Improving classification of low-resource COVID-19 literature by using Named Entity Recognition

Automatic document classification for highly interrelated classes is a demanding task that becomes more challenging when there is little labeled data for training. Such is the case of the coronavirus disease 2019 (COVID-19) Clinical repository—a repository of classified and translated academic articles related to COVID-19 and relevant to the clinical practice—where a 3-way classification scheme is being applied to COVID-19 literature. During the 7th Biomedical Linked Annotation Hackathon (BLAH7) hackathon, we performed experiments to explore the use of named-entity-recognition (NER) to improve the classification. We processed the literature with OntoGene’s Biomedical Entity Recogniser (OGER) and used the resulting identified Named Entities (NE) and their links to major biological databases as extra input features for the classifier. We compared the results with a baseline model without the OGER extracted features. In these proof-of-concept experiments, we observed a clear gain on COVID-19 literature classification. In particular, NE’s origin was useful to classify document types and NE’s type for clinical specialties. Due to the limitations of the small dataset, we can only conclude that our results suggests that NER would benefit this classification task. In order to accurately estimate this benefit, further experiments with a larger dataset would be needed.

An n-gram based approach to the automatic classification of schoolchildren’s writing

This article focuses on the analysis of schoolchildren’s writing (throughout the whole primary school period) using sets of morphological labels (n-grams). We analyzed the sets of bigrams and trigrams from a group of literary texts written by Catalan schoolchildren in order to identify which bigrams and trigrams can help discriminate between texts from the three cycles into which the Spanish primary education system is divided: lower cycle (6- and 7-year-olds), middle cycle (8- and 9-year- olds) and upper cycle (10- and 11-year-olds). The results obtained are close to 70% of correct classifications (77.5% bigrams and 68.6% trigrams), making this technique useful for automatic document classification by age.

Automatic classification of journalistic documents on the Internet1

Online journalism is increasing every day. There are many news agencies, newspapers, and magazines using digital publication in the global network. Documents published online are available to users, who use search engines to find them. In order to deliver documents that are relevant to the search, they must be indexed and classified. Due to the vast number of documents published online every day, a lot of research has been carried out to find ways to facilitate automatic document classification. The objective of the present study is to describe an experimental approach for the automatic classification of journalistic documents published on the Internet using the Vector Space Model for document representation. The model was tested based on a real journalism database, using algorithms that have been widely reported in the literature. This article also describes the metrics used to assess the performance of these algorithms and their required configurations. The results obtained show the efficiency of the method used and justify further research to find ways to facilitate the automatic classification of documents.