Hybrid Adaptive Neuro-fuzzy Inference Research Articles

Soil moisture simulation using hybrid artificial intelligent model: Hybridization of adaptive neuro fuzzy inference system with grey wolf optimizer algorithm

Accurate estimation of soil moisture content is necessary for optimal management of water and soil resources. Soil moisture is an important variable in the hydrologic cycle, which plays an important role in the global water and energy balance due to its impact on hydrological, ecological, and meteorological processes. The purpose of the present study was to explore a newly developed hybrid intelligent model for simulating soil moisture content. A hybrid adaptive neuro fuzzy inference system (ANFIS) model-grey wolf optimization (GWO) algorithm was designed here and validated against the neural network (ANN), support vector regression (SVR) and standalone ANFIS models. The models input parameters were di-electric constant, soil bulk density, clay content and organic matter of 1155 soil samples. Various statistical indices were employed to evaluate the performances of the applied models. For a reliable ranking of models, the Global Performance Indicator (GPI) was utilized, which is a 5-agent index. The results evidenced the feasibility of the developed hybrid ANFIS-GWO model with superior simulation results. At the testing stage, the MAE and SI values for the ANFIS-GWO were 1.468% and 0.098, respectively, which indicated the superiority of the ANFIS-GWO compared to the ANFIS-Fuzzy C mean (MAE = 6.427%, SI = 0.354), and ANFIS-sub clustering (MAE = 2.137%, SI = 0.141) models. Based on the GPI, the ANFIS-GWO model was ranked as the best model, followed by the standalone ANFIS and SVR models, while the worst accuracy was attained through ANN model. The ANFIS-GWO model improved the simulation accuracy by 48% and 50%, respectively, compared to the standalone ANFIS and SVR models. In addition, based on the GPI, the ANFIS-GWO model presented an enhancement of about 77 percent compared to the ANN model. The high accuracy of the ANFIS-GWO model compared to the standalone ANFIS model represents the performance of the GWO algorithm for escaping local optima, which makes the ANFIS-GWO as a powerful tool for estimating soil moisture. Overall, the explored hybrid intelligent models demonstrated a reliable pedotransfer function of soil moisture where it can contribute to several geo-sciences engineering principles.

Forecasting Copper Prices Using Hybrid Adaptive Neuro-Fuzzy Inference System and Genetic Algorithms

An accurate forecasting model for the price volatility of minerals plays a vital role in future investments and decisions for mining projects and related companies. In this paper, a hybrid model is proposed to provide an accurate model for forecasting the volatility of copper prices. The proposed model combines the adaptive neuro-fuzzy inference system (ANFIS) and genetic algorithm (GA). Genetic algorithms are used for estimating the ANFIS model parameters. The results of the proposed model are compared to other models, including ANFIS, support vector machine (SVM), generalized autoregressive conditional heteroscedasticity (GARCH), and autoregressive integrated moving average (ARIMA) models. The empirical results confirm the superiority of the hybrid GA–ANFIS model over other models. The proposed model also improves the forecasting accuracy obtained from the ANFIS, SVM, GARCH, and ARIMA models by a 62.92%, 36.38%, 91.72%, and 42.19% decrease in mean square error, respectively.

EXPLORING THE NATURE OF CREDIT SCORING A NEURO FUZZY APPROACH

Neural Network (NN) is one of the most commonly used method in credit scoring. On the other hand Fuzzy Logic (FL) cannot be used alone in credit scoring due to lack of learning ability. However this is possible in hybrid models where NN and FL are used together. This study investigates the possible nonlinear relationship between the characteristics of loan applicant and default probability by using hybrid Adaptive Neuro Fuzzy Inference System (ANFIS) which combines NN and FL. We used statistical techniques such as discriminant analysis, logistic regression and multivariate adaptive regression splines as feature selection before developing credit scoring models. Then we developed 83 credit scoring models by using NN and 15 credit scoring models by using ANFIS. While the optimal ANFIS model has 78.5% classification accuracy, NN has 77.5%. In addition to higher classification accuracy the ANFIS model is not a black box, in contrast to NN. The ANFIS model can explain the reasons for the credit decision. So this study reveals the nature of personal loan decisions with the German credit data set.

