<![CDATA[This study addresses the problem of unpredictable customer surges at Manggala Motor Workshop, which often lead to long queues, inefficient resource allocation, and reduced service quality. To overcome this problem, the Support Vector Machine (SVM) algorithm was applied to classify workdays into two categories: busy and not busy. The dataset consisted of 400 simulated data points designed to represent real workshop operational conditions by incorporating attributes such as day, weather, promotions, holidays, number of bookings, and number of vehicles. The data acquisition process was carried out through simulation based on average service capacity and external factors that typically influence customer arrivals. Before modeling, preprocessing steps were performed, including one-hot encoding for categorical features and normalization for numerical features. The dataset was then split into 80% training data (320 entries) and 20% test data (80 entries). Using a linear kernel, the SVM model was implemented in Google Colab with the Scikit-learn library. The results showed an accuracy of 96.25%, with high precision and recall scores in both classes. These findings indicate that SVM is effective for binary classification of busy and non-busy days, enabling Manggala Motor Workshop to optimize technician scheduling, manage workloads, and allocate resources more efficiently, thereby improving service quality and customer satisfaction.]]>
Copyrights © 2025
