O presente trabalho, resultado das pesquisas do Grupo de Morfologia Histórica do Português, procura estudar os aspectos semânticos do sufixo -ista no português e no galego, baseando-se na afirmação de VIARO (2005), segundo a qual a semântica da base é diferente da semântica do sufixo e da semântica da palavra formada pela derivação com este, e ressaltando, que não necessariamente a semântica da palavra formada é a soma das semânticas do sufixo e base. Desde esta ótica, traça-se, inicialmente, um panorama histórico e um mapeamento geral da semântica das palavras derivadas com o sufixo proveniente da terminação grega -#953;στης em várias línguas, e, posteriormente, faz-se um estudo em corpora das línguas portuguesa e galega, propondo-se uma classificação semântica das palavras formadas com o sufixo -ista, a qual, detalhada por perspectivas históricas e etimológicas, conduz a uma classificação semântica do sufixo -ista, no português e no galego. / The present work, which is the result of the researches carried out by the Grupo de Morfologia Histórica do Português [\"Group of Historical Morphology of the Portuguese Language\"], studies the semantic aspects of the suffix -ista, in Portuguese and Galician, based on VIARO\'s (2005) assertion, according to which the base semantics is different from the suffix semantics and also different from the semantics of the word formed by the derivation with this suffix, and the semantics of the word formed isn\'t necessarily the sum of the base semantics with the suffix semantics. From this standpoint, we initially draw a historical overview and a general map of the semantics of the words derived with the suffix originating from the Greek ending -ιστης, in several languages. Based on an analysis of the Portuguese corpora and the Galician corpora, we subsequently propose a semantic classification of the words formed by the suffix -ista, which, after providing a detailed historical and etymological perspective, leads to a semantic classification of the suffix -ista, in the Portuguese and Galician languages.

Suffix tree is a data structure which enables the performing of fast search-like operations on the text. In order to be used efficiently, it must be created quickly. In this thesis, we focus on the new kind of suffix link simulation called "minimized branching", which aims to increase the speed of suffix tree construction by reducing the number of branching operations. Our main goal is to present a comparison of the currently used methods for the suffix tree construction and to point out some advantages and disadvantages of the individual approaches. We introduce, implement and practically evaluate multiple variations of the standard McCreight's and Ukkonen's algorithms, as well as Partition and Write Only Top Down (PWOTD) algorithm, originally developed for disk-based construction. Our main result is the integrated description and implementation of these algorithms, which are both well-suited to be further built upon. We also present a simple recommendations on when it is advisable to use a particular algorithm's variation and why.

The present thesis treats one of the word-formative processes, derivation, in two Indo-European languages, Latin and Slavic. It focuses on the description and comparison of nominal (substantive) derivative suffixes and whole word-formative types that in these languages correspond to each other. The main focus of the thesis is to make a classification of these parallels from a functional point of view and to describe particular word-formative types from a scope of the onomasiologic categories. Since Latin and Slavic do not display similar parallels regarding other language levels, the considerable similarity of the derivative suffixes (and whole word-formative types) is a noticable phenomenon. One of its possible explanations is the fact that both languages have retained the original Indo-European word structure. Therefore this thesis uses as a base the Latin language as the older one which is thus formally and functionally more likely closer to the original (Indo-European) forms than Slavic, and traces particular suffixes (word-formative types) throughout their developement and attempts to reveal their formal and semantic shifts.