A Hybrid ANFIS-ABC Based MPPT Controller for PV System With Anti-Islanding Grid Protection: Experimental Realization

This paper introduces a novel control system with maximum power point tracker (MPPT) for the photovoltaic system with grid integration. Hybrid adaptive neuro-fuzzy inference system (ANFIS) and artificial bee colony (ABC) algorithm employed to optimize the membership function. Hence, for minimizing the root mean square error (RMSE), this controls the SEPIC-based MPPT algorithm to achieve rapid PV power tracking. The system performance is improved by fuzzy logic control (FLC), which generates the switching signal to the power switches of the inverter. A dSPACE (DS1104) control board employed for experimental validation of MPPT and inverter control strategies. The novelty of the proposed hybrid MPPT controller is the optimal tuning of ANFIS membership function with the ABC algorithm and been neither discussed before for PV power applications. The experimental responses completely validate the reliability of the PV grid integration with anti-islanding protection. The recentness of this research work is PV MPPT functioning using the hybrid ANFIS-ABC-based algorithm, been not described practically by any researchers in the past works.

GA-ANFIS modeling of higher heating value of wastes: Application to fuel upgrading

ABSTRACTCombustion of wastes is a promising source for energy recovery because of having appropriate higher heating value (HHV) in order to use as fuel. The present study was aimed to estimate HHV value using a hybrid adaptive neuro-fuzzy inference system and genetic algorithm called GA-ANFIS. This model can predict HHV as a function of carbon (%C), hydrogen (%H), oxygen (%O), nitrogen (%N), and sulfur (%S) mass percentages. This suggested model has been also compared with other published correlations, and based on obtained results, great accuracy of our model was confirmed. The obtained values of Mean Squared Error (MSE) and R-squared were 0.236 and 0.9983, respectively. Consequently, this model can be very valuable to have accurate prediction of waste HHV value.

A hybrid-adaptive neuro-fuzzy inference system for multi-objective regression test suites optimization

Regression testing is a mandatory activity of software development life cycle, which is performed to ensure that modifications have not caused any adverse effects on the system’s functionality. With every change in software in the maintenance phase, the size of regression test suite grows as new test cases are written to validate changes. The bigger size of regression test suite makes the testing expensive and time-consuming. Optimization of regression test suite is a possible solution to cope with this problem. Various techniques of optimization have been proposed; however, there is no perfect solution for the problem and therefore, requires better solutions to improve the optimization process. This paper presents a novel technique named as hybrid-adaptive neuro-fuzzy inference system tuned with genetic algorithm and particle swarm optimization algorithm that is used to optimize the regression test suites. Evaluation of the proposed approach is performed on benchmark test suites including “previous date problem” and “Siemens print token.” Experimental results are compared with existing state-of-the-art techniques, and results show that the proposed approach is more effective for the reduction in a regression test suites with higher requirement coverage. The size of regression test suites can be reduced up to 48% using the proposed approach without reducing the fault detection rate.

Comparison of Three Intelligent Techniques for Runoff Simulation

In this study, performance of a feedback neural network, Elman, is evaluated for runoff simulation. The model ability is compared with two other intelligent models namely, standalone feedforward Multi-layer Perceptron (MLP) neural network model and hybrid Adaptive Neuro-Fuzzy Inference System (ANFIS) model. In this case, daily runoff data during monsoon period in a catchment located at south India were collected. Three statistical criteria, correlation coefficient, coefficient of efficiency and the difference of slope of a best-fit line from observed-estimated scatter plots to 1:1 line, were applied for comparing the performances of the models. The results showed that ANFIS technique provided significant improvement as compared to Elman and MLP models. ANFIS could be an efficient alternative to artificial neural networks, a computationally intensive method, for runoff predictions providing at least comparable accuracy. Comparing two neural networks indicated that, unexpectedly, Elman technique has high ability than MLP, which is a powerful model in simulation of hydrological processes, in runoff modeling.

Hybrid ANFIS‐GA‐based control scheme for performance enhancement of a grid‐connected wind generator

This study presents a novel application of a hybrid adaptive neuro-fuzzy inference system (ANFIS)-genetic algorithm (GA)-based control scheme to enhance the performance of a variable-speed wind energy conversion system. The variable-speed wind turbine drives a permanent-magnet synchronous generator, which is connected to the power grid through a frequency converter. A cascaded ANFIS-GA controller is introduced to control both of the generator-side converter and the grid-side inverter. ANFIS is a non-linear, adaptive, and robustness controller, which integrates the merits of the artificial neural network and the FIS. A GA-based learning design procedure is proposed to identify the ANFIS parameters. Detailed modelling of the system under investigation and its control strategies are demonstrated. For achieving realistic responses, real wind speed data extracted from Zaafarana wind farm, Egypt, are considered in the analyses. The effectiveness of the ANFIS-GA controller is compared with that obtained using optimised proportional-integral controllers by the novel grey wolf optimiser algorithm taking into consideration severe grid disturbances. The validity of the ANFIS-GA control scheme is verified by the extensive simulation analyses, which are performed using MATLAB/Simulink environment. With the ANFIS-GA controller, the dynamic and transient stability of grid-connected wind generator systems can be further enhanced.

