• Digital Humanities and Cultural Heritage
• Publicationsclose
• Other research productsclose
• 2019-2023close
• Open Access close
• Danish close
• Roskilde Universitetscenter's Digitale Arkiv close

This study explores the use of Natural Language Processing as a tool for online reviews. By drawing on theories about machine learning, this article investigates how to develop a Natural Language Processing model in Python. The authors present two machine learning models which were developed through an iterative design process and by implementing various approaches from the TRIN-model. This study concludes that machine learning models like ours might not have a significant impact on the future of reviews, but we are optimistic about the potential of Natural Language Processing as a tool for categorizing specific aspects of a company’s services. This conclusion is based on the analysis and discussion of the results we gathered from our Natural Language Processing models.

This paper examines how Russian President Vladimir Putin incorporates the use of history in his speeches and articles with regards to how he positions Russia in its relation to Ukraine. The analysis is structured around three central places of remembrance (danish: erindringssteder): The Kyivan Rus which focuses on the close historical relation between Ukraine and Russia; the heritage of the Sovietunion in relation to how the union defined the borders of Soviet-Ukraine; World War II, ukrainian nationalism and its relation to nazism which centers around how Putin relates nazism to the current ukrainian political elite. The analysis concludes that Putin primarily utilizes the three places of remembrance to legitimize Russia's current invasion of Ukraine. Putin finds the distribution of territories during the soviet era to have been theft, and a complete violation of Russia's integrity. Furthermore, he seeks to protect ethnic russians within the borders of Ukraine from a genocide, instigated by ukrainian nationalists and neo-nazis, who continue the tradition of atrocities commited during World War II. Finally, Putin perceives Ukrainians and Russians as a single people, basing his claim on common history, language, and culture. Thus he implies that ukrainians should unite under Russia, as Russia is the more legitimate state.

This paper revolves around the development of an LSTM multiclass classifier, constructed using Keras as framework and CRISP-DM as project process, with the purpose of classifying natural language into varying degrees of toxicity. The model takes a starting point in an existing toxic comment classification challenge from Kaggle.com, and makes a first iteration, engineered towards the requirements in the challenge. In this first iteration, several measures are taken to avoid common pitfalls of neural networks. The model is then held up against principles of freedom of speech including The Harm Principle and The Offence Principle by John Stuart Mill and Joel Feinberg respectively. After evaluating upon the models performance in the light of these principles, a second iteration is constructed with some design changes. For reasons i.a. related to the dataset, this operation is less successful. The paper concludes that it is possible to make a good multiclassification tool for shallow NLP problem, but gets less efficient in later iterations as we try to apply it to more concrete purposes.