Speaker
Iulia-Maria Florea FLOREA
(Politehnica University of Bucharest)
Description
Recent work in machine learning has provided promising results regarding short text classification. However, in case of technical abstracts and datasets, there are some cases when unsupervised algorithms fail to reach the optimal results. This paper presents a detailed description of a text classification system for short technical texts and datasets from a single domain, together with experiments. In order to be effective, it has been built as a mix between machine learning algorithms and keywords determined by an expert’s point of view.
Authors
Iulia-Maria Florea FLOREA
(Politehnica University of Bucharest)
Răzvan Victor RUGHINIȘ
(University Politehnica of Bucharest)
Stefan-Dan CIOCIRLAN
(Politehnica University of Bucharest)
