Water quality monitoring is one of several methods to control the risk in shrimp farming. Despite its importance, monitoring water quality during shrimp farming can be costly. This research was conducted to develop prediction models that would give farmers insight about water quality characteristic of their ponds. The data were collected from 31 ponds that used the JALA platform. This research tried to predict biological and chemical properties based on physical properties of water such as temperature, dissolved oxygen, salinity, and pH. This research combined Kernel Principal Component Analysis and machine learning algorithms (Random Forest and Gradient Boosting). The results of this research showed biological and chemical conditions of water (Total Organic Matter, Hardness, Calcium, Magnesium) with R2 Score higher than 0.8 on most parameters. This study also found that the Gradient Boosting model performed better in predicting water chemical properties than Random Forest model.
Keywords: Water quality prediction, monitoring in aquaculture, machine learning for aquaculture