i ACKNOWLEDGEMENTS ii TABLE OF CONTENTS iv LIST OF FIGURES vi LIST OF TABLES vii LIST OF ABBREVATIONS viii PUBLICATIONS AND RESEARCH x

With the rapid growth over the past few decades, people are consuming more and more electrical energies. In order to solve the contradiction between supply and demand to minimize electricity cost, it is necessary and useful to predict the electricity demand. In this paper, we apply an improved neural network algorithm to forecast the electricity, and we test it on a collected electricity demand data set in Queensland to verify its performance. There are two contributions in this paper. Firstly, comparing with backpropagation (BP) neural network, the results show a better performance on this improved neural network. Secondly, the performance on various hidden layers shows that different dimension of hidden layer in this improved neural network has little impact on the Queensland's electricity demand forecasting.

With the growing number and increasing availability of shared-use instruments and observatories, observational data is becoming an essential part of application workflows and contributor to scientific discoveries in a range of disciplines. However, the corresponding growth in the number of users accessing these facilities coupled with the expansion in the scale and variety of the data, is making it challenging for these facilities to ensure their data can be accessed, integrated, and analyzed in a timely manner, and is resulting significant demands on their cyberinfrastructure (CI). In this paper, we present the design of a push-based data delivery framework that leverages emerging in-network capabilities, along with data pre-fetching techniques based on a hybrid data management model. Specifically, we analyze data access traces for two large-scale observatories, Ocean Observatories Initiative (OOI) and Geodetic Facility for the Advancement of Geoscience (GAGE), to identify typical user access patterns and to develop a model that can be used for data pre-fetching. Furthermore, we evaluate our data pre-fetching model and the proposed framework using a simulation of the Virtual Data Collaboratory (VDC) platform that provides in-network data staging and processing capabilities. The results demonstrate that the ability of the framework to significantly improve data delivery performance and reduce network traffic at the observatories’ facilities.

With the advancement of manufacturing technology and the diversification of customer needs, product sales forecasting is critical to the company in a small-volume, multivariety production model. Based on the traditional ARIMA model and the deep neural network LSTM model, this paper constructs the ARIMA-LSTM hybrid model to predict the sales of transmission parts manufacturing enterprises. The model uses the LSTM neural network to describe the linear and nonlinear components of the time series rather than the simple additive relationship in the traditional hybrid model. Therefore, the model overcomes the problems of low prediction accuracy that may exist in the traditional hybrid model. This paper selects the historical sales data of a certain manufacturer of a large transmission parts manufacturer in Sichuan, and carries out experimental verification on the constructed ARIMA-LSTM hybrid model. The results show that the model has high prediction accuracy and can provide reference for enterprise decision-making. The research work of this paper is supported by Sichuan Science & Technology Program under Grant No.2017GZ0144 & 2017GZ0047. Keywords—product sales forecasting; ARIMA model; LSTM model; ARIMA-LSTM hybrid model

Wireless Sensor Networks (WSNs) are composed of many sensor nodes using limited power resources. Therefore efficient power consumption is the most important issue in such networks. One way to reduce power consumption of sensor nodes is reducing the number of wireless communication between nodes by dual prediction. In this approach, the sink node instead of direct communication, exploits a time series model to predict local readings of sensor nodes with certain accuracy. There are different linear and non-linear models for time series forecasting. In this paper we will introduce a hybrid prediction model that is created from combination of ARIMA model as linear prediction model and neural network that is a non-linear model. Then, we will present a comparison between effectiveness of our approach and previous hybrid models. Experimental results show that the proposed method can be an effective way to reduce data transmission compared with existing hybrid models and also either of the components models used individually.

