes. Another possibility is that aspects related to territorial dominance influenced the male competitors but not the female competitors. Thus, the existence of home advantage was observed in the competitions that used the judo RS in the achievement of medals for males and females, and in the number of matches won, only for males. Thus, it is likely that athletes from countries that host competitions using the RS have an advantage for obtaining the Olympic classification

Instagram is one of the main social platforms for business promotion. Millions of potential customers and endless visual marketing opportunities makes Instagram a perfect place to increase online sales. There are many tools and mechanisms to promote brands on Instagram such as paid advertising or using a pre-generated set of popular hashtags. In this regard, the presence and content of users’ comments becomes an important socio-psychological factor in the motivation to buy or use a product or service. The goal of this degree project is to investigate natural language processing techniques applied to users’ comments on Instagram in order to determine a new algorithm that will include content analysis to the list of feed ranking factors. As it is now, the user has to read through posts on Instagram to get an idea of the quality of a product or service. Therefore, a way to classify and rank products and services is needed. We propose a new algorithm called "Wed 2.0" that can assist consumers in their search of wedding services and products on Instagram. Data mining techniques and sentiment analysis are used to define the mood of the comments and structure user opinions as well as to rank accounts based on this knowledge.

One of the greatest challenges to data mining is erroneous or noisy data. Several studies have noted the weak performance of classification models trained from low quality data. This dissertation shows that low quality data can also impact the effectiveness of feature selection, and considers the effect of class noise on various feature ranking techniques. It presents a novel approach to feature ranking based on ensemble learning and assesses these ensemble feature selection techniques in terms of their robustness to class noise. It presents a noise-based stability analysis that measures the degree of agreement between a feature ranking techniques output on a clean dataset versus its outputs on the same dataset but corrupted with different combinations of noise level and noise distribution. It then considers classification performances from models built with a subset of the original features obtained after applying feature ranking techniques on noisy data. It proposes the focused ensemble feature ranking as a noise-tolerant approach to feature selection and compares focused ensembles with general ensembles in terms of the ability of the selected features to withstand the impact of class noise when used to build classification models. Finally, it explores three approaches for addressing the combined problem of high dimensionality and class imbalance. Collectively, this research shows the importance of considering class noise when performing feature selection. / by Wilker Altidor. / Thesis (Ph.D.)--Florida Atlantic University, 2011. / Includes bibliography. / Electronic reproduction. Boca Raton, Fla., 2011. Mode of access: World Wide Web.

Possibly the largest problem when working in bioinformatics is the large amount of data to sift through to find useful information. This thesis shows that the use of feature selection (a method of removing irrelevant and redundant information from the dataset) is a useful and even necessary technique to use in these large datasets. This thesis also presents a new method in comparing classes to each other through the use of their features. It also provides a thorough analysis of the use of various feature selection techniques and classifier in different scenarios from bioinformatics. Overall, this thesis shows the importance of the use of feature selection in bioinformatics. / by David Dittman. / Thesis (M.S.C.S.)--Florida Atlantic University, 2011. / Includes bibliography. / Electronic reproduction. Boca Raton, Fla., 2011. Mode of access: World Wide Web.

