摘要：The analysis of emotions is still considered a rather difficult procedure and it is constrained by the limitations related to the languages available for this analysis. The Romanian language is one of the richest and most complex languages with different dialects, regionalisms, archaisms and a powerful and expressive poetic language. Starting from the original vision of the poets, the expressiveness of the Romanian language has given birth to special metaphors that often lose their meaning if one tries to translate them. For this reason, the need arose to develop a lexicon specific to the Romanian language that would allow the analysis of emotions from different poems and to successfully identify the emotions presented by an author through the artistic images created in his work. This paper presents the first steps in the development of a Romanian lexicon based on methodologies employed in other languages, mainly the English language. The data set was subject to a filtering and refining process after which it was enriched by using specific tailored web crawlers for adding all the forms and conjugations of the respective words.
其他关键词：Lexicon, Natural language processing, Emotion analysis, Crawlers, INTELLIT, Romanian language.
