<![CDATA[Micro, Small, and Medium Enterprises (SMEs) have a vital role in Indonesia’s economy. However, IT-based marketing strategies among SMEs receive limited support from the government due to the lack of sufficient data to inform policy. This study aims to (1) identify the needs of SMEs for social media promotion training as part of their digital capacity building, (2) develop and compare the effectiveness of classification models that combine Fuzzy C-Means and K-Means clustering algorithms with the Naïve Bayes algorithm to group SMEs based on business characteristics, (3) analyze the relationships between business variables—such as business type, marketing media, funding sources, and financial aspects—and SME performance through regression analysis, and (4) provide data-driven foundations for designing targeted digital interventions and policy strategies to support SME development in Indonesia. This study used UPPKS data from 133 SMEs in seven districts in the Special Region of Yogyakarta. Data analysis covered business types, marketing platforms used, funding sources, and financial performance indicators. Data pre-processing involved cleaning, normalization, and integration to ensure consistency and readiness for analysis. The researcher used the Elbow method to determine the optimal number of clusters. Then, it also used Fuzzy C-Means (FCM) and K-Means to categorize SMEs into three groups: high, medium, and low. The classification was based on the Naïve Bayes algorithm. The evaluation of the model performance used a confusion matrix, cross-validation, and regression analysis to examine inter-variable relationships. The results showed that the combination of FCM and Naïve Bayes achieved an accuracy of 85% based on the confusion matrix and 97% based on cross-validation. Meanwhile, the combination of K-Means and Naïve Bayes respectively achieved an accuracy of 96% and 94.7%. These findings demonstrate the effectiveness of the proposed approaches in classifying SMEs based on their characteristics and performance. This research provides important insights for policymakers and SME development agencies in designing more targeted digital training and support programs. Future studies should explore the integration of other algorithms, such as Support Vector Machines (SVM) and Decision Trees, while incorporating market trends and customer engagement to enhance SME classification and provide ongoing support.]]>
