<![CDATA[Chinese Ceramics constitute one of the most frequently encountered archaeological artifacts in Indonesia and play a pivotal role in chronological studies and the reconstruction of historical maritime trade networks. Their presence is commonly employed as a relative dating indicator for archaeological sites and as evidence of interregional cultural interactions. Consequently, identifying the dynasty of origin of Chinese ceramics represents a critical aspect of archaeological and maritime historical research in Asia Conventionally, such identification relies on visual analysis conducted by experts, based on morphological and decorative characteristics including glaze color, decorative motifs, vessel shape, and surface texture. However, this approach is inherently subjective, time-consuming, and prone to inconsistencies particularly when applied to large assemblages of ceramic finds. This study aims to implement a machine learning (ML) approach to classify Chinese ceramic dynasties using photographic images. The research dataset comprises labeled photographs of ceramics from various dynastic periods, annotated according to their distinctive visual features. The methodological framework encompasses data collection and image preprocessing, dynasty labeling, model training via the Teachable Machine platform, and performance evaluation through classification accuracy assessment. The results demonstrate that the machine learning model developed using Teachable Machine effectively recognizes the characteristic visual patterns associated with each dynasty, achieving a satisfactory level of classification accuracy. The results demonstrate that Teachable Machine can identify the unique visual patterns of each dynasty with high precision, achieving an overall accuracy of 91%. High classification stability and performance were observed for the Qing Dynasty, with Precision of 0.95, Recall of 0.95, F1-Score of 0.95, and a matrix value of 0.94. Conversely, the lowest classification performance was recorded for the Yuan Dynasty, with Precision of 0.90, Recall of 0.84, and an F1-Score of 0.86. These findings indicate that image-based machine learning holds significant potential as a supportive analytical tool in digital archaeology particularly in enhancing the objectivity, consistency, and efficiency of Chinese ceramic identification, documentation, and data management processes.]]>
