Doctor of Philosophy (PhD)
Electrical and Computer Engineering
First Committee Member
Second Committee Member
Third Committee Member
Fourth Committee Member
Number of Pages
Water is one of the most important natural resources for all living organisms on earth. The monitoring of treated wastewater discharge quality is vitally important for the stability and protection of the ecosystem. Collecting and analyzing water samples in the laboratory consumes much time and resources. In the last decade, many machine learning techniques, like multivariate linear regression (MLR) and artificial neural network (ANN) model, have been proposed to address the problem. However, simple linear regression analysis cannot accurately forecast water quality because of complicated linear and nonlinear relationships in the water quality dataset. The ANN model also has shortcomings though it can accurately predict water quality in some scenarios. For example, ANN models are unable to formulate the non-linear relationship hidden in the dataset when the input parameters are ambiguous, which is common in water quality dataset.
The adaptive neuro-fuzzy inference system (ANFIS) has been proven to be an effective tool in formulating the complicated linear and non-linear relationship hidden in datasets. Although the ANFIS model can achieve good performance in the water quality prediction, it has some limitations. Firstly, the size of the training dataset should not be less than the number of training parameters required in the model. Secondly, when the data distribution in the testing dataset is not reflected in the training dataset, the ANFIS model may generate out-of-range errors. Lastly, a strong correlation is required between input and target parameters. If the correlation is weak, the ANFIS model cannot accurately formulate the hidden relationship.
In this dissertation, several methods have been proposed to improve the performance of ANFIS-based water quality prediction models. Stratified sampling is employed to cover different kinds of data distribution in the training and testing datasets. The wavelet denoising technique is iv used to remove the noise hidden in the dataset. A deep prediction performance comparison between MLR, ANN, and ANFIS model is presented after stratified sampling and wavelet denoising techniques are applied. Because water quality data can be thought as a time series dataset, a time series analysis method is integrated with the ANFIS model to improve prediction performance. Lastly, intelligence algorithms are used to optimize the parameters of membership functions in the ANFIS model to promote the prediction accuracy. Experiments based on water quality datasets collected from Las Vegas Wash since 2007 and Boulder Basin of Lake Mead, Nevada, between 2011 and 2016 are used to evaluate the proposed models.
Adaptive neuro-fuzzy inference system; Artificial neural network; Machine learning; Particle swarm optimization; Time-series analysis; Water quality prediction
Computer Engineering | Electrical and Computer Engineering
University of Nevada, Las Vegas
Fu, Zhao, "Water Quality Prediction Based on Machine Learning Techniques" (2020). UNLV Theses, Dissertations, Professional Papers, and Capstones. 3994.
IN COPYRIGHT. For more information about this rights statement, please visit http://rightsstatements.org/vocab/InC/1.0/