Application of hybrid adaptive neuro-fuzzy inference system in well placement optimization

Well placement optimization as a key factor in field development, is one of the reservoir management cycles, which aims to economize oil and gas production from the reservoir. One of the reservoir development scenarios is determining parts of the reservoir in which hydrocarbons have not been drained completely. After determination of these areas, some new wells are drilled to improve oil production. In this study, a new model based on adaptive neuro-fuzzy inference system (ANFIS) for optimum well placement is proposed. Also, an application of opportunity index (OI) concept is used to develop this model. Using particle swarm optimization (PSO) and genetic algorithm (GA) as population based algorithms, the ANFIS model was optimized. In order to develop and validate the proposed AFNIS-PSO model, a large number of simulation runs were performed in synthetic models. In this study, the proposed model was applied in both synthetic and standard models to find the best location for new production wells. The Tenth SPE model as a standard model was used to evaluate the performance and accuracy of ANFIS-PSO model. The performance of new method compared to well placement optimizer algorithms, which is the combination of commercial flow simulator and evolutionary algorithms, was investigated in terms of the number of simulations and maximum net present value (NPV). Finally, the simulation results show that the proposed ANFIS model was able to achieve better results with fewer simulation runs and therefore improving the computational efficiency.

New type side weir discharge coefficient simulation using three novel hybrid adaptive neuro-fuzzy inference systems

In many hydraulic structures, side weirs have a critical role. Accurately predicting the discharge coefficient is one of the most important stages in the side weir design process. In the present paper, a new high efficient side weir is investigated. To simulate the discharge coefficient of these side weirs, three novel soft computing methods are used. The process includes modeling the discharge coefficient with the hybrid Adaptive Neuro-Fuzzy Interface System (ANFIS) and three optimization algorithms, namely Differential Evaluation (ANFIS-DE), Genetic Algorithm (ANFIS-GA) and Particle Swarm Optimization (ANFIS-PSO). In addition, sensitivity analysis is done to find the most efficient input variables for modeling the discharge coefficient of these types of side weirs. According to the results, the ANFIS method has higher performance when using simpler input variables. In addition, the ANFIS-DE with RMSE of 0.077 has higher performance than the ANFIS-GA and ANFIS-PSO methods with RMSE of 0.079 and 0.096, respectively.

Performance evaluation of hybrid adaptive neuro-fuzzy inference system models for predicting monthly global solar radiation

Solar energy plays a vital role in the field of sustainable energy by providing clean, efficient and reliable alternative source of energy. Where, the output of solar energy systems is highly dependent on the solar radiation. Thus, accurate prediction of solar radiation is considered as a very important factor for such applications. In this paper, standalone adaptive neuro-fuzzy inference system and hybrid models have been developed to predict monthly global solar radiation from different meteorological parameters such as sunshine duration S(h), and air temperature. The proposed hybrid models include particle swarm optimization, genetic algorithm and differential evolution. To evaluate the capability and efficiency of the proposed models, several statistical indicators such as; root mean square error, co-efficient of determination and mean absolute bias error are used. All prediction models’ results showed good agreements with measured datasets. The performance evaluation over different statistical indicators showed high correlation for all developed modules. Whereas, hybrid particle swarm optimization has achieved the best statistical indicators over all models in training and testing models. A detailed comparison with other studies is carried out to validate the prediction accuracy and suitability of the proposed models. The results showed that the developed hybrid models have the most reliable and accurate estimation capability and deemed to be the efficient methods for predicting global solar radiation for various applications.