Made available in DSpace on 2016-08-10T10:44:43Z (GMT). No. of bitstreams: 1 BRUNO FRAUZINO RIBEIRO CAMILO.pdf: 1862817 bytes, checksum: ba7c8b48dadf912fb820375d1be59050 (MD5) Previous issue date: 2014-12-15 / The objective was to study the effects on the ordering of the animals by their genetic value of the inclusion of direct maternal additive genetic covariance models of prediction of breeding values for weaning weight in beef cattle using simulated data. The data were simulated considering three different values of maternal direct additive genetic correlation (-0.25, -0.50 and +0.25) and three reasons direct additive variance and maternal additive (1: 1, 1: 2 and 2: 1), making a total of nine scenarios. In each scenario, the breeding values were predicted including whether or not the covariance in the model. The values of the Spearman correlation between animal clinics, considering the predicted genetic value, the scenarios of direct maternal additive genetic correlation -0.50 in the ratios of variances 1: 1 and 2: 1 were equal to 0.84 and 0, 85, respectively; -0.25 scenario the correlation variance in the ratios 1: 1, 1: 2 and 2: 1 were 0.96, 0.95 and 0.97 respectively, and the positive correlation scenario in +0.25 Variance ratios 1: 1, 1: 2 and 2: 1 with values of 0.92, 0.91 and 0.96 respectively. The values were high and significant, suggesting greater consistency between the animal merit orders. The average value for the Spearman correlation estimated in scenario - 0.50 1: 2 was equal to 0.46 indicating lesser proximity between merit orders of individuals in this scenario. The results indicate include the covariance in prediction models for the classification of animals are closer to real. / Objetivou-se estudar os efeitos, sobre a ordenação dos animais por seu valor genético, da inclusão da covariância genética aditiva direta maternal nos modelos de predição dos valores genéticos de peso a desmama de bovinos de corte utilizando dados simulados. Os dados foram simulados considerando-se três valores diferentes de correlação genética aditiva direta maternal (-0,25, -0,50 e +0,25) e três razões de variância aditiva direta e aditiva maternal (1:1, 1:2 e 2:1), perfazendo um total de nove cenários. Em cada cenário, os valores genéticos foram preditos incluindo-se ou não a covariância no modelo. Os valores da correlação de Spearman entre os postos dos animais, considerando o valor genético predito, nos cenários de correlação genética aditiva direta maternal -0,50 nas razões de variâncias 1:1 e 2:1 foram iguais a 0,84 e 0,85, respectivamente; no cenário de correlação -0,25 nas razões de variâncias 1:1, 1:2, e 2:1 foram iguais a 0,96, 0,95 e 0,97 respectivamente e no cenário de correlação positiva +0,25 nas razões de variâncias 1:1, 1:2 e 2:1 com valores de 0,92, 0,91 e 0,96 respectivamente. Os valores foram altos e significativos, o que sugere maior concordância entre as ordens de mérito dos animais. O valor médio para a correlação de Spearman estimados no cenário -0,50 1:2 foi igual a 0,46 indicando menor proximidade entre as ordens de mérito dos indivíduos neste cenário. Os resultados indicam incluir a covariância nos modelos de predição para que as classificações dos animais estejam mais próximas do real.

In this thesis, we explore the issues of uncertain data management in several different aspects. First, we propose a novel linear time algorithm to compute the positional probability, the computation of which is a primitive operator for most of the ranking definitions. Our algorithm is based on the conditional probability formulation of positional probability and the system of linear equations. Based on the formulation of conditional probability, we also prove a tight upper bound of the top-k probability of tuples, which is then used to stop the top-k computation earlier. Second, we study top-k probabilistic ranking queries with joins when scores and probabilities are stored in different relations. We focus on reducing the join cost in probabilistic top-k ranking. We investigate two probabilistic score functions, namely, expected rank value and probability of highest ranking. We give upper/lower bounds of such probabilistic score functions in random access and sequential access, and propose new I/O efficient algorithms to find top-k objects. Third, we extend the possible worlds semantics to probabilistic XML ranking query, which is to rank top-k probabilities of the answers of a twig query in probabilistic XML data. The new challenge is how to compute top-k probabilities of answers of a twig query in probabilistic XML in the presence of containment (ancestor/descendant) relationships. We focus on node queries first, and propose a new dynamic programming algorithm which can compute top-k probabilities for the answers of node queries based on the previously computed results in probabilistic XML data. We further propose optimization techniques to share the computational cost. We also show techniques to support path queries and tree queries. Fourth, we study how to rank documents using a set of keywords, given a context that is associated with the documents. We model the problem using a graph with two different kinds of nodes (document nodes and multi-attribute nodes), where the edges between document nodes and multi-attribute nodes exist with some probability. We discuss its score function, cost function, and ranking with uncertainty. We also propose new algorithms to rank documents that are most related to the user-given keywords by integrating the context information. / Uncertain data management has received a lot of attentions recently due to the fact that data obtained can be incomplete or uncertain in many real applications. Ranking of uncertain data becomes an important research issue, the possible worlds semantics-based ranking makes it different from the ranking of deterministic data. In the traditional deterministic data, we can compute a score for each object, and then the objects are ranked based on the computed scores. However, in the scenario of uncertain data, each object has a probability to be the true answer (or the existence probability), besides the computed score. A probabilistic top-k ranking query ranks objects by the interplay of score and probability based on the possible worlds semantics. Many definitions have been proposed in the literature based on the possible worlds semantics. / Chang, Lijun. / Advisers: Hong Cheng; Jeffrey Xu Yu. / Source: Dissertation Abstracts International, Volume: 73-06, Section: B, page: . / Thesis (Ph.D.)--Chinese University of Hong Kong, 2011. / Includes bibliographical references (leaves 131-139). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [201-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract also in Chinese.

