Renato Rocha Souza

Semantics in the Semantic Web: A Critical Evaluation †

In recent years, the term “semantics” has been widely used in various fields of research and particularly in areas related to information technology. One of the motivators of such an appropriation is the vision of the Semantic Web, a set of developments underway, which might allow one to obtain better results when querying on the web. However, it is worth asking what kind of semantics we can find in the Semantic Web, considering that studying the subject is a complex and controversial endeavour. Working within this context, we present an account of semantics relying on the main linguist approaches, in order to then analyze what semantics is within the scope of information technology. We critically evaluate a spectrum, which proposes the ordination of instruments (models, languages, taxonomic structures, to mention but a few) according to a semantic scale. In addition to proposing a new extended spectrum, the article suggests alternative interpretations with the aim of clarifying the use of the term “semantics” in different contexts. Finally, we offer our conclusions regarding the semantic in the Semantic Web and mention future directions and complementary works.

Towards a Taxonomy of KOS: Dimensions for Classifying Knowledge Organization Systems

This paper analyzes previous work on the classifica tion of Knowledge Organization Systems (KOS), discusses strengths and weaknesses, and proposes a new and integrative framework. It argues that current a alyses of the KOS tend to be idiosyncratic and incomplete, relying on a limited number of dimensions of analysis. The paper discusses why and how KOS should be class ified on a new basis. Based on the available literature and previous work, the aut hors propose a wider set of dimensions for the analysis of KOS. These are repre sent d in a taxonomy of KOS. Issues arising are discussed. 1. Representation and Knowledge Organization System s Although central to many scientific fields, or perh aps because of that, the process of representing knowledge is not a simp le atter. There are many distinct theories, models, methodologies and p roducts; all influenced by specific applications, backgrounds an d purposes. Knowledge representation artifacts are produced in related fields, such as Artificial Intelligence, Semiotics, Computer Sci ence and Cognitive Science; but are also widely used in a myriad of le ss related areas, such as Education, Mathematics, Business Modeling, Lingu istics, and many more. The field of Library and Information Science (LIS), however, considers these representational artifacts as a par amount matter. In fact, LIS takes upon itself the task of organizing and fa cilit ting the retrieval of the wealth of information that arises from the k nowledge produced in all other fields, and this involves the creation of epistemological and ontological surrogates. Many of the LIS theories, p rocesses and instruments are dependent on representation product s, modeled through successive abstractions over the relevant character istics of a chosen world or domain, or the information gathered and pr ocessed about these, registered in information systems and documents. Th ose representations, known as Knowledge Organization Sy stems (KOS), vary enormously in format and display, but they sha re the general characteristic of aiding knowledge elicitation and organization, aiming at promoting the retrievability of information. KOS are not new to librarians or biologists, who ha ve been using 1 This is an extended and revised version of a paper presented at the 2010 ISKO

O uso das informações contidas em documentos de patentes nas práticas de Inteligência Competitiva: apresentação de um estudo das patentes da UFMG.

Este trabalho apresenta a Inteligencia Competitiva (IC) como ferramenta minimizadora dos riscos durante a tomada de decisao nas empresas. Apresenta o conceito de IC, seu historico e seus processos. Apresenta o documento de patente como fonte de informacao de expressiva importância para recuperacao de informacoes tecnologicas e, consequentemente, como ferramenta para as praticas de IC. Apresenta a analise do conhecimento protegido atraves dos documentos de patentes depositados pela UFMG no Brasil e no exterior, publicados ate o ano de 2008.

