Grantee Research Project Results
2003 Progress Report: Using Neural Networks to Create New Indices and Classification Schemes
EPA Grant Number: R829784Title: Using Neural Networks to Create New Indices and Classification Schemes
Investigators: Brion, Gail M.
Current Investigators: Brion, Gail M. , Lingireddy, Srinivasa
Institution: University of Kentucky
EPA Project Officer: Page, Angela
Project Period: July 1, 2002 through June 30, 2005 (Extended to June 30, 2006)
Project Period Covered by this Report: July 1, 2003 through June 30, 2004
Project Amount: $523,938
RFA: Microbial Risk in Drinking Water (2001) RFA Text | Recipients Lists
Research Category: Water , Drinking Water , Human Health
Objective:
The objective of this research project is to use advanced computer modeling techniques to predict enhanced risk, as represented by peak numbers of potential pathogens, in a drinking water source from other more easily measured surrogate parameters.
Our hypothesis is neural network models can reliably mathematically relate shifts in indicator and indigenous bacterial populations to the presence, concentration, age, and source of microbial pathogens.
The specific objectives of this research project are to:
Riverine-Scale Objectives
- Develop a tool that provides early warning of potentially “risky” conditions in source waters for water treatment plants.
- Collect and analyze surface water samples for a host of surrogate indices and probable human pathogens over a multiyear period to create a large, multiparameter database from a single, well-characterized surface water source (Kentucky River).
- Apply traditional modeling techniques and statistical analysis to the multiparameter database to illuminate relationships between indices, pathogen presence, and origination source in a predictive manner.
- Define the correlation between atypical coliforms and the presence/concentration of other surrogate indicators, enteric viruses, and protozoa.
- Confirm the relationship between the ratio of atypical coliforms to total coliforms (AC/TC) with the time of fecal material in the environment.
- Apply neural network modeling to identify the important indices and combinations of indices related to pathogen presence, concentration, and probable source.
- Gain further insight into neural network architecture and training schemes that are appropriate for surface water quality modeling of indicator microbes.
Laboratory-Scale Objectives
- Obtain kinetic data on relationships between fecal and indigenous bacterial concentrations and pathogens as fecal material ages in surface water.
- Fit appropriate models to these results.
- Predict predominant fecal source and fecal age using neural network modeling from the data obtained from laboratory experiments.
- Predict the initial concentration of fecal bacteria from sampling of aged surface waters.
Progress Summary:
Analysis of the river data by artificial neural network (ANNs) is difficult at this time because of the limited amount of data collected. There are some interesting trends, however, in the data. In the ongoing Kentucky River Study, initial multiple logistic regression on a 30-observation database has shown that three input variables can be used to predict enteric virus presence with 93.3 percent accuracy. The AC/TC ratio is significant for viral presence prediction, modifying the load signal from indicator bacteria and bacteriophage concentrations with respect to the average age of fecal contamination in the river system. The preliminary model was: Logit P = - 4.204 + (5.940 * log fecal coliforms) - (4.471 * log [AT/TC]) + (3.786 * male specific coliphage). The relative importance of the input variables is log AT/TC (p = 0.060), log fecal coliforms (p = 0.073), and male-specific coliphage (p = 0.081).
Giardia cysts are always present, but Cryptosporidium oocysts are only present during times when the river turbidity was more than 200. Enterococci and fecal coliforms are highly correlated (p<0.001). The AC/TC ratio (the new indicator under development by this study) is not linearly related to the non-transformed values of bacterial indicators, but is showing the expected decreases during times of rain that indicates the presence of fresher fecal material and is critical to modeling efforts. This supports the findings obtained from three other local watersheds. Coprostanol and epicoprostanol are present in detectable amounts in all samples, but the long extraction procedure has these data lagging behind the other more immediately measured analytes. All positive, and confirmed, enteric virus flasks are being frozen and archived for future identification by genetic methods.
The data collection effort is expected to yield the necessary data for preliminary ANN analysis by the end of September 2004. In related work, ANNs have been applied to a database consisting of indicators and enteric viruses recovered from shellfish and their surrounding waters from four countries in Europe. It was shown that ANNs outperformed logistic regression for the classification of viral presence and type, and site-specific differences in the relationships between indicators and potential pathogens were indicated. The efficacy of ANNs in backfilling missing microbial data on the river system under study has been demonstrated. Through these efforts, more robust and efficient ANN training termination criteria have been developed. The new bacterial ratio proposed (AC/TC) has been shown to be superior to fecal coliform and Enterococci levels for indicating inputs of fresh human fecal contamination in a related study.
