Application of Gene Expression Programming to water dissolved oxygen concentration prediction

Abstract

This research based on record and collected data from four stations at Eymir Lake, Turkey, which are monitored daily in seven months. Water quality monitoring using former methods are time-needed and expensive, while the application of gene expression programming is more understandable, rapid, and reliable which is used in this article to provide a prediction for dissolved oxygen. The concentration of oxygen is one of the most important factors of water quality identification, which shows if water has proper ability for aquatic life, agriculture, sanitary and drink, or not. Therefore, the concentration of oxygen is one of the most important parameters, which cannot be calculated by mathematical analyses directly. Phosphor, nitrate, phosphate, dissolved nitrogen, water alkalinity, water temperature, dissolved chlorophyll, electrical conductivity, precipitation rate, wind velocity and environment temperature are parameters which used as correlated factors to better prediction of dissolved oxygen in this paper. In the best model determination coefficient and root mean square error values respectively, were found to be 0.8031 and 0.0937. Finally, the assessment of forecasted data showed that the proposed approach produces satisfactory results.