Ensemble of naïve Bayesian approaches for the study of biofilm development in drinking water distribution systems

Abstract

The survival and regrowth of microorganisms in drinking water distribution systems (DWDSs) can be affected not only by biological aspects but also by the interaction of various other factors. Some of these factors have been found to be clearly related to biofilm development in DWDSs. However, the complexity of the microenvironment under study and the biofilm growth characteristics have so far led the various methodologies applied to produce ambiguous or not easily comparable, and thus not very useful, results. In this study we compile the information currently available on biofilm ecology in DWDSs and apply various machine learning algorithms based on naïve Bayesian networks. In addition, as a step forward, we also use ensemble methods. These methods have been widely adopted to improve the robustness and the overall prediction accuracy of single models through their accumulative experience on the performance of multiple applications in learning systems. We claim that they also reduce the high uncertainty associated with the process of biofilm development in DWDSs. Specifically, we propose alternatives for the base naïve Bayesian model to outperform its individual results while maintaining its simplicity. These alternatives include augmentation of the arcs in the graph and bagging initial approaches. Then, both ideas are combined by a hybrid algorithm that produces explanatory decision trees. As a result, it is possible to achieve a deeper understanding of the consequences that the interaction of the relevant hydraulic and physical factors of DWDSs has for biofilm development.

Which method to use? An assessment of data mining methods in Environmental Data Science

“…In geoscience and remote sensing, some works use association rules for geographic data (Rodman et al, 2006), for landscape analysis (Ferrarini and Tomaselli, 2010), for finding relations between biophysical/social parameters and urban land surface temperature (Rajasekar and Weng, 2009) or adaptations to climate change (Lynam, 2016), and for image processing for urban environmental analysis (Du et al, 2007). Regarding water topics there are some works that used association rules for coastal water classification (Pereira and Ebecken, 2009), lake sediments analysis (Annoni and Brüggemann, 2008), water resource management (Castelletti et al, 2007), biofilm development in water supply systems (Ramos- Martínez et al, 2014), and fault detection in WWTP (Ruiz et al, 2011). Cloud screening for meteorological purposes was also investigated with Markov Random Fields in Cadez and Smyth (1999).…”

“…Agriculture-related applications include Holmes et al (1998) for apple bruising, Yeates and Thomson (1996) for bull castration and venison analysis, and the Michalski and Chilausky's soybean disease diagnosis work (Michalski and Chilausky 1980), which is a classic benchmark problem in machine learning. Considerable efforts are recorded in the water-related fields, using rule-based reasoning (Zhu and Simpson, 1996;Dzeroski et al, 1997;Comas et al, 2003;Spate, 2005;Ramos-Martínez et al, 2014), decisiontrees (Kokotos et al, 2011), regression-trees (Dseroski et al, 2003), Support Vector Machines (SVM) (Kanevski et al, 2002), case-based reasoning (Martínez et al 2006 ;Wong et al, 2007), regression trees (Dzeroski and Drumm, 2003) or hybrid techniques (Cortés et al, 2002, Yang et al 2012. In the study of air quality, classification has been used for air quality data assurance issues (Athanasiadis and Mitkas, 2004) and the operational estimation of pollutant concentrations (Athanasiadis et al, 2003;Stebel et al, 2013;Yeganeh et al, 2012).…”

Graph constrained label propagation on water supply networks

“…For the experimental study we use the biofilm database of 210 pipes belonging to different WSNs, proposed and studied by the authors in [17]. In that paper, a clustering of the biofilm database was approached, among other analysis.…”

“…In that paper, a clustering of the biofilm database was approached, among other analysis. For our new interests, we keep on using the clustering medoids obtained in [17], which Table 2 shows. The corresponding clustering membership will be the labels for the elements of this study.…”

Assessing biofilm development in drinking water distribution systems by Machine Learning methods

“…This is a good value and higher than the one obtained with the RT. The good performance of the ensemble techniques on this approach has been already observed when applying them to biofilm metadata [32] (This work has been published as a journal paper and a summarized version is presented in Appendix C).…”