Application of Statistical Model in Wastewater Treatment Process Modeling Using Data Analysis

• Main Authors: Alireza Raygan Shirazinezhad, Morteza Zare, Fahime Zare, Mohammad Mehdi Baneshi, Soheila Rezaei
• Format: Article
• Language: fas
• Published: Alborz University of Medical Sciences 2015-06-01
• Series: Muhandisī-i Bihdāsht-i Muḥīṭ
• Subjects: Wastewater Treatment Process Modeling; Data Analyzing; Classification; Kohgiluyeh and Boyer Ahmad
• Online Access: http://jehe.abzums.ac.ir/browse.php?a_code=A-10-111-4&slc_lang=en&sid=1

Background: Wastewater treatment includes very complex and interrelated physical, chemical and biological processes which using data analysis techniques can be rigorously modeled by a non-complex mathematical calculation models. Materials and Methods: In this study, data on wastewater treatment processes from water and wastewater company of Kohgiluyeh and Boyer Ahmad were used. A total of 3306 data for COD, TSS, PH and turbidity were collected, then analyzed by SPSS-16 software (descriptive statistics) and data analysis IBM SPSS Modeler 14.2, through 9 algorithm. Results: According to the results on logistic regression algorithms, neural networks, Bayesian networks, discriminant analysis, decision tree C5, tree C & R, CHAID, QUEST and SVM had accuracy precision of 90.16, 94.17, 81.37, 70.48, 97.89, 96.56, 96.46, 96.84 and 88.92, respectively. Discussion and conclusion: The C5 algorithm as the best and most applicable algorithms for modeling of wastewater treatment processes were chosen carefully with accuracy of 97.899 and the most influential variables in this model were PH, COD, TSS and turbidity.