Prediction analysis of oxygen content in the water for the fish farm in southern Taiwan

This paper is using artificial neural network (ANN) to predict oxygen content in the water for the fish farm, so that decrease times of starts of oxygen suppliers. In Southern Taiwan, aquaculture is one of major economic industries. Especially, the important issue is how to effectively monitor the oxygen content in the water, so that the fish will not die and start the oxygen suppliers for the minimum of times. According to experience of aquaculture practitioners, the impact factors of oxygen content in the water include temperature, pH, conductivity, salinity and last monitored oxygen content. And ANN is one of frequently used tools about analysis and prediction. In ANN, there are three parts, including input layer, hidden layer and output layer. Input layer and output layer are given by users and build relations between them by hidden layer. In this paper, data provided by Ecotek company divided into training data and testing data. The experimental process is as following: corrected data, set parameters, separated into training data and testing data, and executed neural network. From the experimental result, although it is not possible to achieve a complete positive correlation, but the oxygen content can be kept between 3 and 7 ppm.