An ensemble method for predicting biochemical oxygen demand in river water using data mining techniques

ABSTRACT

Biochemical oxygen demand (BOD) is used to determine the amount of dissolved oxygen used by microbial oxidation of organic content. BOD is a parameter describing the quality of water, especially its extent of pollution. Water from wastewater treatment plants have high BOD values and as such require to be treated. For this purpose, BOD is used as an indicator in determining the quality of water being discharged. The standard method for measuring BOD is a 5-day process. Dilution of sample, constant pH and nutrient content besides temperature of 20°C and dark area are required for correct results. High levels of nitrogen compounds yield false BOD results. Winkler titration, which is also used to measure DO (as part of BOD measurement), is a chemical-intensive process. Hence an automatic prediction model for BOD is required for accurate, cost-effective and time-saving measurement. Based on data available for BOD measurements, the present study focuses on devising a prediction model for BOD using ensemble techniques in data mining. A correlation coefficient of 0.9541 and a root mean-squared error of 0.4679 were obtained for the proposed BOD prediction model on river water quality data. Comparative analysis of the proposed model with existing models built for the same data set was also performed.

Keywords:

• Biochemical oxygen demand
• data mining
• ensemble techniques
• bagging
• K-star algorithm
• correlation coefficient
• RMSE