Introduction: Ted Underwood tests a new language representation model called “Bidirectional Encoder Representations from Transformers” (BERT) and asks if humanists should use it. Due to its high degree of difficulty and its limited success (e.g. in questions of genre detection) he concludes, that this approach will be important in the future but it’s nothing to deal with for humanists at the moment. An important caveat worth reading.
Introduction: Named Entity Recognition (NER) is used to identify textual elements that gives things a name. In this study, four different NER tools are evaluated using a corpus of modern and classic fantasy or science fiction novels. Since NER tools have been created for the news domain, it is interesting to see how they perform in a totally different domain. The article comes with a very detailed methodological part and the accompanying dataset is also made available.
Introduction: Studying n-grams of characters is today a classical choice in authorship attribution. If some discussion about the optimal length of these n-grams have been made, we have still have few clues about which specific type of n-grams are the most helpful in the process of efficiently identifying the author of a text. This paper partly fills that gap, by showing that most of the information gained from studying n-grams of characters comes from the affixes and punctuation.