Fish Disease Prediction Using Machine Learning on Water Quality Data

Fish diseases are a serious threat to the farming industry. To detect these diseases in the initial stage, a different machine learning model is proposed. This model has focused on identifying the fish diseases based on the water quality. To do this, we have used the “Aquaculture–Water Quality Dataset” data set havi15 physico-chemical parameters like PH, DO, hardness, solids, chloramines, iron, ammonia, nitrite, nitrate, phosphate, silica and the like. These attributes are linked with the existence of a disease. From the data set, we have created the Random Forest Classifier using the Python language and the Scikit library. The implementation has provided an efficiency of 89.25%. The metrics of precision, recall, and F1-score were employed to determine the efficiency. The evaluation has revealed that the proposed model has displayed the appropriate efficiency. The important feature analysis has shown that the dissolved oxygen , temperature ,PH, and ammonia are the critical factors used for the detection of the disease. Through the proposed method, a proper mechanism has been developed for detecting the fish diseases. This model can be used effectively to create proper system for the detection of fish diseases.