Microblogging has become one of the prevalent sources of communication. A large volume of data is available online in the form of text across the digital world on social media websites, blogs, news, articles, digital libraries, comments and reviews. The aim is to organize and manage data and obtain valuable information from it. Data volume has grown exponentially over the past few years and it has become impossible for humans to retrieve the required information easily. For this purpose, various text mining techniques have been proposed and used including sentiment analysis. Websites have become a rich source of data and the increasing use of acronyms in text has emerged as a new challenge which affects the results of the sentiment analysis. The case worsens in the informal text like opinion and reviews. This paper presents a novel idea of enhancing the accuracy of sentiment analysis by dealing with acronyms in a particular domain. The proposed system was used for many domains and it shows significant improvement in overall accuracy of the text analysis. The proposed system was evaluated using metrics of information retrieval, i.e., Precision, Recall, and F-Measure. The experimental results show that the system achieves 57% Precision, 78% Recall and 64% F-Measure before mapping the long form of the acronyms. Moreover, the system performance is increased by achieving 89% Precision, 91% Recall and 90% F-Measure after mapping the long-form of acronym into the text.