Wind speed modeling and prediction plays a critical role in wind related engineering studies. With the integration of wind energy into electricity grids, it is becoming increasingly important to obtain accurate wind speed forecasts. Accurate wind speed forecasts are necessary to schedule dispatchable generation and tariffs in the electricity market. In this paper a hybrid model named EMD-ANN for wind speed prediction is proposed based on the Empirical Mode Decomposition (EMD) and the Artificial Neural Networks (ANN) for renewable energy systems. All the models are analyzed with real data of wind speeds in Bilecik, Turkey using data measurement from the Turkish State Meteorological Service. Accuracy of the forecasting is evaluated in terms of MAE and MSE.

While lacking previous knowledge, general users often have difficulty in using nonlinear statistical time series methods (e.g., bilinear model, threshold autoregressive model, smoothing transition autoregressive model, autoregressive conditional heteroscedastic model and generalized autoregressive conditional heteroscedastic model) and spectral analytical approaches. Moreover, standard back-propagation neural networks (BPNNs) using the steepest descent algorithm have the disadvantage that they converge to local optima. To overcome the above limitations, this work develops a self-organizing polynomial neural network (SOPNN) based on the group method of data handling (GMDH) algorithm, which is a statistical learning network. The proposed SOPNN scheme can create an optimal network topology by using an evolutionary process and is easily to be used. Performance of the proposed SOPNN scheme is also measured by using the Mackey-Glass chaotic time series problem. Additionally, numerical results of the proposed SOPNN scheme are compared with those of conventional BPNN, cerebellar model articulation controller NN (CMAC NN), advanced simulated annealing-based BPNN (ASA-BPNN) and artificial immune algorithm-based BPNN (AIA-BPNN). Experimental results indicate that the optimum training and test capabilities of the proposed SOPNN are superior to those of a conventional BPNN. Furthermore, the generalization capability and computational CPU time of the proposed SOPNN scheme are competitive for those of CMAC NN, AIA-BPNN and ASA-BPNN.

We propose in this paper an original technique to predict global radiation using a hybrid ARMA/ANN model and data issued from a numerical weather prediction model (NWP). We particularly look at the multi-layer perceptron (MLP). After optimizing our architecture with NWP and endogenous data previously made stationary and using an innovative pre-input layer selection method, we combined it to an ARMA model from a rule based on the analysis of hourly data series. This model has been used to forecast the hourly global radiation for five places in Mediterranean area. Our technique outperforms classical models for all the places. The nRMSE for our hybrid model MLP/ARMA is 14.9% compared to 26.2% for the naive persistence predictor. Note that in the standalone ANN case the nRMSE is 18.4%. Finally, in order to discuss the reliability of the forecaster outputs, a complementary study concerning the confidence interval of each prediction is proposed.

We propose a time series prediction system of agricultural products price based on recurrent neural networks and their time alignment. The purpose of the system is to provide a way to make the price movements stable by sharing the information of price between producers and consumers of agricultural products. For the purpose, we designed a time alignment mechanism of recurrent neural networks to predict time series of prices, where each predicted results is evaluated not only by the accuracy but also by probability distribution of price movement (statistical characteristic). For the stability of agricultural products, price movements have the same importance as accuracy of predicted prices. This is the reason why we focus on not only the accuracy but also probability distribution. Although we show the results where only past price information is used as input of learning for the first step, the framework is designed to reflect other related information such as meteorological and social information on learning in future.

