<![CDATA[The lack of clarity in koi fish (Cyprinus rubrofuscus) pricing within the Indonesian ornamental fish market, driven by subjective valuation practices and information asymmetry, remains a primary challenge creating significant price disparities. The primary objective of this research is to address this challenge by designing and evaluating an objective predictive model. As its main contribution, this study develops the first Random Forest Regressor (RFR) based price prediction model. This model is specifically designed to handle complex non-linear relationships by integrating three main feature groups: morphometric parameters (species, size) and phenotypic characteristics (color patterns). Using a dataset of 800 samples collected from koi breeding centers in East Java, the optimized model achieved solid predictive performance, indicated by a Coefficient of Determination (R²) of 0.85 and a Root Mean Squared Error (RMSE) of IDR 265,000. Feature importance analysis revealed the significant finding that fish variety (one of the three analyzed feature groups) is the most dominant price determinant (62% contribution). The model quantitatively validates that rare varieties (such as Tancho/Utsuri) are valued 3 to 5 times higher than common varieties of the same size. Comparative analysis with traditional linear regression models (R² 0.61) also demonstrated the RFR's superiority in capturing complex morphological feature interactions. A critical finding indicates that the model's accuracy, already strong in the non-premium segment, can be improved by up to 15% through the quantification of qualitative aesthetic attributes (such as kiwa or gradation) using computer vision. The implementation of this model has the potential to standardize koi valuation, reduce market information asymmetry by up to 40%, and serve as a foundation for the development of the first AI-based price recommendation system in Indonesia's aquaculture industry.]]>
