application/pdf 146642 bytes Abstract: In this paper, the sentiment dictionary is presented as one of the lexical resources necessary for the sentiment analysis. The use of this resource is related to the existence of other specific resources that are adapted to the ultimate goal - in this case, sentiment analysis. Thus, in order to normalize documents for sentiment analysis, a specific stop word dictionary is needed (for example, it does not contain negation signals), then a stemmer or a lemmatizer to transform the word to the basic form, to which the words from the sentence must be reduced. Informal writing on social networks leads to the impossibility of applying standardized lexical resources. Therefore, in order to improve the results parallel lexical resources are made. 5th International Conference „Contemporary Problems of Mathematics, Mechanics and Informatics“ Towards the lexical resources for sentiment-reach informal texts – the Serbian language case 2018 https://phaidrabg.bg.ac.rs/o:30275 Marovac, Ulfeta Ljajić, Adela Kajan, Ejub Avdić, Aldina Keywords: sentiment analysis, social media, n-grams, lexicon All rights reserved info:eu-repo/semantics/conferenceProceedings eng