Stabilization and Control of Elastic Inverted Pendulum System (EIPS) Using Adaptive Fuzzy Inference Controllers

Elastic Inverted Pendulum system (EIP) are very popular objects of theoretical investigation and experimentation in field of control engineering. The system becomes highly nonlinear and complex due to transverse displacement of elastic pole or pendulum. This paper presents a comparison study for control of EIP using fuzzy and hybrid adaptive neuro fuzzy inference system (ANFIS) controllers. Initially a fuzzy controller was designed, which was used for training and tuning of ANFIS controller using gbell shape membership functions (MFs). The performance of complete system was evaluated through output responses of settling time, steady state error and maximum overshoot. The study also highlights effect of varying number of MFs on training error of ANFIS. The results showed better performance of ANFIS controller compared to fuzzy controller.

Oil flooded scroll compressors: Predicting the energy performance and evaluating the experimental data

The annual or seasonal performance evaluation of Oil Flooded Scroll Compressor (OFSC) is essential to maximize its performance and increase vapor compression cycle efficiency. An experimental data bank for detail evaluation of this component in the vapor compression systems has need of a highly equipped laboratory and proper instruments. The previous researchers have developed different theoretical, empirical, and semi-empirical models to solve this problem, which are not highly effective under different situations. We propose a Hybrid-Adaptive Neuro Fuzzy Inference System (Hybrid-ANFIS) for the fast and precise estimation of the discharge temperature, refrigerant mass flow rate, and electrical power of the OFSCs. The performance of the Hybrid-ANFIS is checked against the other well-known models, including the Particle Swarm Optimization-Artificial Neural Network, Coupled Simulated Annealing-Least Square Support Vector Machine, Genetic Algorithm-Least Square Support Vector Machine, and the existing correlations in the open literature. The results indicated that the proposed Hybrid-ANFIS model is more accurate and has the promising potential for estimating the desired parameters by introducing a coefficient of determination higher than 0.998. The application of the suggested model is further illustrated against the experimental data and new test conditions are evaluated. Finally, the Leverage algorithm, which is a novel statistical approach, is implemented to assess the quality of the experimental data and diagnose the probable doubtful data samples. The results of the data analysis and the outlier detection method clarified that the presented Hybrid-ANFIS model is statistically valid and four experimental data samples are doubtful reported for the OFSC.

Hybrid ANFIS with ant colony optimization algorithm for prediction of shear wave velocity from a carbonate reservoir in Iran

Shear wave velocity (Vs) data are key information for petrophysical, geophysical and geomechanical studies. Although compressional wave velocity (Vp) measurements exist in almost all wells, shear wave velocity is not recorded for most of elderly wells due to lack of technologic tools. Furthermore, measurement of shear wave velocity is to some extent costly. This study proposes a novel methodology to remove aforementioned problems by use of hybrid adaptive neuro fuzzy inference system (ANFIS) with ant colony optimization algorithm (ACO) based on fuzzy c–means clustering (FCM) and subtractive clustering (SCM). The ACO is combined with two ANFIS models for determining the optimal value of its user–defined parameters. The optimization implementation by the ACO significantly improves the generalization ability of the ANFIS models. These models are used in this study to formulate conventional well log data into Vs in a quick, cheap, and accurate manner. A total of 3030 data points was used for model construction and 833 data points were employed for assessment of ANFIS models. Finally, a comparison among ANFIS models, and six well–known empirical correlations demonstrated ANFIS models outperformed other methods. This strategy was successfully applied in the Marun reservoir, Iran.

Design of a Hybrid Adaptive Neuro Fuzzy Inference System (ANFIS) Controller for Position and Angle control of Inverted Pendulum (IP) Systems

This paper illustrates a comparison study of Fuzzy and ANFIS Controller for Inverted Pendulum systems. IP belongs to a class of highly non-linear, unstable and multi-variable systems which act as a testing bed for many complex systems. Initially, a Matlab-Simulink model of IP system was proposed. Secondly, a Fuzzy logic controller was designed using Mamdani inference system for control of proposed model. The data sets from fuzzy controller was used for development of a Hybrid Sugeno ANFIS controller. The results shows that ANFIS controller provides better results in terms of Performance parameters including Settling time(sec), maximum overshoot(degree) and steady state error.

Integrating metaheuristics and ANFIS for daily mean temperature forecasting

Weather forecasting is considered as a key to successful planning for various applications such as agricultural industries. Having accurate weather forecasting allows people to make better decision on managing day to day activities. Also, it has to be underlined that forecasting is important to cope with impacts of extreme events and to adapt to climatic changes. To improve weather forecasting, we used hybrid adaptive neuro-fuzzy inference system (ANFIS), which consist in exploiting capabilities of harmony search (HS) and genetic algorithm (GA), for selecting the most relevant weather variables and simultaneously searching the most appropriate structure of ANFIS. Proposed methods are applied for six different cities of Turkey which are determined according to Aydeniz's climate classification. The results of the study showed that GA-ANFIS and HS-ANFIS yield remarkable results in daily mean temperature forecasting due to the ability of capturing the advantages of both types of methods simultaneously.

