Hybrid approach for water quality UV-Vis absorbance time series forecasting in sewer systems

Andres Eduardo Torres Abello; Leonardo Plazas-Nossa; Thomas Hofer; ESCOBAR Jorge; Leonardo Florez Valencia; Günter Gruber

Hybrid approach for water quality UV-Vis absorbance time series forecasting in sewer systems

Andres Eduardo Torres Abello, Leonardo Plazas-Nossa, Thomas Hofer, ESCOBAR Jorge, Leonardo Florez Valencia, Günter Gruber

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

This work presents a hybrid approach based on seven methodologies combined with Principal Component Analysis (PCA) for UV-Vis absorbance time series forecasting. It was applied to four absorbance data sets in four different study sites. Thus, it is important to determine which forecasting methodology is best suited for different wavelengths, according to the target water quality. The Mean Absolute Percentage Error (MAPE) values were obtained in a range between 0% and 57% for all the study sites. Results shown that is not possible to have a best forecasting methodology among the proposed ones, because all of them would complement each other for different forecasting time steps and spectra range.

Original language	English
Title of host publication	14th International Conference on Urban Drainage-ICUD 2017
Number of pages	9
State	Published - 2017

Keywords

Artificial Neural Networks
Discrete Fourier Transform
Polynomial transforms
Principal Component Analysis
UV-Vis time series forecasting
Water Quality

Cite this

@inproceedings{1ae4978bb8394768bf734eecd7f4d2f9,

title = "Hybrid approach for water quality UV-Vis absorbance time series forecasting in sewer systems",

abstract = "This work presents a hybrid approach based on seven methodologies combined with Principal Component Analysis (PCA) for UV-Vis absorbance time series forecasting. It was applied to four absorbance data sets in four different study sites. Thus, it is important to determine which forecasting methodology is best suited for different wavelengths, according to the target water quality. The Mean Absolute Percentage Error (MAPE) values were obtained in a range between 0% and 57% for all the study sites. Results shown that is not possible to have a best forecasting methodology among the proposed ones, because all of them would complement each other for different forecasting time steps and spectra range.",

keywords = "Artificial Neural Networks, Discrete Fourier Transform, Polynomial transforms, Principal Component Analysis, UV-Vis time series forecasting, Water Quality",

author = "{Torres Abello}, {Andres Eduardo} and Leonardo Plazas-Nossa and Thomas Hofer and ESCOBAR Jorge and {Florez Valencia}, Leonardo and G{\"u}nter Gruber",

year = "2017",

language = "English",

booktitle = "14th International Conference on Urban Drainage-ICUD 2017",

}

TY - GEN

T1 - Hybrid approach for water quality UV-Vis absorbance time series forecasting in sewer systems

AU - Torres Abello, Andres Eduardo

AU - Plazas-Nossa, Leonardo

AU - Hofer, Thomas

AU - Jorge, ESCOBAR

AU - Florez Valencia, Leonardo

AU - Gruber, Günter

PY - 2017

Y1 - 2017

N2 - This work presents a hybrid approach based on seven methodologies combined with Principal Component Analysis (PCA) for UV-Vis absorbance time series forecasting. It was applied to four absorbance data sets in four different study sites. Thus, it is important to determine which forecasting methodology is best suited for different wavelengths, according to the target water quality. The Mean Absolute Percentage Error (MAPE) values were obtained in a range between 0% and 57% for all the study sites. Results shown that is not possible to have a best forecasting methodology among the proposed ones, because all of them would complement each other for different forecasting time steps and spectra range.

AB - This work presents a hybrid approach based on seven methodologies combined with Principal Component Analysis (PCA) for UV-Vis absorbance time series forecasting. It was applied to four absorbance data sets in four different study sites. Thus, it is important to determine which forecasting methodology is best suited for different wavelengths, according to the target water quality. The Mean Absolute Percentage Error (MAPE) values were obtained in a range between 0% and 57% for all the study sites. Results shown that is not possible to have a best forecasting methodology among the proposed ones, because all of them would complement each other for different forecasting time steps and spectra range.

KW - Artificial Neural Networks

KW - Discrete Fourier Transform

KW - Polynomial transforms

KW - Principal Component Analysis

KW - UV-Vis time series forecasting

KW - Water Quality

M3 - Conference contribution

BT - 14th International Conference on Urban Drainage-ICUD 2017

ER -

Hybrid approach for water quality UV-Vis absorbance time series forecasting in sewer systems

Abstract

Keywords

Other files and links

Fingerprint

Cite this