Vocabulary is considered the building block of language. Without it, communication breaks down. This degree project investigates the vocabulary types, progression, lexical coverage andacademic words in EFL upper secondary textbooks in Sweden. This was done by a corpus-based approach using corpus tools such as Range and AntWord Profiler. The entire corpus contained six textbooks from two different series: Solid Gold and Viewpoints. These textbooks are utilised in English 5, 6, and 7 upper secondary courses. The analysis focused on Nation's word types and Coxhead’s Academic World List, including progression and coverage. The results revealed that high-frequency vocabulary is maintained throughout the series, while mid-frequency vocabulary considerably increases as the textbook levels progress. Results also show that around 3,000 word families are needed for 95% lexical coverage and 6,000 word families for 98% lexical coverage. Each textbook increases around 1,000 word families per level, but is less proportional in relation to the word count. The highest increase of word families is 32% and the lowest of 17%. Furthermore, the study found that the AWL covers 7% of the vocabulary, and 80% of the AWL items are used. Pedagogical recommendations from this study include establishing a vocabulary input threshold per level to ensure consistent progression. Additionally, academic written genres such as Introduction-Method-Results-Discussion (IMRaD) structures can be included in the textbooks in addition to the existing literary genres to help prepare Swedish L2 learners to cope with university academic texts.

We explore in this study the different Natural Language Processing state-of-the art technologies that allow transforming textual data into numerical representation. We go through the theory of the existing traditional methods as well as the most recent ones. This thesis focuses on the recent advances in Natural Language processing being developed upon the Transfer model. One of the most relevant innovations was the release of a deep bidirectional encoder called BERT that broke several state of the art results. BERT utilises Transfer Learning to improve modelling language dependencies in text. BERT is used for several different languages, other specialized model were released like the french BERT: Camembert. This thesis compares the language models of these different pre-trained models and compares their capability to insure a domain adaptation. Using the multilingual and the french pre-trained version of BERT and a dataset from AXA France’s emails, clients’ messages, legal documents, insurance documents containing over 60 million words. We fine-tuned the language models in order to adapt them on the Axa insurance’s french context to create a French AXAInsurance BERT model. We evaluate the performance of this model on the capability of the language model of predicting a masked token based on the context. BERT proves to perform better : modelling better the french AXA’s insurance text without finetuning than Camembert. However, with this small amount of data, Camembert is more capable of adaptation to this specific domain of insurance. / I denna studie undersöker vi de senaste teknologierna för Natural Language Processing, som gör det möjligt att omvandla textdata till numerisk representation. Vi går igenom teorin om befintliga traditionella metoder såväl som de senaste. Denna avhandling fokuserar på de senaste framstegen inom bearbetning av naturliga språk som utvecklats med hjälp av överföringsmodellen. En av de mest relevanta innovationerna var lanseringen av en djup dubbelriktad kodare som heter BERT som bröt flera toppmoderna resultat. BERT använder Transfer Learning för att förbättra modelleringsspråkberoenden i text. BERT används för flera olika språk, andra specialmodeller släpptes som den franska BERT: Camembert. Denna avhandling jämför språkmodellerna för dessa olika förutbildade modeller och jämför deras förmåga att säkerställa en domänanpassning. Med den flerspråkiga och franska förutbildade versionen av BERT och en dataset från AXA Frankrikes epostmeddelanden, kundmeddelanden, juridiska dokument, försäkringsdokument som innehåller över 60 miljoner ord. Vi finjusterade språkmodellerna för att anpassa dem till Axas försäkrings franska sammanhang för att skapa en fransk AXAInsurance BERT-modell. Vi utvärderar prestandan för denna modell på förmågan hos språkmodellen att förutsäga en maskerad token baserat på sammanhanget. BERTpresterar bättre: modellerar bättre den franska AXA-försäkringstexten utan finjustering än Camembert. Men med denna lilla mängd data är Camembert mer kapabel att anpassa sig till denna specifika försäkringsdomän.

No description available.

No description available.