La théorie de l'argumentation abstraite de Dung est un formalisme permettant d'utiliser un système d'argumentation afin de représenter des informations conflictuelles. Des sémantiques à base d'extensions ont d'abord été introduites dans le but de déterminer quels arguments peuvent être conjointement acceptés. Cependant, ces sémantiques ne sont pas appropriées pour certaines applications, c'est pourquoi des sémantiques à base de classement, permettant de classer les arguments du plus acceptable au moins acceptable, ont été introduites. Le but de cette thèse est donc de proposer et d'étudier ces sémantiques à base de classement dans le contexte de l'argumentation abstraite.Nous définissons d'abord une nouvelle famille de sémantiques à base de classement basées sur un principe de propagation permettant de contrôler l'influence des arguments non-attaqués sur l'acceptabilité des arguments. Nous étudions les propriétés de ces sémantiques, les relations entre elles ainsi qu'avec d'autres sémantiques existantes.Nous proposons ensuite deux méthodes pour comparer les sémantiques à base de classement. La première est une comparaison empirique sur des systèmes d'argumentation générés aléatoirement donnant un aperçu des similitudes et des différences entre ces sémantiques. La seconde est une comparaison axiomatique de toutes ces sémantiques à la lumière des propriétés proposées visant à mieux comprendre le comportement de chaque sémantique.Enfin, nous remettons en question la capacité des sémantiques existantes à capturer certains principes de persuasion et introduisons une nouvelle sémantique paramétrée à base de classement plus appropriée pour ce contexte précis. / Dung’s theory of abstract argumentation is a formalism that represents conflicting information using an argumentation framework. Extension-based semantics have been introduced to determine, given an argumentation framework, the justifiable points of view on the acceptability of the arguments. However, these semantics are not appropriate for some applications. So alternative semantics, called ranking-based semantics, have recently been evolved. Such semantics produces, for a given argumentation framework, a ranking on its arguments from the most acceptable to the least one(s). The overall aim of this thesis is to propose and study ranking-based semantics in the context of abstract argumentation. We first define a new family of ranking-based semantics based on a propagation principle which allow us to control the influence of non-attacked arguments on the acceptability of arguments. We investigate the properties of these semantics, the relationships between them but also with other existing semantics. Then, we provide a thorough analysis of ranking-based semantics in two different ways. The first one is an empirical comparison on randomly generated argumentation frameworks which reveals insights into similarities and differences between ranking-based semantics. The second one is an axiomatic comparison of all these semantics with respect to the proposed properties aiming to better understand the behavior of each semantics. At last, we question the ability of the existing ranking-based semantics to capture persuasion settings and introduce a new parametrized ranking-based semantics which is more appropriate in this context.

Data from rating scale assessments have rank-invariant properties only, which means that the data represent an ordering, but lack of standardized magnitude, inter-categorical distances, and linearity. Even though the judgments often are coded by natural numbers they are not really metric. The aim of this thesis is to further develop the nonparametric rank-based Svensson methods for paired ordinal data that are based on the rank-invariant properties only. The thesis consists of five papers. In Paper I the asymptotic properties of the measure of systematic disagreement in paired ordinal data, the Relative Position (RP), and the difference in RP between groups were studied. Based on the findings of asymptotic normality, two tests for analyses of change within group and between groups were proposed. In Paper II the asymptotic properties of rank-based measures, e.g. the Svensson’s measures of systematic disagreement and of additional individual variability were discussed, and a numerical method for approximation was suggested. In Paper III the asymptotic properties of the measures for paired ordinal data, discussed in Paper II, were verified by simulations. Furthermore, the Spearman rank-order correlation coefficient (rs) and the Svensson’s augmented rank-order agreement coefficient (ra) were compared. By demonstrating how they differ and why they differ, it is emphasized that they measure different things. In Paper IV the proposed test in Paper I for comparing two groups of systematic changes in paired ordinal data was compared with other nonparametric tests for group changes, both regarding different approaches of categorising changes. The simulation reveals that the proposed test works better for small and unbalanced samples. Paper V demonstrates that rank invariant approaches can also be used in analysis of ordinal data from multi-item scales, which is an appealing and appropriate alternative to calculating sum scores.