We propose a time series forecasting method for the future prices of agricultural products and present the criteria by which forecasted future time series are evaluated in the context of statistical characteristics. Time series forecasting of agricultural products has the basic importance in maintaining the sustainability of agricultural production. The prices of agricultural products show seasonality in their time series, and conventional methods such as the auto-regressive integrated moving average (ARIMA or the Box Jenkins method) have tried to exploit this feature for forecasting. We expect that recurrent neural networks, representing the latest machine learning technology, can forecast future time series better than conventional methods. The measures used in evaluating the forecasted results are also of importance. In literature, the accuracy determined by the error rate at a specific time point in the future, is widely used for evaluation. We predict that, in addition to the error rate, the criterion for conservation of the statistical characteristics of the probability distribution function from the original past time series to the future time series in the forecasted future is also important. This is because some time series have a non-Gaussian probability distribution (such as the Levy stable distribution) as a characteristic of the target system; for example, market prices on typical days change slightly, however on certain occasions, change dramatically. We implemented two methods for time series forecasting based on recurrent neutral network (RNN), one of which is called time-alignment of time point forecast (TATP), and another one is called direct future time series forecast (DFTS). They were evaluated using the two aforementioned criteria consisting of the accuracy and the conservation of the statistical characteristics of the probability distribution function. We found that after intensive training, TATP of LTSM shows superior performance in not only accuracy, but also the conservation compared to TATP of other RNNs. In DFTS, DFTS of LSTM cannot show the best performance in accuracy in RMS sense, but it shows superior performance in other criteria. The results suggest that the selection of forecasting methods depends on the evaluation criteria and that combinations of forecasting methods is useful based on the application. The advantage of our method is that the required length of time series for training is enough short, namely, we can forecast the whole cycle of future time series after training with even less than the half of the cycle, and it can be applied to the field where enough numbers of continuous data are not available.

We propose a new approach to forecasting total port container throughput: to generate forecasts based on each of the port’s terminals and aggregate them into the total throughput forecast. We forecast the demand for total container throughput at the Indonesia’s largest seaport Tanjung Priok Port, employing SARIMA, the additive and multiplicative Seasonal Holt-Winters (MSHW) and the Vector Error Correction Model (VECM) on the monthly port and individual terminal container throughput time series between 2003 and 2013. The performance of forecasting models is evaluated based on mean absolute error and root mean squared error. Our results show that the MSHW model produces the most accurate forecasts of total container throughput, whereas SARIMA generates the worst in-sample model fit. The VECM provides the best model fits and forecasts for individual terminals. Our results report that the total container throughput forecasts based on modelling the total throughput time series are consistently better than those obtained by combining those forecasts generated by terminal-specific models. The forecasts of total throughput until the end of 2018 provide an essential insight into the strategic decision-making on the expansion of port’s capacity and construction of new container terminals at Tanjung Priok Port.

We develop multiple resolution river flow time series model using machine learning methods at three locations in South Australia. In multiple resolution river flow models, we identify the best method from a set of widely used machine learning methods. We also identify optimized lag values of river flow and rainfall as input time series in predicting river flow multiple days ahead. The best models are ranked based on mean absolute error. Experimental results demonstrate that M5P method provides the lowest error over artificial neural network, linear regression and support vector regression in modelling river flow for multiple days ahead prediction for all three locations. Although M5P gives better accuracy over other methods for these locations as also found in the recent research in hydrological time series modelling, it may not be the best method for other geographical locations. Detailed evaluation of statistical and machine learning methods may be needed in predicting river flow for any location of interest.

We applied the Box-Jenkins time series model and artificial neural network (ANN) in the framework of a multilayer perceptron (MLP) to predict the total dissolved solids (TDS) in the Zāyandé-Rūd River, Esfahan province, Iran. The MLP inputs were total hardness (TH), bicarbonate (HCO3 ), sulfate (SO4 2 ), chloride (Cl ), Sodium (Naþ), and Calcium (Ca2þ), which were monitored over 9 years by the Esfahan Water Authority. The Autoregressive Integrate Moving Average (ARIMA) (2, 0, 3) (2, 0, 2) time series model with the lowest Akaike factor was selected. The coefficient of determination (R) and index of agreement (IA) between the measured and predicted data of the ARIMA (2, 0, 3) (2, 0, 2) time series model were 0.78 and 0.73, respectively. Using Tansig transfer functions, the Levenberg-Marquardt algorithm for training and an MLP neural network with 10 neurons in a hidden layer were developed. R and IA between the measured and predicted data were 0.94 and 0.91, respectively. Consequently, the results of the MLP were more reliable than the Box-Jenkins time series to predict TDS in the river. doi: 10.2166/aqua.2018.108 ://iwaponline.com/aqua/article-pdf/67/7/673/511999/jws0670673.pdf G. Asadollahfardi (corresponding author) H. Zangooi M. Tayebi Jebeli M. Meshkat-Dini N. Roohani Department of Civil Engineering, Faculty of Technical and Engineering, Kharazmi University, 43 Mofateh Avenue, Tehran, Iran E-mail: fardi@khu.ac.ir M. Asadi Department of Mechanical Engineering, University of Saskatchewan, Saskatoon, Canada