1 Solving strategies of word problems depending on the context topicality ABSTRACT The thesis deals with word problems solvable by linear equations. The aim of the work is to show if the context, on which the word problem depends, influences how pupils participating in the research solve it and verify if word problem phrasing and modern language usage help pupils solve the word problem. It is accomplished by assigning pupils a set of varied word problems and assessing their solutions. More results were gained by assessing questionnaires related to the context of the word problems being calculated. The second aim of the thesis is to explain various definitions of word problems, putting the term into the context of school mathematics and a brief summary of the historical development of mathematics focusing on word problems. The work consists of three parts. In the first part, problems and word problems are defined and it is shown how word problems are included in the Framework educational programme. Word problems are put into historical context. The second part shows the conception of motion word problems, word problems on joint work and word problems on dividing a whole into parts in the present-day mathematical textbooks in the second stage of elementary schools and in the lower classes of secondary grammar...

CoNLL 2015 featured a shared task on shallow discourse parsing. In 2016, the efforts continued with an increasing focus on sense classification. In the case of implicit sense classification, there was an interesting mix of traditional and modern machine learning classifiers using word representation models. In this thesis, we explore the performance of a number of these models, and investigate how they perform using a variety of word representation models. We show that there are large performance differences between word representation models for certain machine learning classifiers, while others are more robust to the choice of word representation model. We also show that with the right choice of word representation model, simple and traditional machine learning classifiers can reach competitive scores even when compared with modern neural network approaches.

This master thesis deals with the construction of a system for automatic classification of~unknown words based on derivation bonds. For this purpose, the system was designed to~extract derivative links based on electronic dictionaries and to create word-forming models from them. Based on this knowledge, it is then possible to incorporate unclassified words into existing nests formed from the obtained bonds, and their models, or create new ones. The reader will be gradually acquainted with the reasons that lead to the continuous transformation or expansion of the lexicon, the ways in which the words in~the~Czech language are derived and how to obtain information about the changes caused by this derivation process. This system builds on and extends the research of the branch of morphology in~the~project of a morphological analyzer of the Research Group of Knowledge Technologies, working at the Faculty of Information Technology of the Brno University of~Technology.

Thesis (B. Tech.) - Central University of Technology, Free State, 2007

E-commerce is today a very attractive market with growth prospects and opportunities for new companies to establish themselves and start operations. Since the establishment of Internet and the evolution of how Internet is used today, there has been an increasing possibility for consumer’s to buy products online. The more expanding E-commerce business today makes companies and entrepenuers eager to establish markets online. There is though a opportunity with the expanding of E-commerce today and that is for companies to investigate how consumer’s values the risks of shopping online. This opportunity is something that companies should take in care before starting an E-commerce market. This study aims to identify consumers' perceived risk in purchasing decisions in relation to E-commerce. The study is intended to give a broader perspective to be helpful for companies who are in business with e-commerce or supposed to conduct and begin sales activities through Ecommerce. This will be achieved by showing which factors consumers find to be most affecting to their uncertainty and risk when buying online. The survey methodology used in this study was a quantitative method where we deduced hypotheses to ensure on previous research and theories that the results of our investigation was consistent with what we concuded in our study. To come up with the answers we received, we used questionnaires in which respondents were asked to answer on a seven-point scale how they felt the risk of shopping online today. The results of this study shows us that consumer’s perceived risk will grow if the purchase is of an high engagement product, we can also conclude that there will be a decline in perceived risk if the websites have good design, user friendly navigation, clear information about consumer’s warranty and right of return.

<p>In recent years, more and more people have started to live double lives. Blogs, Facebook, YouTube and Twitter have created conditions so that we increasingly live virtual lives on the Internet. The gathering of people in this way entails both risks and opportunities for businesses. This study aims to explain how the phenomenon of social media has affected trade and service companies in a marketing and communication perspective. The idea for the essay came to us after we had observed how different behaviors in companies’ concerning social media could vary a lot, and how we felt that this affected our personal view of the brands. Through a qualitative approach, where we interviewed six people with different backgrounds, we have examined the subject in order to explain how companies can use the characteristics of social media; conversation and interaction. Our results are based on that the social media entails both opportunities and challenges for companies, which are presented in detail in Chapter five.</p>