Adaptive Neural-based Fuzzy Inference System Model Research Articles

ANFIS based system identification of underactuated systems

Abstract In this work, the effectiveness of the adaptive neural based fuzzy inference system (ANFIS) in identifying underactuated systems is illustrated. Two case studies of underactuated systems are used to validate the system identification i. e., linear inverted pendulum (LIP) and rotary inverted pendulum (RIP). Both the systems are treated as benchmark systems in modeling and control theory for their inherit nonlinear, unstable, and underactuated behavior. The systems are modeled with ANFIS using the input-output data acquired from the dynamic response of the nonlinear analytical model of the systems. The dynamic response of the ANFIS model is simulated and compared to the nonlinear mathematical model of the inverted pendulum systems. In order to check the effectiveness of the ANFIS model, mean square error is used as the performance index. From the obtained simulation results, it has been perceived that the ANFIS model performed satisfactorily within the trained operating range while a minor deviation is seen outside the trained operating region for both the case studies. Furthermore, the experimental validation of the of the proposed ANFIS model is done by comparing it with the experimental model of the rotary inverted pendulum. The obtained results show that the response of ANFIS model is in close agreement to the experimental model of the rotary inverted pendulum.

Modelling water quantity parameters using Artificial Intelligence techniques, A case study Abu-Ziriq Marsh in south of Iraq.

The low water quantity has a significant impact on the ecosystem and the food chain of living organisms, thus causing a loss of biodiversity and a lack of natural food sources. Abu-Ziriq Marsh, located in the south of Iraq, is chosen as the case study for the application of the proposed methodology. The aim of this study was to assess the ability of using three different models of Artificial Intelligence (AI) techniques: Adaptive Neural-based Fuzzy Inference System (ANFIS), Artificial Neural Networks (ANN) and Multiple Regression Model (MLR) to predict and estimate the discharge of Abu-Ziriq Marsh by depending on flow release from upstream Al-Badaa regulator. Daily discharge of Al-Badaa regulator(QB ) and Abu-Ziriq Marsh(Qz ) were used in this study. The water quantity data, consisting of 720 records of daily data between the years 2017 and 2018, were used for training and testing the models. The training and testing data were randomly partitioned into 515 (70.5 %) and 215 (29.5 %) datasets, respectively. The performance of all models was assessed through the correlation coefficient (R), root mean square error (RMSE) and Nash–Sutcliffe efficiency coefficient (NSE). Results of RMSE, R and NSE for the calibration (validation) of ANFIS model were 4.11 (4.17), 0.87 (0.83) and 0.76 (0.70), respectively. The evaluation of the results indicates that ANFIS model is superior to other models. The identified ANFIS models can be used as tools for the computation of water quantity parameter(Qz ) in Iraqi Marshes.

Modeling Water Quality Parameters Using Data-Driven Models, a Case Study Abu-Ziriq Marsh in South of Iraq

Total dissolved solids (TDS) and electrical conductivity (EC) are important parameters in determining water quality for drinking and agricultural water, since they are directly associated to the concentration of salt in water and, hence, high values of these parameters cause low water quality indices. In addition, they play a significant role in hydrous life, effective water resources management and health studies. Thus, it is of critical importance to identify the optimum modeling method that would be capable to capture the behavior of these parameters. The aim of this study was to assess the ability of using three different models of artificial intelligence techniques: Adaptive neural based fuzzy inference system (ANFIS), artificial neural networks (ANNs) and Multiple Regression Model (MLR) to predict and estimate TDS and EC in Abu-Ziriq marsh south of Iraq. As so, eighty four monthly TDS and EC values collected from 2009 to 2018 were used in the evaluation. The collected data was randomly split into 75% for training and 25% for testing. The most effective input parameters to model TDS and EC were determined based on cross-correlation test. The three performance criteria: correlation coefficient (CC), root mean square error (RMSE) and Nash–Sutcliffe efficiency coefficient (NSE) were used to evaluate the performance of the developed models. It was found that nitrate (NO3), calcium (Ca+2), magnesium (Mg+2), total hardness (T.H), sulfate (SO4) and chloride (Cl−1) are the most influential inputs on TDS. While calcium (Ca+2), magnesium (Mg+2), total hardness (T.H), sulfate (SO4) and chloride (Cl−1) are the most effective on EC. The comparison of the results showed that the three models can satisfactorily estimate the total dissolved solids and electrical conductivity, but ANFIS model outperformed the ANN and MLR models in the three performance criteria: RMSE, CC and NSE during the calibration and validation periods in modeling the two water quality parameters. ANFIS is recommended to be used as a predictive model for TDS and EC in the Iraqi marshes.

