• Digital Humanities and Cultural Heritage
• Publicationsclose
• 2019-2023close
• Digitala Vetenskapliga Arkivet - Academic Archive On-line close

Macroeconomic forecasting is a classic problem, today most often modeled using time series analysis. Few attempts have been made using machine learning methods, and even fewer incorporating unconventional data, such as that from social media. In this thesis, a Generative Adversarial Network (GAN) is used to predict U.S. unemployment, beating the ARIMA benchmark on all horizons. Furthermore, attempts at using Twitter data and the Natural Language Processing (NLP) model DistilBERT are performed. While these attempts do not beat the benchmark, they do show promising results with predictive power. The models are also tested at predicting the U.S. stock index S&P 500. For these models, the Twitter data does improve the accuracy and shows the potential of social media data when predicting a more erratic index with less seasonality that is more responsive to current trends in public discourse. The results also show that Twitter data can be used to predict trends in both unemployment and the S&P 500 index. This sets the stage for further research into NLP-GAN models for macroeconomic predictions using social media data. Makroekonomiska prognoser är sedan länge en svår utmaning. Idag löses de oftast med tidsserieanalys och få försök har gjorts med maskininlärning. I denna uppsats används ett generativt motstridande nätverk (GAN) för att förutspå amerikansk arbetslöshet, med resultat som slår samtliga riktmärken satta av en ARIMA. Ett försök görs också till att använda data från Twitter och den datorlingvistiska (NLP) modellen DistilBERT. Dessa modeller slår inte riktmärkena men visar lovande resultat. Modellerna testas vidare på det amerikanska börsindexet S&P 500. För dessa modeller förbättrade Twitterdata resultaten vilket visar på den potential data från sociala medier har när de appliceras på mer oregelbunda index, utan tydligt säsongsberoende och som är mer känsliga för trender i det offentliga samtalet. Resultaten visar på att Twitterdata kan användas för att hitta trender i både amerikansk arbetslöshet och S&P 500 indexet. Detta lägger grunden för fortsatt forskning inom NLP-GAN modeller för makroekonomiska prognoser baserade på data från sociala medier.

AbstractThousands of Roma were killed in Ukraine by the Nazis and auxiliary police on the spot. There are more than 50,000 Roma in today’s Ukraine, represented by second and third generation decendants of the genocide survivors. The discussion on Roma identity cannot be isolated from the memory of the genocide, which makes the struggle over the past a reflexive landmark that mobilizes the Roma movement. About twenty Roma genocide memorials have been erected in Ukraine during last decade, and in 2016 the national memorial of the Roma genocide was opened in Babi Yar. However, scholars do not have a clear picture of memory narratives and memory practices of the Roma genocide in Ukraine. A comprehensive analysis of the contemporary situation is not possible without an examination of the history and memory of the Roma genocide before 1991.

The purpose of this thesis is to investigate different scales of mobility and social networks in Southern Scandinavia around 7000 BC. An essential basis for this work is the hypothesis that the conical core pressure blade technology, a specific method for producing lithic blades, reached Scandinavia from the east. The process has been discussed and verified in several earlier studies. The study is based on investigations of lithic material and bone tools, as well as human skeletal remains; a multidisciplinary strategy has been applied, which combines technological analyzes of archaeological material with isotope analyzes of human teeth. Materials and data have been collected from a total of 111 sites located mainly in Scandinavia, but also in Finland, the Baltic countries and Russia. The Mesolithic site Norje Sunnansund in Blekinge, southern Sweden is in focus. Analyzes of the chaîne opératoire of lithic and osseous production make it possible to study technological processes as the transmission of culturally conditioned patterns; these patterns underlie the formative principles of each technology complex. By identifying specific traditional knowledge built into the material process, it becomes possible to define prehistoric human traditions and thereby study human interactions and migrations between geographical regions. The analyzes of strontium isotopes in human teeth from Norje Sunnansund enables an additional individual provenancing since the isotopes reflect a geographical-geological variation. The study provides a higher resolution of the arrival and spreading of the pressure blade technology from the northeast. It also strengthens the picture of a Mesolithic society in Southern Scandinavia based on geographically extensive social networks. An increased regionalization and territorialization has previously been proposed for western Scandinavia around 8800-7500 BC. The study confirms that during this period the population seem to consist of several subgroups related to specific territories, but with close and regular contact within a more extensive social network. This fits well with the picture of a group that may have inhabited Norje Sunnansund for extended periods, while social contacts were maintained over longer distances.

