Ensemble of ANN and ANFIS for Water Quality Prediction and Analysis - A Data Driven Approach
The consequences of un-clean water are some of the direst issues faced by humanity today. These concerns can be addressed efficiently if data is pre-analyzed and water quality is predicted before its effects occur. The aim of this research is to develop a novel ensemble of Artificial Neural Netw...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | en |
| Published: |
Universiti Teknikal Malaysia (UTEM)
2017
|
| Subjects: | |
| Online Access: | http://ir.unimas.my/id/eprint/21937/6/Ensemble.pdf http://ir.unimas.my/id/eprint/21937/ http://journal.utem.edu.my/index.php/jtec/article/view/2685 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The consequences of un-clean water are some of the
direst issues faced by humanity today. These concerns can be
addressed efficiently if data is pre-analyzed and water quality is
predicted before its effects occur. The aim of this research is to
develop a novel ensemble of Artificial Neural Network (ANN)
and Adaptive Neuro-Fuzzy Inference System (ANFIS) models
using averaging ensemble technique, producing improved
prediction accuracy. Measurements of different water quality
parameters have been used for predicting the overall water
quality, applying ANN, ANFIS and ANN-ANFIS ensemble and
their results have been compared. The data used in this study is
obtained by USGS online repository for the year of 2015, with a
30-minutes time interval between measurements. Root Mean
Squared Error (RMSE) has been used as the main performance
measure. The results depict a significant improvement in the
Ensemble ANN-ANFIS model (RMSE: 0.457) as compared to
both the ANN model (RMSE: 2.709) and the ANFIS model
(1.734). The study concludes that the ensemble of ANN and
ANFIS model shows significant improvement in prediction
performance as compared to the individual models. The
research can prove to be beneficial for decision making in terms of water quality improvement. |
|---|