<![CDATA[General Background: Evaluating public satisfaction with government services is vital to ensuring transparency and continuous improvement in public administration. Specific Background: At the Investment and One-Stop Integrated Services Office (DPMPTSP) of Minahasa Regency, satisfaction assessment has been limited by manual data processing and a lack of integrated systems, leading to inefficiencies in monitoring and classification. Knowledge Gap: Existing approaches to measuring the Public Satisfaction Index (IKM) have not effectively utilized machine learning to automate classification and provide real-time recommendations. Aims: This study aims to implement the Support Vector Machine (SVM) algorithm to classify public satisfaction levels and support service evaluation at DPMPTSP Minahasa. Results: Using 182 testing datasets, the system successfully categorized satisfaction into four levels—very satisfied, satisfied, less satisfied, and dissatisfied—with the majority of respondents classified as satisfied. The developed web-based system also provided actionable recommendations for each satisfaction level. Novelty: This study presents an integrated and automated framework that applies SVM to the public service domain, enabling efficient, accurate, and real-time evaluation. Implications: The findings demonstrate that machine learning can enhance public service management by facilitating data-driven decision-making and promoting service quality improvements. Highlight : The SVM algorithm effectively classifies public satisfaction levels into four categories. The web-based system improves efficiency and accuracy in service evaluation. Recommendations from the system support continuous service quality improvement. Keywords : Public Satisfaction Index, Support Vector Machine, Classification, Service Quality, DPMPTSP Minahasa]]>
