Impact of Negation and AnA-Words on Overall Sentiment Value of the Text Written in the Bosnian Language

In this manuscript, we present our efforts to develop an accurate sentiment analysis model for Bosnian-language tweets which incorporated three elements: negation cues, AnA-words (referring maximizers, boosters, approximators, relative intensifiers, diminishers, and minimizers), sentiment-labeled words from a lexicon. We used several machine-learning techniques, including SVM, Naive Bayes, RF, CNN, with different input parameters, such as batch size, number of convolution layers, type layers. addition these BOSentiment is provide initial value each tweet, then CNN. Our best-performing model, combined CNN 256 filters size 4×4, 10, achieved accuracy over 92%. results demonstrate the effectiveness approach in accurately classifying Bosnian using lexicons, pre-trained models. This study makes significant contribution field under-researched languages Bosnian, could be extended other social media platforms gain insight into public opinion.