<![CDATA[Motor vehicle users experience rapid growth every year. The increasing number of vehicles contributes to one of the state revenues: taxes. SAMSAT is a state institution with the authority to regulate motor vehicle tax (PKB). As technology develops, SAMSAT innovates through the SIGNAL application, which allows people to make motor vehicle tax payments safely via cell phone. Social media such as Instagram and X have great potential for collecting data to understand public reactions to the SIGNAL application. Comments on social media regarding the SIGNAL application raise pros and cons from the public; therefore, it is necessary to carry out sentiment analysis through a text mining approach using the Support Vector Machine (SVM) algorithm following the Cross Industry Standard Process for Data Mining (CRISP-DM) methodology. This research was carried out through several stages: data collection, preprocessing, modeling with the Support Vector Machine (SVM), and evaluation with a confusion matrix. Data in the research were collected from Instagram social media comments from September 20, 2023, until. 16 April 2024 as many as 3,543 records and 1,335 comments on X's social media from 31 May 2023 until March 27, 2024, with the keyword "SIGNAL application". After the preprocessing stage, the data used was reduced to 3,911 because there were duplicate and irrelevant reviews. based on 3,911 data, it produced 773 positive comments, 1991 negative, and 1147 neutral comments. This research aims to identify public sentiment towards SIGNAL services via social media, such as Instagram. We prepared a dataset of two and three sentiment classes for research modeling needs. Based on the application of the model, a Support Vector Machine (SVM) with a linear kernel produces better scores than the Naïve Bayes and KNN models with accuracy values of 0.88, precision of 0.88, recall of 0.81, and AUC of 0.92 using a 10-fold cross-validation on training data and test data.]]>
