In this article, we propose a method for the automatic retrieval of a set of semantic primitive words from an explanatory dictionary and a novel evaluation procedure for the obtained set of primitives. The approach is based on the representation of the dictionary as a directed graph with a single-objective constrained optimization problem via a genetic algorithm with the PageRank scoring model. The problem is defined as a subset selection. The algorithm is fit to search for the sets of words that should fulfil several requirements: the cardinality of the set should not exceed empirically selected limits and the PageRank word importance score is minimized with cycle prevention thresholding. In the experiments, we used the WordNet dictionary for English. The proposed method is an improvement over the previous state-of-the-art solutions.
In computational argument mining, the final constructed argument graph is usually the center of attention, with researchers analyzing its structure and developing methods to extract valuable insights from it. In this paper, we argue that analyzing only the final argument graph is not enough. We demonstrate that the same argument graphs do not imply the same argument construction styles (ways of populating them), which are an important abstract argumentation concept. Therefore, both final structure and the construction style should be considered when analyzing arguments. To address the analysis of argument construction styles, we present the DRACS: Diachronic Representation of Argument Construction Styles framework to capture argument construction style features and enable their analysis alongside the graph structures. We evaluate our methodology on three distinct datasets of pre-defined similar and differing construction styles: the AIF-labeled corpora UNSC and US2016, as well as a custom corpus of synthetic tree-based argument graphs. We apply our method to highlight similarities and differences of argument construction styles on US2016 datasets. Our findings suggest that even if the final argument graphs display a high degree of similarity, their construction styles could differ significantly.