Application of Natural Language Processing to Search for Business Opportunities for the raw Glycerol in Biodiesel Production

Resumo

Recently, in Brazil, the level of biodiesel added to oil diesel fuel has expanded because of compulsory rules. The
increase of biodiesel production has provided an excess of crude glycerol which is its primary co-product. This
paper intends to search business opportunities for the excess of crude glycerol by a Natural Language Processing
Approach. This process of exploring and analyzing unstructured data was performed by text mining thousands of
recovered abstracts from the scientific literature available on online research platforms over ten years. The
methodology involved the information retrieval and compilation of a database for the textual content analysis and
the use of word embedding concepts including Word2vec and Bert. The results indicated the tendencies for
producing value-added substances from crude glycerol over ten years of research.