Wavelet artificial neural networks and remote sensing techniques can be used to estimate water quality variables such as Chlorophyll-a, turbidity and suspended solids. This paper describes empirical algorithms for the estimation of these variables incorporating information from the Operational Land Imager Sensor on board the Landsat-8 satellite. Neural networks are seasonally trained using data from the Cefni reservoir (Anglesey, U.K.), covering a variety of physical trophic status. Chlorophyll-a levels and the suspended solids are estimated from the reflectance in band-2 and band-4, while the turbidity values from reflectance in band-4. Experimental results show the potential of Landsat-8 as a substitute of Landsat-7 in water bodies quality monitoring. Moreover, predicted values obtained by using wavelet artificial neural networks fit well measured data and hence, such models provide accurate results therefore improving the efficiency in monitoring water quality parameters and contributing to possible decision making processes in the environmental management.

Water quality prediction plays an important role in environmental monitoring, ecosystem sustainability, and aquaculture. Traditional prediction methods cannot capture the nonlinear and non-stationarity of water quality well. In recent years, the rapid development of artificial neural networks (ANNs) has made them a hotspot in water quality prediction. We have conducted extensive investigation and analysis on ANN-based water quality prediction from three aspects, namely feedforward, recurrent, and hybrid architectures. Based on 151 papers published from 2008 to 2019, 23 types of water quality variables were highlighted. The variables were primarily collected by the sensor, followed by specialist experimental equipment, such as a UV-visible photometer, as there is no mature sensor for measurement at present. Five different output strategies, namely Univariate-Input-Itself-Output, Univariate-Input-Other-Output, Multivariate-Input-Other(multi), Multivariate-Input-Itself-Other-Output, and Multivariate-Input-Itself-Other (multi)-Output, are summarized. From results of the review, it can be concluded that the ANN models are capable of dealing with different modeling problems in rivers, lakes, reservoirs, wastewater treatment plants (WWTPs), groundwater, ponds, and streams. The results of many of the review articles are useful to researchers in prediction and similar fields. Several new architectures presented in the study, such as recurrent and hybrid structures, are able to improve the modeling quality of future development.

Water quality prediction is the basis of water environmental planning, evaluation, and management. In this work, a novel intelligent prediction model based on the fuzzy wavelet neural network (FWNN) including the neural network (NN), the fuzzy logic (FL), the wavelet transform (WT), and the genetic algorithm (GA) was proposed to simulate the nonlinearity of water quality parameters and water quality predictions. A self-adapted fuzzy - means clustering was used to determine the number of fuzzy rules. A hybrid learning algorithm based on a genetic algorithm and gradient descent algorithm was employed to optimize the network parameters. Comparisons were made between the proposed FWNN model and the fuzzy neural network (FNN), the wavelet neural network (WNN), and the neural network (ANN). The results indicate that the FWNN made effective use of the self-adaptability of NN, the uncertainty capacity of FL, and the partial analysis ability of WT, so it could handle the fluctuation and the nonseasonal time series data of water quality, while exhibiting higher estimation accuracy and better robustness and achieving better performances for predicting water quality with high determination coefficients over 0.90. The FWNN is feasible and reliable for simulating and predicting water quality in river.

