Abstract
RKUHP (Rancangan undang-undang kitab hukum pidana) renewal of criminal law was making a huge controversy because considered to have an over criminalization value. The critics were mostly given in microblog social media. This research will be done by using the data collected from the social media users in Indonesia about the given topics to retrieve an information about the the sentiment of Indonesian people towards RKUHP. The purpose of this research was to make a classification model which will classify the sentiment from the collected data into three classes : positive, neutral, and negative. This research will also aim to evaluate the performance of the model. The data that will be used in this research is a tweet from the period of september to november 2019. The labelling process in this research was done by crowdsourcing method. In which the majority result of the label will be set as the label. All the labelled data will be weighted using TF-IDF and sentiment dictionary and later on the weighted matrix will be used to build a machine learning model. The evaluation result of the machine using cross validation with the value of K equals to 10 using mean approach shows that the model reaches the highest accuracy with 95% using radial basis function kernel, C=1000 and gamma=0.0001.
Keywords: sentiment analysis, support vector machine, prediction, classification