Collaboration with a group of researchers in Europe that offered a large bacterial, phage, and viral database; analysis of an in-house microbial water quality database from the Kentucky River; and a database from a smaller local watershed have resulted in the publication of two papers in an international journal and four national/international conference presentations, two of which received awards. In addition, five new journal articles and two national/international conference presentations have been submitted.
The principal investigators have given a seminar on emerging pathogens, jointly hosted with the Kentucky Watershed Watch Group for outreach to county agents, faculty, and other interested parties. This event was broadcast remotely to three local universities and advertised through the University of Kentucky Tracy Farmer Center for the Environment and the State Division of Water for the widest reach to potentially interested people. In addition, two workshops were arranged for the graduate students at the University of Kentucky on the use of the in-house ANN software package NeuroSort.
Future Activities:
We will continue river sampling on a weekly basis, compile the already acquired Kentucky River data into a format quickly amenable for the ANN analysis towards the verification of the proposed hypothesis, continue comparison studies of the ANN analysis against logistic regression and other methods of data analysis on multiparameter databases, and investigate the applicability of ANNs to bootstrapping and data expansion procedures. The principal investigators are promoting the use of the new AC/TC bacterial ratio with other researchers.
Journal Articles on this Report : 5 Displayed | Download in RIS Format
| Other project views: | All 21 publications | 9 publications in selected types | All 9 journal articles |
|---|
| Type | Citation | ||
|---|---|---|---|
|
|
Booth J, Brion GM. The utility of the AC/TC ratio for watershed management: a case study. Water Science & Technology 2004;50(1):199-203. |
R829784 (2003) R829784 (Final) |
|
|
|
Brion GM. The AC/TC bacterial ratio: a tool for watershed quality management. Journal of Water and Environment Technology 2005;3(2):271-277. |
R829784 (2003) R829784 (2004) R829784 (Final) |
Exit Exit |
|
|
Brion G, Lingeriddy S, Neelakantan TR, Wang M, Girones R, Lees D, Allard A, Vantarakis A. Probing Norwalk-like virus presence in shellfish, using artificial neural networks. Water Science & Technology 2004;50(1):125-129. |
R829784 (2003) R829784 (Final) |
|
|
|
Brion G, Viswanathan C, Neelakantan TR, Lingireddy S, Girones R, Lees D, Allard A, Vantarakis A. Artificial neural network prediction of viruses in shellfish. Applied and Environmental Microbiology 2005;71(9):5244-5253. |
R829784 (2003) R829784 (2004) R829784 (Final) |
Exit Exit |
|
|
Chandramouli V, Brion G, Neelakantan TR, Lingireddy S. Backfilling missing microbial concentrations in a riverine database using artificial neural networks. Water Research 2007;41(1):217-227. |
R829784 (2003) R829784 (2004) R829784 (Final) |
Exit Exit |
Supplemental Keywords:
water quality, pathogens, indicators, modeling, artificial neural networks, ANNs, risk management, remediation, Giardia, Cryptosporidium, enteric viruses, EPA Region 4, alternative disinfection methods, bacteria, drinking water contaminants, early warning, ecological risk, emerging pathogens, environmental monitoring, microbial contamination, microbial pathogens, microbial risk assessment, microbiological organisms, water contaminants, water contamination detection, geographic area, water, drinking water, ecological risk assessment, engineering, chemistry, physics, environmental chemistry, environmental engineering, water contaminants, water contamination detection,, RFA, Scientific Discipline, Water, Geographic Area, Environmental Chemistry, Ecological Risk Assessment, Ecology and Ecosystems, Drinking Water, Engineering, Chemistry, & Physics, Environmental Engineering, EPA Region, microbial risk assessment, alternative disinfection methods, microbial contamination, environmental monitoring, water contamination detection, region 4, bacteria, microbiological organisms, early warning, microbial pathogens, cryptosporidium , neural networks, emerging pathogens, water quality, ecological risk, drinking water contaminantsProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.