DALGACIK-ADAPTIF AĞ TEMELLI BULANIK ÇIKARIM SISTEMLERI ILE DALAMAN ÇAYI AKIMLARININ MODELLENMESİ ÜZERİNE BİR ÇALIŞMA

The object of the study is to investigate a flow estimation model by using a combination of Wavelet Transform Technique (W) and Adaptive Neural Based Fuzzy Inference System (ANFIS). Many models has been applied in recent years for the prediction of Dalaman Stream flow in the south of Turkey. One of these studies was AR-ANFIS models which developed by Taylan (2008), its training data set was extended with synthetic series produced by autoregressive processes. In this study, W-ANFIS models were developed with sub-series generated by wavelet analysis by using extended training set of ANFIS models. It is seen that increasing the number of input data in training increases model performance. Compared with the developed models, it has been shown that the W-ANFIS hybrid models have a better predictive power than the AR-ANFIS models. Consequently, the W-ANFIS hybrid model could be used successfully in predicting of flow.

Application of an Adaptive Neural-Based Fuzzy Inference System Model for Predicting Leaf Area

ABSTRACTLeaf Area (LA) is a key index of plant productivity and growth. A multiple linear regression technique is commonly applied to estimate LA as a non-destructive and quick method, but this technique is limited under the realistic situation. Thus, it is indispensable to elaborate new models for estimation. In this research, the performance of the Adaptive Neural-Based Fuzzy Inference System (ANFIS) in predicting the LA of 61 plant species (C) was investigated. Four parameters including leaf length (L), leaf width (W), C, and specific coefficient (K) for each plant were selected as input data to the ANFIS model and the LA as the output. Seven different ANFIS models including different combinations of input data were constructed to reveal the sensitivity analysis of the models. The normalized root mean square error (NRMSE), mean residual error (MRE), and linear regression were applied between observed LA and estimated LA by the models. The results indicated that ANFIS4-K2min which employed all input data was the most accurate (NRMSE = 0.046 and R2 = 0.997) and ANFIS1 which employed only the K input was the worst (NRMSE = 0.452 and R2 = 0.778). In ranking, ANFIS4-K2ave, ANFIS4-K1min, ANFIS4-K1ave, ANFIS3, and ANFIS2 ranked second, third, fourth, fifth, and sixth, respectively. The sensitivity analysis indicated that the predicted LA is more sensitive to the K, followed by L, W, and C. The results displayed that estimations are slightly overestimated. This study demonstrated that the ANFIS model could be accurate and faster alternative to the available laborious and time-consuming methods for LA prediction.

A comparative study between popular statistical and machine learning methods for simulating volume of landslides

This study attempts to compare popular statistical methods (linear, logarithmic, quadratic, power and exponential functions) with machine learning methods (multi-layer perceptron (MLP), radial base function (RBF), adaptive neural-based fuzzy inference system (ANFIS) and support vector machine (SVM)) for simulating the volume of landslides based on their surface area (VL~AL) in the Kurdistan province, Iran. Performances of the models were validated using some commonly error functions including the Adjusted R2, F-test and AIC (Akaike Information Criteria). The results showed that the power model demonstrates the best performance compared to other statistical methods whereas the ANFIS model outperforms other machine learning approaches. Furthermore, the comparative results showed that machine learning methods indicate better performances than simple statistical methods for simulating the volume of landslides in the study area. In practice, the outputs of this research can help managers and investigators decrease the cost of field surveys and measurements of volumes of landslides in landslide hazard management projects.

