Seawater salinity modeling using bivariate probit regression

Salt is one of the marine resources that is quite a lot needed as a supplementary food for the people of Indonesia. However, efforts to increase salt production have not been in demand, including in efforts to improve its quality, because many factors affect sea salt content or salinity, including the evaporation process, location and size of the sea, wind, air humidity and sea water temperature in this study are expected to produce the best salinity modeling by taking into account the factors that affect salinity. In this study, the method used was probit bivariate. The parameter estimation method used in the bivariate probit is the Maximum Likelihood Estimation (MLE). After the initial bivariate probit regression model is formed, then testing is carried out to determine the significance of each predictor variable to the response variable. After that the model that is formed identifies the criteria of goodness using the smallest Akaike Information Criterion (AIC) value of -9.03 so that the modeling results are good.