Dans cette étude trilingue, nous abordons le passage d’un signe linguistique du lexique courant à la terminologie de la macroéconomie. En effet, la sémantique d’un signe linguistique dans le lexique courant et dans la terminologie semble, de prime abord, identique. Cependant, à l'analyse, des différences sémantiques commencent à émerger. Nous proposons des méthodes d’analyse pour relever les traits sémantiques différents d’un signe linguistique quand il passe du lexique courant à la terminologie de la macroéconomie. D'ailleurs, nous avons porté notre attention sur les termes composés de deux éléments. Le traitement commence par l’élément de base selon des catégories et sous-catégories basées sur les traits sémantiques généraux ou sur le domaine d’origine de l’élément de base. De plus, nous divisons les termes composés en fonction de la structure syntaxique. Ainsi, un chapitre est consacré au traitement des termes composés d’un nom et d'un adjectif, et un autre au traitement des termes composés de deux noms. La structure différente des termes composés nous incite à aborder les types de composés différemment. Pour explique cela, nous remarquons que l'élément de base, dans le composé N1-A1, tende à s'associer avec un adjectif de manière préférentielle ou obligatoire. Ainsi, nous adaptons le concept de la « polarisation » pour traiter le comportement du composé N1-A1 où l'élément de base polarise les adjectifs associés. Par contre, le composé N1-N2 subit une sorte de dilution où l'élément de base peut être remplacé par un synonyme sans changer le concept. D'ailleurs, le composé N1-N2 peut changer de structure sans changer le concept. par conséquent, nous employons le terme « dilution » pour décrire ce comportement.D’un autre côté, nous proposons quelques analyses pour comprendre la sémantique de l’état d’annexion en ce qui concerne les termes composés en arabe N1-N2. Pour ce faire, nous remplaçons l’état d’annexion par des tournures prépositionnelles pour expliciter la sémantique de l’état d’annexion. Cela permet aussi d'établir une catégorisation sémantique de l’état d’annexion. / In this trilingual study, we discuss the passage of a linguistic sign from the current lexicon to the terminology of macroeconomics. Indeed, the semantics of a linguistic sign in the current lexicon and in terminology seem, at first sight, identical. Nevertheless, it is only a detailed analysis, semantic differences begin to emerge. We propose methods in order to analyze the different semantic features of a linguistic sign as it passes from the current lexicon to the terminology of macroeconomics.As for compound terms, we only deal with terms composed of two elements. The treatment begins with the base element according to categories and subcategories based on the general semantic features or on the origin of the base element. In addition, we divide the compound terms by using models based on the syntactic structure. Thus, a chapter is devoted to the treatment of compound terms composed of a noun and an adjective and another is dedicated to the treatment of compound terms composed of tow nouns. The different structure of compound terms encourages us to approach them differently. To explain this, we note that the base element, in compound N1-A1, tends to associate with an adjective in a preferential or obligatory way. Thus, we use the concept of "polarization" to deal with the behavior of compound terms N1-A1 where the base element polarizes the associated adjectives. On the other hand, the compound terms N1-N2 undergoes a kind of dilution where the basic element can be replaced by a synonym without changing concept. Moreover, the compound N1-N2 can change its structure without changing its concept. So we use the term "dilution" to describe this behavior.On the other hand, we offer some analysis to understand the semantics of the genitive case regarding the compound terms in Arabic N1-N2. To do this, we replace the genitive case by prepositional phrases to clarify the semantics of the structure of annexation. This also allows making a semantic categorization for the genitive case.

No description available.

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第19107号 / 情博第553号 / 新制||情||98(附属図書館) / 32058 / 京都大学大学院情報学研究科知能情報学専攻 / (主査)教授 黒橋 禎夫, 教授 石田 亨, 教授 河原 達也 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