The consumer litigation industry: chasing dragon kings in lawyer-client networks

Abstract The community of lawyers and their clients form a scale-free bipartite network that develops naturally as the outcome of the recommendation process through which lawyers form their client base. This process is an example of preferential attachment where lawyers with more clients are more likely to be recommended to new clients. Consumer litigation is an important market for lawyers. In large consumer societies, there always a significant amount of consumption disputes that escalate to court. In this paper we analyze a dataset of thousands of lawsuits, reconstructing the lawyer–client network embedded in the data. Analyzing the degree distribution of this network we noticed that it follows that of a scale-free network built by preferential attachment, but for a few lawyers with much larger client base than could be expected by preferential attachment. Incidentally, most of these also figured on a list put together by the judiciary of lawyers which openly advertised the benefits of consumer litigation. According to the code of ethics of their profession, lawyers should not stimulate clients into litigation, but it is not strictly illegal. From a network formation point of view, this stimulation can be seen as a separate growth mechanism than preferential attachment alone. In this paper we find that this composite growth can be detected by a simple statistical test, as simulations show that lawyers which use both mechanisms quickly become the “dragon-kings” of the distribution of the number of clients per lawyer.

Comparing three different techniques to retrieve documents using multiwords expressions

This paper aims to compare the results obtained in the retrieval of documents in acorpus using as calculating the similarity function three different techniques adapted for useas descriptors Multi Words Expressions (MWE) extracted from a reference document. Thetechniques tested were Term Frequency-Inverse Document Frequency (TF-IDF), CosineSimilarity Vector (CVS) and Okapi BM25. The goal is to propose an alternative means ofInformation Retrieval Systems (IRS) find in a digital database compound of thesis,dissertations and articles, similar documents from a reference document informed by user. Inthis sense, are considered some structural features of the document. As results, are presented acomparison and analysis of empirical results obtained.

Sentiment Analysis and Knowledge Organization: An Overview of the International Literature

Knowledge organization (KO) as an activity is, among other meanings, a process for conceptual modeling of knowledge domains that produces a consensual abstraction model of this domain with a particular purpose. It adopts a myriad of techniques to analyze and build efficient knowledge organization systems, and one of these techniques is called sentiment analysis (SA) or opinion mining, which is emerging as promising and useful in a variety of ways. It is based in NLP and AI algorithms, and aims at identifying opinions and emotions toward any person, organization or subject; evaluating them as positive or negative, in both binary and graded fashions. This study sought to show various aspects of the implementation of SA for knowledge organization tasks as registered in the scientific literature. We began with exploratory bibliographic research and built a corpus of 91 scientific papers, written in English, selected in the LISA Database, between 2000 to 2016. We analyzed these papers and extracted title, year of publication, author(s) and institution(s), title of the journal where they were published, keywords, the LISA classification code, methods/techniques adopted and its application areas. Our main findings are that theoretical papers still prevail, which may indicate a field in the early stages. We found many institutions and authors from Asia, which points to a new shift in world expertise. We concluded that SA is still a novelty in the KO field, being slowly adopted as an aid to the main tasks, as document classification. Received: 25 November 2016; Revised 8 March 2017; Accepted: 9 March 2017-03-11

A linguagem da ciência: prospecção de dados baseados em corporas

Neste artigo, discutimos brevemente a trajetoria da linguistica de corpus e suas aplicacoes aos estudos da linguagem cientifica. Oferecemos algumas exemplificacoes de ferramentas de exploracao de corpora, aplicacoes de corpora ao estudo da linguagem cientifica e empregos liga n do-os a garimpagem de dados e processamento da linguagem natural.

Sistema de recuperação da informação por busca comparada, que utiliza como descritores expressões multipalavras obtidas através de uma técnica que avalia a estrutura do documento.

Este trabalho visa propor um metodo alternativo para recuperacao de documentos que utiliza Expressoes Multiplavras (EM) extraidas de um documento base, para serem utilizadas como descritores de busca em um Sistema de Recuperacao da Informacao (SRI). Neste sentido, diferentemente dos metodos que consideram o texto como um conjunto de palavras, do ingles bag of words, propusemos um metodo que leva em consideracao as caracteristicas da estrutura fisica do documento no processo de extracao de EM. A partir desse conjunto de termos pre-processados comparamos o uso de tecnica algoritmica exaustiva proposta com o resultado obtido por treze diferentes medidas de associacao estatisticas geradas pelo software Ngram Statistics Package (NSP). Para realizar esse experimento foi constituido um corpus com documentos em formato digital.