SUMMARY
Water condition is the main factor that affects the success rate of aquaculture, especially in shrimp cultivation. However, the farmer often experiences difficulties in determining the condition which is stated based on the measurement of various water parameter. Therefore, a proper classification model is needed to help the farmer in classifying the water condition in a pond. By knowing the condition, then proper and correct treatment can be given. In this research, a machine learning algorithm called SVM is used to make a model from an aquaculture dataset. Another processing technique like data normalization and the usage of optimization algorithm named grid search is also performed to improve the modelling result. Furthermore, a test scheme with using k-fold cross-validation is performed to know the performance of the model which is measured by the value of accuracy, precision, recall, f-measure, and AUROC. Then, the SVM model is compared with several models which are made by using another machine learning algorithm such as KNN, CNB, RF, MLP, and LR in order to know the best model to be implemented on cultivation process. From the experiment results, the model which is made with SVM and grid search optimization has the best performance in the validation process with the performance score of 3.54383.