Evolutionary computing based approach for the removal of ECG artifact from the corrupted EEG signal.

Electroencephalogram (EEG) is an important tool for clinical diagnosis of brain-related disorders and problems. However, it is corrupted by various biological artifacts, of which ECG is one among them that reduces the clinical importance of EEG especially for epileptic patients and patients with short neck. To remove the ECG artifact from the measured EEG signal using an evolutionary computing approach based on the concept of Hybrid Adaptive Neuro-Fuzzy Inference System, which helps the Neurologists in the diagnosis and follow-up of encephalopathy. The proposed hybrid learning methods are ANFIS-MA and ANFIS-GA, which uses Memetic Algorithm (MA) and Genetic algorithm (GA) for tuning the antecedent and consequent part of the ANFIS structure individually. The performances of the proposed methods are compared with that of ANFIS and adaptive Recursive Least Squares (RLS) filtering algorithm. The proposed methods are experimentally validated by applying it to the simulated data sets, subjected to non-linearity condition and real polysomonograph data sets. Performance metrics such as sensitivity, specificity and accuracy of the proposed method ANFIS-MA, in terms of correction rate are found to be 93.8%, 100% and 99% respectively, which is better than current state-of-the-art approaches. The evaluation process used and demonstrated effectiveness of the proposed method proves that ANFIS-MA is more effective in suppressing ECG artifacts from the corrupted EEG signals than ANFIS-GA, ANFIS and RLS algorithm.

Adaptive neuro‐fuzzy based solar cell model

The modelling of photovoltaic (PV) solar cells using a hybrid adaptive neuro-fuzzy inference system (ANFIS) algorithm is presented. It is based on the decomposition of the cell output current into photocurrent and junction current. The photocurrent is linearly dependent on solar irradiance and cell temperature; consequently, its analytical computation is done easily. However, the junction current is highly non-linear and depends on cell voltage and temperature. Therefore, its analytical computation is complicated and the manufacturers do not supply any information about this parameter. Moreover, there is no way to measure it physically. Therefore, it is proposed to use the ANFIS algorithm as a powerful technique in order to estimate this current and reconstruct the output PV cell current using the photocurrent. The model validation is based on the gradient descent and chain rule applied to a set of data different than the one used for training process. The advantage of the proposed model is that only one climatic parameter is used as the input to the ANFIS algorithm, which makes it less sensitive to climatic variations.

An empirical comparison of conventional techniques, neural networks and the three stage hybrid Adaptive Neuro Fuzzy Inference System (ANFIS) model for credit scoring analysis: The case of Turkish credit card data

The number of Non-Performing Loans has increased in recent years, paralleling the current financial crisis, thus increasing the importance of credit scoring models. This study proposes a three stage hybrid Adaptive Neuro Fuzzy Inference System credit scoring model, which is based on statistical techniques and Neuro Fuzzy. The proposed model’s performance was compared with conventional and commonly utilized models. The credit scoring models are tested using a 10-fold cross-validation process with the credit card data of an international bank operating in Turkey. Results demonstrate that the proposed model consistently performs better than the Linear Discriminant Analysis, Logistic Regression Analysis, and Artificial Neural Network (ANN) approaches, in terms of average correct classification rate and estimated misclassification cost. As with ANN, the proposed model has learning ability; unlike ANN, the model does not stay in a black box. In the proposed model, the interpretation of independent variables may provide valuable information for bankers and consumers, especially in the explanation of why credit applications are rejected.

Modeling and prediction of activity anti-HIV molecules using soft computing techniques

Several works quantitative structure-activity relationships (QSAR) of anti-Human Immunodeficiency Virus (HIV) molecules were studied by different statistical methods and non-linear models. But few studies have used the heuristic methods. In this paper, a hybrid decision trees (DT) and adaptive neuro-fuzzy inference system (ANFIS) is used of the prediction of inhibitory activity of anti-VIH molecules. DT algorithm is utilized to select the most important variables in QSAR modeling and then these variables were used as inputs of ANFIS to predict the anti-HIV activity. The model's predictions were compared with other methods and the results indicated that the proposed models in this work is superior over the others.