The use cases of Information Extraction (IE) are more or less endless, often consisting of a combination of Named Entity Recognition (NER) and Relation Extraction (RE). One use case of IE is the extraction of transaction information from Norwegian insider transaction Press Releases (PRs), where a transaction consists of at most four entities: the name of the owner performing the transaction, the number of shares transferred, the transaction date, and the price of the shares bought or sold. The relationships between the entities define which entity belongs to which transaction, and whether shares were bought or sold. This report has investigated how a pair of supervised NER and RE models extract this information. Since these Norwegian PRs were not labeled, two different approaches to annotating the transaction entities and their associated relations were investigated, and it was found that it is better to annotate only entities that occur in a relation than annotating all occurrences. Furthermore, the number of PRs needed to achieve a satisfactory result in the IE pipeline was investigated. The study shows that training with about 400 PRs is sufficient for the results to converge, at around 0.85 in F1-score. Finally, the report shows that there is not much difference between a complex RE model and a simple rule-based approach, when applied on the studied corpus.

‘Public health’ investigates the determinants of health, born during the Enlightenment in the seventeenth/eighteenth century. But ‘public health’ is also policies, aiming at the improvement of a population’s health. There is a mutual interchange between public health as science and as politics. A brief historical background is followed by an analysis of the impacts of political changes during the first two decades of the twenty first century in Sweden. In 2005, a policy document accepted by all political parties except for the Moderate Party highlighted socio-economic factors and structural reforms to decrease the health gaps in the population. The general election in September 2006 resulted in a new majority in the parliament and a center-right coalition government, including the Moderates and three parties that had approved of the 2005 document. In 2007 a “new public health policy” was introduced. Its priority lists stressed individual behavior and the new policy should be incentives to work instead of “allowances”. The Public Health Institute got instructions in accordance with the new policy. The ten years following this policy change has seen public health policies and attitudes to research shifting almost year by year. The new policy met a counter-stream from the very beginning. Influenced by Michael Marmot’s WHO Commission on health inequalities, regional commissions started in Sweden, Recommendations how to decrease social health gaps was adopted with almost no opposition by regional health boards in 2012–2013. But new problems were now occupying politicians and media—how to finance the growth of the old, multi-sick part of the population and increasing costs for new medical technologies and drugs. Public health as an academic discipline was in the middle of this fluctuating political landscape with direct effects on what has been considered worth listening to or support by public money.

During a spectacular excavation in 1921 at the Norse farm of Herjolfsnes (Ikigaat) on the southern tip of Greenland, Poul Nørlund found 58 wooden crosses of drift­wood in the graves at the site. These vary in size from c. 10 to 70 cm. Since then, more crosses have been found in other churchyards, as well as a few in a more “profane” context in dwellings. Nearly all of these crosses are quite simple. But six of them are more elaborately carved with specific traits, which enable closer comparison with prototypes from Europe. Four crosses have Doric capital ends, which must be derived from the design of German and English crosses dating to the beginning of the 11th century and spread to Scandinavia in the 12th century. A regular crucifixion group (Calvary) has English and Norwegian antecedents dating to the mid-13th century, and a panel crucifix displays elements from a period as long c. 1200–1350, suggesting extreme lateness in style. There is nothing to stop us assuming that dissemination of influences essentially occurred through Nor­way and perhaps Iceland. Several stylistic traits, such as the Doric capitals, acanthus leaf and classical drapery, can be traced all the way back to classical anti­quity and represent their earliest occurrence in the western hemisphere. https://doi.org/10.33063/diva-429323