Idag finns det stora mängder användar-skapat data i form av texter från spelomdömen till åsikter i mikro-bloggar som Twitter. Att analysera detta data kan vara utav värde för både företag och akademisk forskning men är väldigt omfattande. Med hjälp av attitydanalysen kan detta utföras automatiskt och spara resurser, men vilka algoritmer presterar bäst? Med hjälp av en förstudie och ett par kvantitativa tester kunde dem mest populära tillvägagångsätten inom attitydanalysen genom att analysera spelomdömen från plattformen Steam. I testning har det visat sig att maskininlärningsalgoritmer både presterar bättre och är enklare att komma igång i jämförelse med lexikonbaserade algoritmer som knappast uppnår tröskelvärdet för pålitlighet vid klassifikation av omdömen som positiva eller negativa. Men det är fortfarande viktigt anpassa attitydanalysen för just det specifika problemet eftersom båda dessa har sina brister eftersom båda dessa tillvägagångsätt hade en dålig prestation i förhållande till sarkastiska omdömen. / Today there exist a huge amount of user created content in the shape of text from game reviews to opinions in microblogs like Twitter. To analyze this data could be of value for both companies and data scientists but remains to be very daunting. With the help of sentiment analysis this could be achieved automatically and save resources, but the question remains which algorithms have the best performance? With the help of a study and a couple of tests the most popular approaches in sentiment analysis could be compared by analyzing game reviews from the platform Steam. Through testing it has showed that machine learning based algorithms have the best performance and are easier to start with in comparison to lexicon-based approaches, which barely even reach the threshold for reliability in classifying reviews to be positive or negative. But it’s still important to plan and consider which algorithm one chooses for sentiment analysis as both approaches have their flaws and had a weak performance regarding sarcastic reviews.

The present study, a quantitative lexical analysis, examines the extent to which vocabulary size and lexical diversity contribute to writing scores on advanced non-native speakers’ and native speakers’ academic compositions. The data consists of essays composed by 104 adult non-native English learners enrolled in advanced second language writing courses and 68 native speaking university students in a first-year composition course. The lexical diversity of the sample essays is quantified by both the Measure of Textual Lexical Diversity (MTLD) and the voc-D while vocabulary size is measured by CELEX word frequency means, three instruments that are available in the computational linguistics program Coh-Metrix 3.0. Writing scores are provided by three independent raters’ evaluations according to the TOEFL iBT Independent Writing Rubric. Results from a binary logistic regression reveal that lexical diversity has a significantly greater impact on writing score than vocabulary size (p < .01). Nevertheless, a series of MANOVAs indicate that vocabulary size initially facilitates writing scores at the lower proficiency levels, but it is an essay’s lexical diversity that promotes it to the higher score levels. Additional findings from the MANOVAs demonstrate that native speakers’ profiles of lexical diversity and size are significantly different from their non-native peers (p < .001). The lexical profiles also differed significantly among the individual score levels of the TOEFL iBT rubric (p < .05). A final outcome from a Pearson’s product moment correlation analysis shows that iii vocabulary size has only a moderate relationship to lexical diversity, suggesting that variation of mid-range vocabulary may be more important to writing proficiency than the use of more sophisticated terms that occur less frequently in natural language. Implications for practice suggest that it is not enough to simply teach vocabulary words in the L2 composition classroom, but also to guide learners in how to employ these words in a varied manner within their writing. Furthermore, the results of this study indicate that teachers should spend more time on helping students use medium frequency words along with synonyms of a similar frequency rather than teaching students infrequent vocabulary, which may appear to sound more advanced.

There has been little research exploring whether different memory processes (i.e. related to short term memory (STM), working memory (WM) and long term memory (LTM) may be differentially sensitive to representation and processing aspects of compound words. This study investigated how compound words are represented in STM in immediate serial recall tasks and in WM in complex span tasks that combine processing and storage demands. The simple span STM task was comprised of solely a list of memory words, whereas the two complex span WM tasks interleaved sentence processing between presentation of memory words. They varied in the presence of a pause after presentation of each memory word and before onset of the following distractor sentence for processing. The absence of a pause was intended to minimize opportunity for subvocal rehearsal, whereas the presence of a pause encouraged rehearsal. To increase chances of recombination errors for error analyses, lists of memoranda were manipulated so that each set (list) of four compound words contained one “lure” pair (e.g. pinstripe + warhead = pinhead) in which the modifier and head constituents from separate compound words could recombine to form a new, legal word. Recall performance was better in the simple span and complex span pause tasks compared to the complex span no pause task. Whole compound and left constituent frequencies played opposite roles, helping and harming, respectively. Error types reflecting decomposition of the compound words to their constituents were more common in simple span than in complex span. Omissions were more common in complex span. We discuss how different memory processes may be differentially sensitive to representation and processing aspects, and how recall of compound words is affected by various lexical variables. / Thesis / Master of Science (MSc)

