Analysis waste water characteristics via data mining: A Muğla province case and external validation

Güller S., SİLAHTAROĞLU G., Akpolat O.

Communications in Statistics Case Studies Data Analysis and Applications, vol.5, no.3, pp.200-213, 2019 (Scopus) identifier


This study is related to the analysis of characteristic parameters of domestic wastewater samples which are collected from municipal wastewater treatment plants located in the provinces of Muğla in Turkey. Used parameters are pH, temperature, conductivity, suspended solids, dissolved oxygen, oxygen saturation, salinity, electrical conductivity, chemical oxygen demand (COD), total phosphorus, total organic carbon, total nitrogen, and biological oxygen demand. As 7 of these 12 parameters selected, pH, temperature, suspended solids, COD, total phosphorus, total nitrogen, and biological oxygen demand, can be measured within a day, chemical laboratory analysis of the BOD5 (biological oxygen demand) parameter lasts 5 days to be tested. The objective of the study is to find a data analysis model to predict BOD5 using other 6 parameters. Our dataset consists of from 334 samples which ware collected on a daily basis. The effects of those 6 parameters on BOD5 were examined by decision tree and artificial neural networks methods using KNIME data mining package. In total 53% of samples in the dataset have a BOD5 value which is lower than 100. About 15.3% are between 100 and 200; 12.6% of samples have got BOD5values which are between 450 and 550, and the rest are higher than 550. Our results suggest that the most important parameter which affects BOD5 value is COD. If the COD is less than or equal to 214.93, the value of BIO5 is between 0 and 100 and, its frequency of occurrence is 98.6%. When COD is greater than 214.93, BOD5 never exceeds 200. In this case, the probability of BOD5 being between 100 and 200 is only 1.4%. In cases the chemical oxygen requirement is between 214.93 and 390.445, the most important parameter in predicting the value of BOD5 is the total phosphorus value. In the study, we present the analysis and discuss the results in detail.