<![CDATA[The SDGs (Sustainable Development Goals) success era was marked by the openness of the government and other stakeholders to implement open and aspirational governance. Through opinion mining, the government, business actors, and other stakeholders can find out the current state of their output achievements to the public. Public opinion on social media is a tool for measuring, evaluating, and planning the success of a product, program and policy on a particular subject. However, quite a lot of opinion domains are shared by the public on social media. The existing opinion mining system only analyzes opinions on one particular opinion domain. This study builds a system that can help users to find out public perceptions about products, figures, and current topics in many analysis domains. Research uses big data on Twitter as a very popular source of opinion data to monitor public opinion. A review of previous related research concludes that the Naive Bayes algorithm has advantages in its simple computation, optimal for clustering few classes, and effective in classifying noisy features. The researcher found that the Naive Bayes algorithm in the Supervised Learning method was quite good in classifying multi-domain data. The implementation of the research resulted in a multi-domain sentiment analysis system. From the results of testing and evaluation, it is concluded that the system is able to identify opinion sentiments and provide a framework to accommodate analysis of the diversity of opinion domains from Twitter social media]]>