Esta tese apresenta, por um lado, uma descrição sumária do funcionamento dos sufixos latinos -men e -mentum em relação à sua produção lexical, às suas características morfológicas e semânticas e a informação semântica estrutural de suas palavras derivadas, no Latim. Por outro, analisa, de modo exaustivo, como o sufixo -mentum desenvolveu-se nas principais línguas românicas: Castelhano, Francês, Italiano e Romeno, para poder contrastar com o Português. Apresentamos primeiramente uma proposta de origem para o sufixo latino - mentum, diferente da bibliografia existente sobre o assunto, que indica que -mentum advém de um alargamento de -men. Para isso, montamos um corpus de análise das palavras latinas, identificando as bases nominal e verbal das palavras em -men e em -mentum; em seguida, analisamos a datação dessas palavras para, ao final, verificar o significado do sufixo em cada uma delas, resultando numa proposta de genealogia dos sufixos -men / -mentum. Na segunda parte desta tese, apresentamos a derivação panromânica do sufixo latino -mentum. Montamos um corpus para cada língua pesquisada, de modo exaustivo, com indicação de data de entrada, étimo, origem e aspecto semântico. Em cada capítulo, analisamos todos os itens, exceto no Português, em que não realizamos pesquisa feita previamente em Freitas (2008), sobre a semântica dessas palavras. Em todas as línguas, propusemos uma árvore genealógica do sufixo estudado sob um aspecto semântico, mostrando o caminho feito pelo sufixo por meio de seus significados. Ao final, propomos um modelo de análise etimológica para as línguas românicas e uma genealogia do sufixo -mentum, do Latim às românicas. / This thesis presents, on the one hand, a brief description of the functioning of Latin suffixes -men and -mentum related to their lexical production, their morphological and semantic characteristics as well as the structural semantic information of their derived words in Latin. On the other hand, it analyses exhaustively the development of the suffix -mentum in the main Romance languages: Spanish, French, Italian and Romanian, so that we can contrast them with Portuguese. First, we present a proposal for the origins of the Latin suffix -mentum, different from the existing bibliography on the subject, which suggests that -mentum comes from an enlargement of -men. Thus, we have organised a corpus for the analysis of the Latin words, identifying the nominal and verbal bases of words ending in -men and -mentum; then, we analyse the datings of these words, so that we can verify the meaning of the suffix in each one of them, which results in a proposal for the genealogy of both suffixes -men / -mentum. In the second part of the thesis, we present the PanRomance derivation of Latin suffix -mentum. We have organized an exhaustive corpus for each language, indicating the first datings, etymon, origin and semantic aspect. In each chapter, we have analysed every item, but for Portuguese, in which we have not made the research previously presented in Freitas (2008) about the semantics of these words. For every language we have proposed a genealogical tree for the suffix, in its semantic aspect, showing the routes taken by the suffix through its meanings. Finally, we have proposed a model for etymological analysis for Romance languages, and a genealogy of the suffix -mentum, from Latin to Romance languages.

In this paper we investigate the potential of the analysis of noisy non-stationary time series by quantizing it into streams of discrete symbols and applying finite-memory symbolic predictors. The main argument is that careful quantization can reduce the noise in the time series to make model estimation more amenable given limited numbers of samples that can be drawn due to the non-stationarity in the time series. As a main application area we study the use of such an analysis in a realistic setting involving financial forecasting and trading. In particular, using historical data, we simulate the trading of straddles on the financial indexes DAX and FTSE 100 on a daily basis, based on predictions of the daily volatility differences in the underlying indexes. We propose a parametric, data-driven quantization scheme which transforms temporal patterns in the series of daily volatility changes into grammatical and statistical patterns in the corresponding symbolic streams. As symbolic predictors operating on the quantized streams we use the classical fixed-order Markov models, variable memory length Markov models and a novel variation of fractal-based predictors introduced in its original form in (Tino, 2000b). The fractal-based predictors are designed to efficiently use deep memory. We compare the symbolic models with continuous techniques such as time-delay neural networks with continuous and categorical outputs, and GARCH models. Our experiments strongly suggest that the robust information reduction achieved by quantizing the real-valued time series is highly beneficial. To deal with non-stationarity in financial daily time series, we propose two techniques that combine ``sophisticated" models fitted on the training data with a fixed set of simple-minded symbolic predictors not using older (and potentially misleading) data in the training set. Experimental results show that by quantizing the volatility differences and then using symbolic predictive models, market makers can generate a statistically significant excess profit. However, with respect to our prediction and trading techniques, the option market on the DAX does seem to be efficient for traders and non-members of the stock exchange. There is a potential for traders to make an excess profit on the FTSE 100. We also mention some interesting observations regarding the memory structure in the studied series of daily volatility differences. (author's abstract) / Series: Report Series SFB "Adaptive Information Systems and Modelling in Economics and Management Science"

Texte remanié de: Diss.--Philosophische Fakultät I--Zürich--Universität, 1973. / Bibliogr. p. 17-22. Notes bibliogr. Index.

Title: Suffix Graphs and Lossless Data Compression Author: Martin Senft Department: Department of Software and Computer Science Education Supervisor of the doctoral thesis: doc. RNDr. Tomáš Dvorˇák, CSc., Depart- ment of Software and Computer Science Education Abstract: Suffix tree and its variants are widely studied data structures that enable an efficient solution to a number of string problems, but also serve for implementation of data compression algorithms. This work explores the opposite approach: design of compression methods, based entirely on prop- erties of suffix graphs. We describe a unified construction algorithm for suf- fix trie, suffix tree, DAWG and CDAWG, accompanied by analysis of implicit suffix link simulation that yields two practical alternatives. Since the com- pression applications require maintaining text in the sliding window, an in- depth discussionof slidingsuffixgraphsisneeded. Fillinggapsin previously published proofs, we verify that suffix tree is capable of perfect sliding in amortised constant time. On the other hand, we show that this is not the case with CDAWG, thus resolving a problem of Inenaga et al. Building on these investigations,we describea family of data compression methods,based on a description of suffix tree construction for the string to be compressed. While some of...

