<![CDATA[This study provides a paradigm for using a digital decision support system to automate OSCE evaluation. The effectiveness of this model is restricted to the scope of small-scale data and particular educational situations at Aisyah University, despite the results demonstrating great accuracy. As a result, additional modifications are needed for its practical implementation at other institutions. However, this research provides a crucial basis for the creation of digital assessment systems that might assist teachers in identifying students who want extra aid prior to final exams. Five machine learning algorithms Neural Network (NN), Support Vector Machine (SVM), Random Forest (RF), Naive Bayes (NB), and K-Nearest Neighbors (kNN) are assessed experimentally in this study. A dataset of 439 clinical competency data from Aisyah Pringsewu University midwifery students was used to create the model. Eight clinical skill factors were used as input, including baby massage, newborn care, and family planning services. To guarantee result stability, the 5-fold cross-validation approach was used for model validation. According to the test findings, every algorithm performs well, with an accuracy of more than 90%. On this particular dataset, SVM achieved a 100% classification accuracy, whereas Random Forest and SVM showed the most efficacy. With an average validation accuracy of 95%, neural networks also demonstrated excellent performance. This study provides a paradigm for using a digital decision support system to automate OSCE evaluation. The effectiveness of this model is restricted to the scope of small-scale data and particular educational situations at Aisyah University, despite the results demonstrating great accuracy. As a result, additional modifications are needed for its practical implementation at other institutions. However, this research provides a crucial basis for the creation of digital assessment systems that might assist teachers in identifying students who want extra aid prior to final exams.]]>
