Prediction of mullet catch in Iranian waters of the Caspian Sea using decision tree algorithm

Every year the trend of catches and the maximum harvestable water from the water resources are studied and studied. Extensive studies of catch data have been conducted to estimate reserves. The main objective of this study was to predict Caspian mullet catch by decision tree algorithm (C&R tree) presented using Clementine 18 data mining software. Decision tree structure in machine learning is a predictive model that contributes to the observed facts about a phenomenon with inferences about the objective value of that phenomenon. For this purpose, 10-year catch data as a dependent variable (year of catch, month of catch, latitude and longitude of catch point, depth of catch, rate of catch) derived from bony fish stock assessment projects and environmental information (temperature, chlorophyll-a) From the NASA site and the factors of pH, transparency, salinity, oxygen, benthic organisms from northern seas hydrology projects) were considered as independent variables that besides specifying the effective factor on fishing, prediction of fish mullet in water The Caspian Sea also occurs. Among the investigated factors, water surface temperature and salinity were identified as the two main factors affecting fish mullet catchment and the prediction value for 1400, 862/484 tons of mullet with an accuracy of 98/42%.