This thesis explores the phenomenon of intramural infant burial during Middle Helladic III–Late Helladic II during the Aegean Bronze Age. Intramural graves of children aged two years or less at Málthi and Ayios Stephanos, two settlements on the Greek mainland, are studied from a number of perspectives: the physical properties of the graves and the buried infants, the spatial and intramural context of the grave, and their relationship to their social and societal context. Some things found to be relevant in the analysis were the importance of kinship and group belonging, as well as shifting funerary practices in a time of large-scale socio-economic change in the region. Denna kandidatuppsats utforskar fenomenet intramurala spädbarnsgravar under Mellanhelladisk III–Senhelladisk II under den egeiska bronsåldern. Intramurala gravar av barn som var två år gamla eller yngre vid Málthi och Ayios Stephanos, två boplatser på det grekiska fastlandet, studeras ur ett antal perspektiv: gravarna och de begravda spädbarnens fysiska egenskaper, gravens rumsliga och intramurala kontext, och dess relation till dess sociala och samhälleliga kontext. Några ting som visade sig vara relevanta i analysen var vikten av släktskap och grupptillhörighet, såväl som föränderliga begravningsskick under en tid med storskaliga socioekonomiska förändringar i regionen.

Named entity recognition (NER) is a task that concerns detecting and categorising certain information in text. A promising approach for NER that recently has emerged is fine-tuning Transformer-based language models for this specific task. However, these models may require a relatively large quantity of labelled data to perform well. This can limit NER models applicability in real-world applications as manual annotation often is costly and time-consuming. In this thesis, we investigate the learning curve of human annotation and of a NER model during a semi-supervised bootstrapping process. Special emphasis is given the dependence of the number of classes and the amount of training data used in the process. We first annotate a set of collected job advertisements and then apply bootstrapping using both annotated and unannotated data and continuously fine-tune a pre-trained Swedish BERT model. The initial class system is simplified during the bootstrapping process according to model performance and inter-annotator agreement. The model performance increased as the training set grew larger with a final micro F1-score of 54%. This result provides a good baseline, and we point out several improvements that can be made to further enhance performance. We further identify classes handled differently by the annotators and potential factors as to why this is. Suggestions for future work include adjusting the current class system further by removing classes that were identified as low-performing in this thesis. Namngiven entitetsigenkänning (eng. named entity recognition) innebär att identifiera och kategorisera nyckelord i text. En ny lovande teknik för identifiering av namngivna enheter är att finjustera Transformerbaserade språkmodeller för denna specifika uppgift. Dessa modeller kräver dock stora mängder märkt data för att prestera väl. Detta kan begränsa antal områden i vilka de kan användas då manuell märkning av data ofta är kostsamt och tidskrävande. I denna avhandling undersöker vi inlärningskurvan för manuell annotering och för en språkmodell under en halvövervakad bootstrapping process. Särskild vikt läggs på hur modellens och annoterarnas inlärning påverkas av antal klasser och mängden träningsdata som används i processen. Vi annoterar först en samling jobbannonser och tillämpar sedan en bootstrapping process med både märkt och omärkt data i vilken en förtränad svensk BERT-modell kontinuerligt finjusteras. Det första klasssystemet förenklas under processens gång beroende på modellprestation och interannoterar-överenskommelse. Modellen presterade bättre med mer träningsdata och uppnådde en slutlig micro F1-score på 54%. Detta resultat ger en bra baslinje, och vi föreslår flera förbättringar som kan göras för att ytterligare förbättra modellprestationen. Vidare identifierar vi även klasser som hanteras olika av annoterare och potentiella faktorer till vad detta beror på. Förslag för framtida arbete inkluderar att justera det nuvarande klasssystemet ytterligare genom att ta bort klasser som identifierades som lågpresterande i denna avhandling.

Osteoarthritis is one of the most common general diseases in modern society Sweden. It is also one of the most common diseases traced within archaeological human remains. The aetiology of osteoarthritis has been widely debated within the field of medicine and paleopathology. The initial claim that the degenerative disease is caused by activity and ageing has been questioned and factors such as environment, diet and genetic markers has been brought up and to some extent studied. Even though osteoarthritis being one of the most common diseases recognized among archaeological human remains, it has during recent years often been neglected within the field and referred to mostly in different palaeopathological atlases. By studying human remains from the medieval churchyard of St: Hans in Visby, Gotland, the notion of osteoarthritis as a general disease in the past is discussed in the following study. The concept of a medieval “general public” is examined by looking at social strata through grave placement on the studied graveyard.