Water quality prediction is an effective method for managing and protecting water resources by providing an early warning against water quality deterioration. In general, the existing water quality prediction methods are based on a single shallow model which fails to capture the long-term dependence in historical time series and is more likely to cause a high rate of false alarms and false negatives in practical water monitoring application. To resolve these problems, a new model combining recurrent neural network (RNN) with improved Dempster/Shafer (D-S) evidence theory (RNNs-DS) is proposed in this paper. Among them, the RNNs which can handle the long-term dependence in historical time series effectively are used to realize the preliminary prediction of water quality. And the improved D-S evidence theory is used to synthesize the prediction results of RNNs. In addition, an improved strategy based on correlation analysis method is presented for evidence theory to obtain the number of evidence, which reduces uncertainty in evidence selection effectively. Besides, a new basic probability assignment function which based on modified softmax function is proposed. The new function can effectively solve the problems of weight allocation failure in the traditional function. Then, data about permanganate index, pH, total phosphorus, and dissolved oxygen from Jiuxishuichang monitoring station near Qiantang River, Zhejiang Province, China is used to verify the proposed model. Compared with support vector regression (SVR) and backpropagation neural network (BPNN) and three RNN models, the new model shows higher accuracy and better stability as indicated by four indices. Finally, the engineering application of the RNNs-DS algorithm has been realized on the self-developed water environmental monitoring and forecasting system, which can provide effective support for early risk assessment and prevention in water environment.

Water quality prediction has a great significance to evalute the quality development trend of water and make the planning of water processing. In the study, a novel hybrid method of grey model and support vector regression is proposed to improve the prediction accuracy of sypport vector regression. COD is the important composition in the polluting water. Thus, COD is used as evaluation index of polluting water in the paper. The quality prediction values of input and output water of BP nerual network and the hybrid method of grey model and support vector regression are computated respectively. It is indicated that the water quality prediction by using the hybrid method of grey model and support vector regression is superior to BP nerual network.

Water quality monitoring using Wireless Sensor Networks (WSNs) is essential in aquaculture water quality management. In the field of water quality monitoring, dissolved oxygen (DO) is a key parameter, and its prediction can provide decision support for aquaculture production, thereby reducing farming risk. However, it is difficult to build a precise prediction model, and existing methods of DO prediction neglect the importance of analyzing DO content. To address this problem, this study proposes a hybrid DO prediction model, named KIG-ELM, which is composed of K-means, improved genetic algorithm (IGA), and extreme learning machine (ELM). This model is based on edge computing architecture, in which data acquisition, processing and dissolved oxygen prediction are distributed in sensing nodes, routing nodes and server respectively. Sensing technique and clustering operation are applied in the process of data acquisition and processing. Meanwhile, an optimized extreme learning machine is implemented for DO prediction. We evaluate the efficiency and accuracy of proposed prediction approach in a practical aquaculture on massive water quality data. Experimental results show that the hybrid model achieves significant prediction results and can meet the needs of practical production and management.

Water quality monitoring (WQM) systems seek to ensure high data precision, data accuracy, timely reporting, easy accessibility of data, and completeness. The conventional monitoring systems are inadequate when used to detect contaminants/pollutants in real time and cannot meet the stringent requirements of high precision for WQM systems. In this work, we employed the different types of wireless sensor nodes to monitor the water quality in real time. Our approach used an energy-efficient data transmission schedule and harvested energy using solar panels to prolong the node lifetime. The study took place at the Weija intake in the Greater Accra Region of Ghana. The Weija dam intake serves as a significant water source to the Weija treatment plant which supplies treated water to the people of Greater Accra and parts of Central regions of Ghana. Smart water sensors and smart water ion sensor devices from Libelium were deployed at the intake to measure physical and chemical parameters. The sensed data obtained at the central repository revealed a pH value of 7. Conductivity levels rose from 196 S/cm to 225 S/cm. Calcium levels rose to about 3.5 mg/L and dropped to about 0.16 mg/L. The temperature of the river was mainly around 35°C to 36°C. We observed fluoride levels between 1.24 mg/L and 1.9 mg/L. The oxygen content rose from the negative DO to reach 8 mg/L. These results showed a significant effect on plant and aquatic life.