This thesis aims to analyse how poverty and social exclusion are visually represented through photographic images in the Greek mainstream media, following relevant articles in 2022. It contributes to the discourse on poverty using visual research methodologies, examining media-generated visuals that accompany articles on the same topic, addressing a gap in similar work within Greek media. Additionally, it seeks to offer insights into the professional use of these visuals. The interpretivist visual analysis is conducted under the ontological assumption of a realist and constructivist epistemological assumption. The analysis reveals that the visual lexicon of poverty in Greek mainstream news outlets in 2022 lacks variation and sustains stereotypical representations, with a focus on energy poverty and housing for which visuals are often inanimate. People are present in less than 50% of the visuals, and among them, figures of authority predominate, while in the rest they do not make direct eye contact with the viewer. Visuals involving refugees are staged by the producer, while stereotypical representations of poverty can be found in the one third of the sample and involve refugees in the most part. This research adopts a qualitative and inductive approach to describe, understand, and interpret the phenomenon. By examining the visual strategies employed by mainstream media outlets to represent poverty, the study aims to contribute to a better understanding of how these topics are constructed and portrayed in the Greek media. The findings hold significance for media practitioners, policymakers, and scholars interested in comprehending the role of media in shaping public perceptions of poverty. / Η μελέτη αυτή έχει ως στόχο να αναλύσει πώς η φτώχεια και ο κοινωνικός αποκλειστισμός αναπαρίστανται οπτικά μέσω φωτογραφικών εικόνων στα κυριότερα ΜΜΕ της Ελλάδας, ακολουθώντας σχετικά άρθρα του 2022. Συμβάλλει στο διάλογο για τη φτώχεια χρησιμοποιώντας μεθοδολογίες οπτικής έρευνας, εξετάζοντας τις οπτικές που παράγονται από τα μέσα που συνοδεύουν άρθρα με το ίδιο θέμα και αντιμετωπίζοντας ένα κενό σε παρόμοιες εργασίες όσον αφορά τα ελληνικά μέσα ενημέρωσης. Επιπλέον, επιδιώκει να ερευνήσει βαθύτερα την επαγγελματική χρήση αυτού του υλικού. Η ερμηνευτική οπτική ανάλυση διεξάγεται υπό την οντολογική υπόθεση μιας ρεαλιστικής και κονστρουκτιβιστικής επιστημολογικής υπόθεσης. Η ανάλυση αποκαλύπτει ότι το οπτικό λεξιλόγιο της φτώχειας στα κυριότερα ελληνικά ειδησεογραφικά μέσα το 2022 δεν διαθέτει ποικιλία και διατηρεί στερεότυπες αναπαραστάσεις, με έμφαση στην ενεργειακή φτώχεια και τη στέγαση, για τις οποίες οι φωτογραφίες συχνά είναι άψυχες. Οι άνθρωποι παρίστανται σε λιγότερο από 50% του υλικού, και μεταξύ αυτών, επικρατούν πρόσωπα εξουσίας, ενώ στις υπόλοιπες δεν έχουν άμεση επαφή με τον θεατή. Οι φωτογραφίες που αφορούν πρόσφυγες είναι σκηνοθετημένες από το δημιουργό τους, ενώ στο ένα τρίτο του δείγματος βρίσκονται στερεότυπες αναπαραστάσεις της φτώχειας που αφορούν κυρίως πρόσφυγες. Αυτή η έρευνα υιοθετεί μια ποιοτική και επαγωγική προσέγγιση για να περιγράψει, να κατανοήσει και να ερμηνεύσει το φαινόμενο. Εξετάζοντας τις οπτικές στρατηγικές που χρησιμοποιούν τα κυριότερα μέσα ενημέρωσης για να αναπαραστήσουν τη φτώχεια, η μελέτη αποσκοπεί στο να συμβάλει στην καλύτερη κατανόηση του τρόπου με τον οποίο αυτά τα θέματα κατασκευάζονται και απεικονίζονται στα ελληνικά μέσα ενημέρωσης. Τα ευρήματα είναι σημαντικά για τους επαγγελματίες των μέσων ενημέρωσης, τους πολιτικούς και τους ερευνητές που ενδιαφέρονται να κατανοήσουν το ρόλο των μέσων στη διαμόρφωση της κοινής γνώμης για τη φτώχεια.

The objectives of this research were to use trained sensory panels to identify key sensory profiling characteristics of freshwater prawns, develop and test price elicitation methods using the trained panels and gather freshwater prawn pricing information. Prawns were procured through a reputable producer and delivered to the Department of Food Science, Nutrition & Health Promotion at Mississippi State University. A trained sensory panel was utilized to develop descriptive terms to express key attributes of freshwater prawns. Trained panelists evaluated both raw and cooked product forms for 40 and 83 attributes, respectively. At the end of the sensory evaluation of the products, trained panelists also completed a willingness-to-pay ballot. Results determined sensory attributes for both raw and cooked prawns and the value of the raw product form.

No description available.

No description available.