Sentiment Analysis of Public Opinion Against Road Infrastructure Development in Manado City Using K-Means Clustering Algorithm

Infrastructure development attempts to enhance community quality of life while also facilitating the distribution of products and services. Constructing road infrastructure elicits comments from the general public. K-Means Clustering is an algorithm data aims to classify observational objects into clusters. The Term Frequency-Inverse Document Frequency method was to transform textual data into numerical data for feature weighting. The methodology used to develop this application was the Waterfall methodology. The train data consisted of 190 instances, while the test data used consisted of 82 instances. The results obtained a silhouette score of 0.13 and a model accuracy of 56%. The applications can categorize into positive, neutral, and negative sentiments from the public toward the development of road infrastructure in carrying out the routine maintenance program of government roads in Manado City.