Prediction of Biochemical Oxygen Demand Using Radial Basis Function Network

Abstract

Biochemical oxygen demand shows the amount of oxygen needed by microorganisms to decompose dissolved organic substances suspended in water. This variable determines water quality. The higher value indicates lower water quality. Obtaining this value requires a lengthy procedure of five days in typical laboratories. This paper proposes to predict biochemical oxygen demand using a radial basis function network with improvement relational fuzzy c-means clustering to set centroid by using 11 parameters that come from water quality records. The dataset used in testing consisting of weekly parameters between 2014-2019. Testing results show performance measurement of mean absolute error, mean square error, root mean square error, mean absolute percentage error, and accuracy using centroid with improvement relational fuzzy c-means 0.15016, 0.3677, 0.19082, 21.64490 and 78.35510 comparing with centroid from fuzzy c-means 0.16002, 0.04021, 0.19963, 22.83184, and 77.16816.