A decision fusion method based on classification models for water quality monitoring

Abstract

-Monitoring of water quality is one of the world's main intentions of countries. In this paper we present the use of Principal Component Analysis (PCA) combined with Support Vector Machines (SVM) and Artificial Neural Network (ANN) based on Decision Templates combination data fusion method. SVM and ANN are employed in classification stage. Decision Templates is applied to increase accuracy of the water quality classification compared to others combination data fusion methods. This work concerned the water quality assessment from Tilesdit dam (Algeria) that it permitted us to acquire additional knowledge and information about study area and to obtain an intelligent monitoring system. The Multi-Layer Perceptron network (MLP) and the One-Against-All strategy for SVM method are have been widely used. The training step is performed in this paper using these techniques to classify water quality from various physicochemical parameters such as temperature, pH, electrical conductivity and turbidity, etc. Eight of them were collected in the period 2009-2018 from the study area. The selection of the excellent parameters of the used models can be improving the performance of classification process. In order to assess their results, an experiment step using collected data set corresponding to the accuracy and running time of training and test phases, and robustness, is carried out. Various scenarios are examined in comparative study to obtain the most results of decision step with and without features selection of the input data. The combination by Decision Templates of two classifiers enhanced expressively the results of the proposed monitoring framework that had prove a considerable ability in surface water quality assessment.

Water Quality Assessment Tool for On-Site Water Quality Monitoring

“…While machine learning algorithms like ANN and SVM offer simplicity, they struggle with non-linear relationships, posing challenges in capturing complex structures. Furthermore, the work of Ladjal et al [13] did not provide an interpretability mechanism for their model classifications. Khelil et al [14] compared the performance of SVM and the Long Short-Term Memory (LSTM) binary classifiers using the Tilesdit dam water quality data.…”

“…In the literature, several methods have been employed to classify water quality samples and determine their suitability for various purposes [13]- [17]. Ladjal et al [13] utilized artificial neural networks (ANN) and support vector machines (SVMs) to assess the quality of water from the Tilesdit dam.…”

Comparative Study of Algorithms Used in Water Pollution Prevention and Control

“…Elkiran [37] analyzed and compared different integration techniques for Backpropagation Neural Network (BPNN), Adaptive Neuro-Fuzzy Inference System (ANFIS), Support Vector Machine (SVM) and Linear Autoregressive Integral Moving Average (ARIMA) models, applied to water quality modeling of Yamuna River in India, it proves that the performance of ANFIS model is better. Ladjal [38] proposed a data fusion method using Principal Component Analysis (PCA) in combination with Support Vector Machines (SVM), Artificial Neural Networks (ANN) and decision templates for water quality monitoring in the Tilesdit dam area in Algeria. The achieved classification accuracy was 98%.…”

Water Pollutant Classification Method Based on Multi-Class Support Vector Machine

“…Solanki [9] analyzed and compared the application of deep learning algorithms and other unsupervised learning algorithms for classification of river pollution problems near Nasik, Maharashtra, India. Ladjal [10] proposed a data fusion method using principal component analysis (PCA) combined with support vector machines, artificial neural networks and decision templates for water quality classification and monitoring in the Tilesdit dam area in Algeria. The research found that SVM, The integration of ANN and PCA achieves a classification accuracy of 98%.…”