Comparative Analysis of the Main SaaS Algorithms for
Named Entity Recognition Applied for Romanian Language

Bogdan IANCU
The Academy of Economic Studies,
6 Piața Romană, 010374 Bucharest, Romania

Abstract: This paper proposes a comparative analysis of the main Name Entity Recognition algorithms available in cloud, applied for texts written in Romanian. The context of this analysis is the one of the semantic web, where the problem of identifying new entities and linking them to existing ontologies persists. There are processes defined that allow the text written in Romanian to be translated in one of the languages supported by the algorithms provided by DBpedia (DBpedia Spotlight), Google (Google Cloud Natural Language API), Microsoft (the NER module from Azure Machine Learning Studio) and IBM (IBM Watson Natural Language Understanding), and afterwards the F1 score is computed in order to identify the optimal process. The article ends with a comparison between the obtained results and the performance achieved by NER algorithms specialized for
English or language independent.

Keywords: Semantic web, NER, LOD, SaaS.

View full article