Evaluation of adaptive neural-based fuzzy inference system approach for estimating saturated soil water content

The saturated soil water content (θ s ) is an important parameter in hydrological studies. In this paper, adaptive neural-based fuzzy inference system (ANFIS) was used for estimation of soil saturation percentage of some flood spreading areas in Iran. Soil particle size distribution (sand%, silt%, and clay%), bulk density and medium porosity (0.2–30 µm) were used to develop saturated soil water content pedotransfer functions (PTFs). Then, contributions of various member functions (MFs) were assessed on estimation of θ s . The results showed that the member function type has an important role in performance of ANFIS approach. In the present investigation, Gaussian curve (gaussmf) was found to be superior over the other MFs in estimating θ s . In all of the θ s PTFs, correlation between estimations of θ s and corresponding observations was the low. R2 values between measured and PTF-estimated θ s using ANFIS approach did not increase as some input predictors were used in the PTFs (from PTF1 to PTF5). Based on the results, it is suggested that ANFIS model can be applied for reasonable estimation of θ s and there is a need for obtain more information of the proposed approach especially for the selection of best member functions. Therefore, a good performance may be obtained when best member function would be selected in addition to the more effective PTF inputs.

Estimation of dissolved oxygen using data-driven techniques in the Tai Po River, Hong Kong

This study investigates the applicability of multilinear regression (MLR), adaptive neural-based fuzzy inference system (ANFIS) and artificial neural networks (ANN) methods from data-driven techniques in estimation of the dissolved oxygen (DO), which is an important parameter in water quality, aquatic life, efficient water management and health care planning studies. The measured parameters covering 21 years of sample data for the years 1991–2011 in the Tai Po River, Hong Kong, were used to develop the models. The input parameters used to estimate DO are chloride (Cl), pH value (pH), electrical conductivity, temperature (Temp), nitrite nitrogen (NO2-N), nitrate nitrogen (NO3-N), ammonia nitrogen (NH4-N) and total phosphorous (T-P). The performance of the developed models was evaluated through the three performance criteria: correlation coefficient, root mean square error and the Nash–Sutcliffe efficiency coefficient. When the results of the developed models were compared with DO measurements using performance criteria, the ANN model shows better performance than the MLR and ANFIS models in estimation of DO concentration. Also, a sensitivity analysis was carried out to evaluate the relative importance of the input parameters in estimation of the DO. The most effective input parameters are determined as Cl, Temp, NO3-N, NO2-N, NH4-N and T-P parameters, respectively. Furthermore, the pH variable has the least contribution on the ANN model.

Development of an adaptive neural-based fuzzy inference system for feeding decision-making assessment in silver perch (Bidyanus bidyanus) culture

The primary objective of this study was to develop an adaptive neural-based fuzzy inference system (ANFIS) for feeding decision-making in aquaculture. Silver perch (Bidyanus bidyanus) were raised under semi-intensive conditions in Taiwan. Because dissolved oxygen (DO) is a key factor that is helpful in detecting the appetite of fish at the initial period of the feeding activity and because the flocking and struggle behaviors of food-searching fish have a transient effect on the measurement of the DO, a simple water-reused rearing tank was prepared to measure the DO to develop a fuzzy logic controller (FLC). In the equivalent ANFIS of the FLC, two linguistic variables were used to describe the food-searching state of the fish population and establish a rule base composed of 15 rules. Furthermore, an alternate hybrid learning approach, which is a fuzzy logic technique based on artificial neural networks, was suggested to quickly model the linguistic variables and evaluate their relative contributions. The results indicated that a decision threshold of 0.17, which was inferred using the fuzzy logic approach, considerably benefits the feeding decision; the high rate of accurate judgments (with an accuracy of 97.89%), which was obtained by the ANFIS model, was close to the actual food searching behaviors of fish. Therefore, the application of the ANFIS model to the feeding decision system in an aquaculture rearing tank has considerable potential for success.