The main topic of this thesis is the analytic predicate with event nouns with the suffix -ement in French. It is divided into two parts, the theoretical and the practical one. In the first theoretical part we concentrate on the description of the analytic predicate and we also introduce some test for its identification. The next parts are dedicated to the particular components of this predicate: to the light verb and the predicative noun with a particular emphasis on event nouns, for which we examine their possible typologies. After this we focus on the suffix ement itself, its characteristics as well as the characteristics of nouns, which are created by this suffix. The last topic which we cover is the concurrence between suffixes -ement and age. In the practical part, we begin by describing used methods thanks to which we received needed data as well as our instruments. Then we divided the extracted nouns with the suffix ement according to their meaning into a few groups and we continued by verifying their eventive meaning. Subsequently we concentrated on the collocability of chosen nouns with verbs and we finished by comparing the suffix ement with suffix -age.

As formações x-inho são descritas, na maioria das gramáticas de Língua Portuguesa como contendo noções dimensiva e afetiva. Entretanto, essas mesmas gramáticas não incluem os fatores extraligüísticos e contextuais nos quais os anunciadores estão inseridos quando optam por uma formação em x-inho. Sob esta perspectiva, tem-se no presente trabalho, o objetivo de investigar a produtividade das formações x-inho em dois corpora eletrônicos: um oral, subdividido em dois subcorpora contendo respectivamente narrativas e descrições e um escrito, oriundo exclusivamente das variadas seções e cadernos de um jornal de grande circulação e qualidade. A dissertação quantifica as ocorrências das formações x-inho em cada um dos corpora. Em seguida cada uma dessas ocorrências é analisada para se verificar que tipo de noção (dimensiva, afetiva positiva ou negativa, intensificadora, etc) ela contem. Por fim são contrastados os dados de freqüência e dispersão de cada uma das noções encontradas para cada um dos corpora. A metodologia de nossa análise está centrada na área de investigação lingüística denominada Lingüística de corpus, que serve de base para que os dados colhidos sejam analisados e interpretados. / The items ending in -inho are described in the majority of grammars of Portuguese as conveying two notions, namely affect and dimension. However, the same grammars do not seem to include either the extralinguistic or contextual factors in which speakers are inserted when they opt for a word ending in -inho. The aim of the present work thus is to investigate the productivity of such items in two electronic corpora: one of an oral nature which is further subdivided into two sub-corpora containing narratives and descriptions and a second one compiled exclusively from the various sections of a widely read quality newspaper. The dissertation quantifies the various instances of items ending in inho in each of the corpora. Next, each of these occurrences is analysed and classified to check which notion (dimentio,positive affect, negative affect, intensification) they convey. Last the results of both frequency and dispension counts are contrasted for each of the corpora. The methodology of our analyses is centered on the area known as Corpus Linguistics, which provides a basis for the data to be compiled and interpreted.

As formações x-inho são descritas, na maioria das gramáticas de Língua Portuguesa como contendo noções dimensiva e afetiva. Entretanto, essas mesmas gramáticas não incluem os fatores extraligüísticos e contextuais nos quais os anunciadores estão inseridos quando optam por uma formação em x-inho. Sob esta perspectiva, tem-se no presente trabalho, o objetivo de investigar a produtividade das formações x-inho em dois corpora eletrônicos: um oral, subdividido em dois subcorpora contendo respectivamente narrativas e descrições e um escrito, oriundo exclusivamente das variadas seções e cadernos de um jornal de grande circulação e qualidade. A dissertação quantifica as ocorrências das formações x-inho em cada um dos corpora. Em seguida cada uma dessas ocorrências é analisada para se verificar que tipo de noção (dimensiva, afetiva positiva ou negativa, intensificadora, etc) ela contem. Por fim são contrastados os dados de freqüência e dispersão de cada uma das noções encontradas para cada um dos corpora. A metodologia de nossa análise está centrada na área de investigação lingüística denominada Lingüística de corpus, que serve de base para que os dados colhidos sejam analisados e interpretados. / The items ending in -inho are described in the majority of grammars of Portuguese as conveying two notions, namely affect and dimension. However, the same grammars do not seem to include either the extralinguistic or contextual factors in which speakers are inserted when they opt for a word ending in -inho. The aim of the present work thus is to investigate the productivity of such items in two electronic corpora: one of an oral nature which is further subdivided into two sub-corpora containing narratives and descriptions and a second one compiled exclusively from the various sections of a widely read quality newspaper. The dissertation quantifies the various instances of items ending in inho in each of the corpora. Next, each of these occurrences is analysed and classified to check which notion (dimentio,positive affect, negative affect, intensification) they convey. Last the results of both frequency and dispension counts are contrasted for each of the corpora. The methodology of our analyses is centered on the area known as Corpus Linguistics, which provides a basis for the data to be compiled and interpreted.