Publication: Forecasting Water Quality, Parameters By Neural Network Technique
Date
2020-02
Authors
Aldobai, Aiad Mazin Deeb
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Throughout this study, an approximate Water Quality Index (WQI) in Kelantan
River, Malaysia, will also be developed and validated for an Artificial Neural Network
(ANN). Digital information from 30 data collection locations was designed and tested
for the ANN model. Two sets for simulation data were separated. For the first set of
ANNs, five independent water quality variables were used as input parameters to be
trained, checked and validated. As a result, multiple linear regression (MLR) was
utilized to replace the independent variables with the lowest variance commitment.
Dissolved Oxygen (DO) cover independent variables representing substantially 75% of
WQI variance; Biochemical Solids (SS), Ammoniacal Nitrate (AN), Biochemical
Oxygen Demand (BOD). Only 8% and 2% of the discrepancy made a significant
contribution to the substance Oxygen Demand (COD) and the pH.
Therefore, only four independent variables for preparation, testing and
recognition of ANN have been used in the 2nd collection of information. In addition,
the statistical significance given by six independent variables (0.92) in prediction of
WQI are only marginally better than the actual WQI (0.91) data sets, including one that
clearly demonstrate the ANN network, whether it is trained by fuzzy inference system
(ANFIS) to redact COD and the pH as independent variables, to investigate WQI
accurately .
Description
FYP Sem 2 2019/2020
Keywords
WQI by IE Moddling , Water Quality