Artificial neural network modeling of dissolved oxygen in reservoir

The water quality of reservoirs is one of the key factors in the operation and water quality management of reservoirs. Dissolved oxygen (DO) in water column is essential for microorganisms and a significant indicator of the state of aquatic ecosystems. In this study, two artificial neural network (ANN) models including back propagation neural network (BPNN) and adaptive neural-based fuzzy inference system (ANFIS) approaches and multilinear regression (MLR) model were developed to estimate the DO concentration in the Feitsui Reservoir of northern Taiwan. The input variables of the neural network are determined as water temperature, pH, conductivity, turbidity, suspended solids, total hardness, total alkalinity, and ammonium nitrogen. The performance of the ANN models and MLR model was assessed through the mean absolute error, root mean square error, and correlation coefficient computed from the measured and model-simulated DO values. The results reveal that ANN estimation performances were superior to those of MLR. Comparing to the BPNN and ANFIS models through the performance criteria, the ANFIS model is better than the BPNN model for predicting the DO values. Study results show that the neural network particularly using ANFIS model is able to predict the DO concentrations with reasonable accuracy, suggesting that the neural network is a valuable tool for reservoir management in Taiwan.

Modeling Pb (II) adsorption from aqueous solution by ostrich bone ash using adaptive neural-based fuzzy inference system

To evaluate the performance of Adaptive Neural-Based Fuzzy Inference System (ANFIS) model in estimating the efficiency of Pb (II) ions removal from aqueous solution by ostrich bone ash, a batch experiment was conducted. Five operational parameters including adsorbent dosage (Cs), initial concentration of Pb (II) ions (Co), initial pH, temperature (T) and contact time (t) were taken as the input data and the adsorption efficiency (AE) of bone ash as the output. Based on the 31 different structures, 5 ANFIS models were tested against the measured adsorption efficiency to assess the accuracy of each model. The results showed that ANFIS5, which used all input parameters, was the most accurate (RMSE = 2.65 and R2 = 0.95) and ANFIS1, which used only the contact time input, was the worst (RMSE = 14.56 and R2 = 0.46). In ranking the models, ANFIS4, ANFIS3 and ANFIS2 ranked second, third and fourth, respectively. The sensitivity analysis revealed that the estimated AE is more sensitive to the contact time, followed by pH, initial concentration of Pb (II) ions, adsorbent dosage, and temperature. The results showed that all ANFIS models overestimated the AE. In general, this study confirmed the capabilities of ANFIS model as an effective tool for estimation of AE.

Modeling for pavement roughness using the ANFIS approach

The term ''present serviceability'' was adopted to represent the momentary ability of pavement to serve traffic, and the performance of the pavement was represented by its serviceability history in conjunction with its load application history. Serviceability was found to be influenced by longitudinal and transverse profile as well as the extent of cracking and patching. The amount of weight that should be assigned to each element in the determination of overall serviceability is a matter of subjective opinion. In this study, an Adaptive Neural-Based Fuzzy Inference System (ANFIS) method is used in modeling the International Roughness Index (IRI) of flexible pavements. Data from the LTPP IMS database, namely, age, cumulative Equivalent Single Axle Loads (ESALs), and Structure Number (SN) were used in the modeling. Results showed that the ANFIS model is successful for the estimation of IRI, and this model can be easily applied in different regions. The model can be further developed by combining expert judgment and newly measured data.

Daily pan evaporation estimation using gene expression programming and adaptive neural-based fuzzy inference system

This study was conducted using gene expression programming (GEP) and an adaptive neural-based fuzzy inference system (ANFIS) as an alternative approach to estimate daily pan evaporation, which is an important parameter in hydrological and meteorological studies. The input parameters used to estimate daily pan evaporation from Lake Egirdir in the southwestern part of Turkey are the daily pan evaporation data of Lake Kovada (Kot) and Lake Karacaoren Dam (Kat) and the previous 1-, 2-, and 3-day pan evaporation values of Lake Egirdir. The various input combinations were tried by using pan evaporation data for the years 1998–2005. The GEP model with the highest Nash–Sutcliffe efficiency and the lowest mean square error have the daily pan evaporation data of Lake Kovada (Kot) and Lake Karacaoren Dam (Kat) and the previous 1-day pan evaporation values of Lake Egirdir. The NSE of the best GEP model was obtained as 0.729, 0.722, and 0.701 for training, testing, and validation sets, respectively. Furthermore, the ANFIS models were developed using the same input combinations. It was seen that the GEP model was more superior to the ANFIS model.

Modelling nitrate concentration of groundwater using adaptive neural-based fuzzy inference system

High nitrate concentration in groundwater is a major problem in agricultural areas in Iran. Nitrate pollution in groundwater of the particular regions in Isfahan province of Iran has been investigated. The objective of this study was to evaluate the performance of Adaptive Neural-Based Fuzzy Inference System (ANFIS) for estimating the nitrate concentration. In this research, 175 observation wells were selected and nitrate, potassium, magnesium, sodium, chloride, bicarbonate, sulphate, calcium and hardness were determined in groundwater samples for five consecutive months. Electrical conductivity (EC) and pH were also measured and the sodium absorption ratio (SAR) was calculated. The five-month average of bicarbonate, hardness, EC, calcium and magnesium are taken as the input data and the nitrate concentration as the output data. Based on the obtained structures, four ANFIS models were tested against the measured nitrate concentrations to assess the accuracy of each model. The results showed that ANFIS1 was the most accurate (RMSE = 1.17 and R<sup>2</sup> = 0.93) and ANFIS4 was the worst (RMSE = 2.94 and R<sup>2</sup> = 0.68) for estimating the nitrate concentration. In ranking the models, ANFIS2 and ANFIS3 ranked the second and third, respectively. The results showed that all ANFIS models underestimated the nitrate concentration. In general, the ANFIS1 model is recommendable for prediction of nitrate level in groundwater of the studied region.

Hydrological time series modeling: A comparison between adaptive neuro-fuzzy, neural network and autoregressive techniques

Summary Time series modeling is necessary for the planning and management of reservoirs. More recently, the soft computing techniques have been used in hydrological modeling and forecasting. In this study, the potential of artificial neural networks and neuro-fuzzy system in monthly reservoir inflow forecasting are examined by developing and comparing monthly reservoir inflow prediction models, based on autoregressive (AR), artificial neural networks (ANNs) and adaptive neural-based fuzzy inference system (ANFIS). To take care the effect of monthly periodicity in the flow data, cyclic terms are also included in the ANN and ANFIS models. Working with time series flow data of the Sutlej River at Bhakra Dam, India, several ANN and adaptive neuro-fuzzy models are trained with different input vectors. To evaluate the performance of the selected ANN and adaptive neural fuzzy inference system (ANFIS) models, comparison is made with the autoregressive (AR) models. The ANFIS model trained with the input data vector including previous inflows and cyclic terms of monthly periodicity has shown a significant improvement in the forecast accuracy in comparison with the ANFIS models trained with the input vectors considering only previous inflows. In all cases ANFIS gives more accurate forecast than the AR and ANN models. The proposed ANFIS model coupled with the cyclic terms is shown to provide better representation of the monthly inflow forecasting for planning and operation of reservoir.

A model of adaptive neural-based fuzzy inference system (ANFIS) for prediction of friction coefficient in open channel flow

Friction coefficient factor in free surface smooth channel with rectangular cross-section is generally affected by Reynolds number and wall roughness and can be determined with laboratory or field measurements. In application, according to researchers, correct selection of friction coefficient is substantially critical to estimate hydraulic problems correctly. In this paper an ANFIS is set up in which Reynolds number, velocity and discharge are used as inputs to estimate friction coefficients of an open channel flow (FC). By using experimental data from the laboratory, learning algorithm and training are applied according to ANFIS model. As a result, simulation results are compared with experimental friction coefficient results. A good correlation is obtained between the experimental data and predicted results. It is shown that when provided with correct and sufficient samples, ANFIS model can be used to predict the non-linear relationship between friction coefficient and the factors which affect it. It is concluded that, in practice, ANFIS model can be used as a suitable and effective method and general hydraulic problems which are mostly based on laboratory tests can be analyzed with ANFIS model. Key words: Open channel, friction coefficient, ANFIS.

Estimation of daily pan evaporation using adaptive neural-based fuzzy inference system

The main aim of this research was to estimate daily pan evaporation (E P ) using adaptive neural-based fuzzy inference system (ANFIS) method in the semi-arid region of Iran. The daily climatic data of the five gauging stations which are located in Isfahan Province, including maximum and minimum temperature, maximum and minimum relative humidity, wind speed and sunny hours are introduced as input data and pan evaporation as output data. Based on the obtained structures, three ANFIS models have been tested against the measured pan evaporation to assess the accuracy of each model. The best estimation pan evaporation values is an ANFIS1 model using five input parameters, including maximum and minimum temperature, mean relative humidity, wind speed and sunny hours were obtained with RMSE = 0.72 mm/day, R 2 = 0.95. In ranking the models, ANFIS2 and ANFIS3 ranked in a second and third place, respectively. Thus, in this study, the ANFIS1 model works well for the dataset used.

Meteorological drought analysis using data-driven models for the Lakes District, Turkey

Droughts may be classified as meteorological, hydrological or agricultural. When meteorological drought appears in a region, agricultural and hydrological droughts follow. In this study, the standardized precipitation index (SPI) was applied for meteorological drought analysis at nine stations located around the Lakes District, Turkey. Analyses were performed on 3-, 6-, 9- and 12-month-long data sets. The SPI drought classifications were modelled by Adaptive Neural-Based Fuzzy Inference System (ANFIS) and Fuzzy Logic, which has the advantage that, in contrast to most of the time series modelling techniques, it does not require the model structure to be known a priori. Comparison of the observed values and the modelling results shows a better agreement with SPI-12 and ANFIS models than with fuzzy logic models.

Estimation of Saturation Percentage of Soil Using Multiple Regression, ANN, and ANFIS Techniques

The saturation percentage (SP) of soils is an important index in hydrological studies. In this paper, arti?cial neural networks (ANNs), multiple regression (MR), and adaptive neural-based fuzzy inference system (ANFIS) were used for estimation of saturation percentage of soils collected from Boukan region in the northwestern part of Iran. Percent clay, silt, sand and organic carbon (OC) were used to develop the applied methods. In additions contributions of each input variable were assessed on estimation of SP index. Two performance functions, namely root mean square errors (RMSE) and determination coefficient (R2), were used to evaluate the adequacy of the models. ANFIS method was found to be superior over the other methods. It is, then, proposed that ANFIS model can be used for reasonable estimation of SP values of soils.

Artificial Models for Interbasin Flow Prediction in Southern Turkey

The aim of this study was to develop an optimum flow prediction method, based on the adaptive neural-based fuzzy inference system (ANFIS) and artificial neural network (ANN). Each methodology was applied to river flow predicting in Manavgat Stream in the southern part of Turkey. In application, Manavgat Stream flows were predicted from Dalaman Stream, Alara Stream, and Göksu Stream flows. Each stream is located in different catchments. For monthly streamflow predictions, data were taken from the General Directorate of Electrical Power Resources Survey and Development Administration. Used data covered a 35-year period (1969–2003) for monthly streamflows. The ANFIS and ANN models had only one output with three input variables. Comparison of the ANFIS and ANN models showed a better agreement between the ANFIS model estimations and measurements of monthly flows than ANN. With the help of the ANFIS model for interbasin flow prediction, it was possible to estimate missing or unmeasured data.