Neural Network-based Fuzzy Inference System Research Articles

Predicting the Compressive Strength of Environmentally Friendly Concrete Using Multiple Machine Learning Algorithms

Machine learning (ML) algorithms have been widely used in big data prediction and analysis in terms of their excellent data regression ability. However, the prediction accuracy of different ML algorithms varies between different regression problems and data sets. In order to construct a prediction model with optimal accuracy for fly ash concrete (FAC), ML algorithms such as genetic programming (GP), support vector regression (SVR), random forest (RF), extremely gradient boost (XGBoost), backpropagation artificial neural network (BP-ANN) and adaptive network-based fuzzy inference system (ANFIS) were selected as regression and prediction algorithms in this study; the particle swarm optimization (PSO) algorithm was also used to optimize the structure and hyperparameters of each algorithm. The statistical results show that the performance of the assembled algorithms is better than that of an NN-based algorithm. In addition, PSO can effectively improve the prediction accuracy of the ML algorithms. The comprehensive performance of each model is analyzed using a Taylor diagram, and the PSO-XGBoost model has the best comprehensive performance, with R2 and MSE equal to 0.9072 and 11.4546, respectively.

Hybrid 5G Positioning Systems with Applications to Mobile Robot Technology

A new concept of 5G/INS integration of artificial intelligence-based systems is proposed, namely the Recurrent Neural Networkbased Fuzzy Inference System (RNNFIS). The 5G system with fuzzy logic was used as a reference during its availability. Data from the 5G system and the inertial navigation system (INS) are used to build a structured knowledge base consisting of the behavior of the INS system in some special scenarios of robotic technology. Using the same data, the proposed fuzzy system is trained to obtain corrected navigation data. In the absence of 5G system information, the system will only utilize data from the INS system and blur correction algorithms to perform its tasks.

Modeling the Moisture Content and Drying Rate of Zucchini (Cucurbita pepo L.) in a Solar Hybrid Dryer Using ANN and ANFIS Methods

Estimating product drying kinetics is critical to obtain the best drying process without compromising product quality and necessitates the development of numerical drying models. This research aims to compare the prediction models developed using artificial neural network (ANN) and adaptive network-based fuzzy inference system (ANFIS), two popular machine learning approaches in the recent years. Zucchini slices were chosen as samples and dried in a solar-assisted microwave belt dryer at 0.245 m/min belt speed and microwave powers of 0.7, 1, and 1.4 kW. On the data set obtained by computing the moisture content and drying rate values, prediction models were developed using ANN and ANFIS approaches. These models were evaluated using the coefficient of determination, mean absolute percent error, and root mean square error data. The ANFIS-based prediction model outperformed the ANN model in terms of drying rate performance, but the ANN model outperformed the ANFIS model in terms of moisture content values. Results showed that both methods established can be utilized to estimate zucchini slices.

The role of artificial intelligence in developing a banking risk index: an application of Adaptive Neural Network-Based Fuzzy Inference System (ANFIS).

Banking risk measurement and management remain one of many challenges for managers and policymakers. This study contributes to the banking literature and practice in two ways by (a) proposing a risk ranking index based on the Mahalanobis Distance (MD) between a multidimensional point representing a bank's risk measures and the corresponding critical ratios set by the banking authorities and (b) determining the relative importance of a bank's risk ratios in affecting its financial standing using an Adaptive Neuro-Fuzzy Inference System. In this study, ten financial ratios representing five risk areas were considered, namely: Capital Adequacy, Credit, Liquidity, Earning Quality, and Operational risk. Data from 45 Gulf banks for the period 2016-2020 was used to develop the model. Our findings indicate that a bank is in a sound risk position at the 99%, 95%, and 90% confidence level if its Mahalanobis distance exceeds 4.82, 4.28, and 4.0, respectively. The maximum distance computed for the banks in this study was 9.31; only five out of the forty-five banks were below the 4.82 and one below the 4.28 and 4.0 thresholds at 3.96. Sensitivity analysis of the risks indicated that the Net Interest Margin is the most significant factor in explaining variations in a bank's risk position, followed by Capital Adequacy Ratio, Common Equity Tier1, and Tier1 Equity in order. The remaining financial ratios: Non-Performing Loans, Equity Leverage, Cost Income Ratio, Loans to Total Assets, and Loans to Deposits have the least influence in the order given; the Provisional Loans Ratio appears to have no influence.

Prediction Analysis of Surface Roughness of Aluminum Al6061 in End Milling CNC Machine Using Soft Computing Techniques

Computer numerically controlled (CNC) milling has been one of the most commonly used manufacturing processes for the performance of multiple operations, from tiny integrated circuits to heavy-duty mining machine gearboxes. It is a well-known machining process that offers close tolerances and repeated operations. However, the choice of machining parameters to achieve a desired part’s surface roughness (SR) remains a challenge. In the present study, artificial neural network (ANN) and adaptive network-based fuzzy inference system (ANFIS) approaches have been used to predict and monitor the surface roughness of aluminum Al6061 machined blocks. Furthermore, both models have been hybridized with genetic algorithm (GA) and particle swarm optimization (PSO) to investigate the potential enhancement in the prediction performance of the hybrid approach. The results show that factors such as the population size, the acceleration values, the choice of membership functions, and the number of neurons and layers significantly influence the prediction performance of the proposed models. Through a parametric analysis, this study demonstrates how the configuration of the models could affect the prediction performance. While exhibiting the impact of models’ hyperparameter combination on the prediction ability, this study provides insight into the development of suitable prediction models and the potential of soft computing techniques to predict the surface roughness of aluminum Al6061 blocks on CNC machines.

RETRACTED ARTICLE: Artificial neural networks (ANN), MARS, and adaptive network-based fuzzy inference system (ANFIS) to predict the stress at the failure of concrete with waste steel slag coarse aggregate replacement

RETRACTED ARTICLE: Artificial neural networks (ANN), MARS, and adaptive network-based fuzzy inference system (ANFIS) to predict the stress at the failure of concrete with waste steel slag coarse aggregate replacement

Prediction of hydraulic conductivity of porous granular media by establishment of random forest algorithm

The present study investigates the possibility of using random forest (RF) algorithm for prediction of hydraulic conductivity of coarse-sized granular material. Meanwhile, the conventional soft computing methods named artificial neural network (ANN) and adaptive network-based fuzzy inference system (ANFIS) are considered to assess the robustness of the established approach. The characterized properties related to the particle size distribution as well as the air void content are utilized to predict the hydraulic conductivity coefficient. In addition, owing to the direct effect of the applied hydraulic head on the measured values of water flow velocity, the hydraulic gradient is also characterized as an influential parameter. The statistical comparison between the prediction capability of the RF and the results of ANN and ANFIS models corroborates the robustness of the developed procedure based on the machine learning algorithm to indirectly predict the hydraulic conductivity of the provided porous media. The results of parametric analysis mainly demonstrate that the air void content among aggregates is by far the most influential factor affecting the permeability of granular media made of coarse-sized aggregate.

NOISE MODELLING AND NOISE MAPPING IN MECHANIZED OPENCAST COAL MINE - A CASE STUDY

Increased mechanization has generated noise and vibration problems in opencast mines. Noise isgenerated by every activity or machinery present in the mines. Artificial Neural Network (ANN)and Adaptive Network-based Fuzzy Inference System (ANFIS) was applied to predictmachinerynoise in a mechanized opencast coal mine. This paper dealt with comparison of theprediction capabilities of ANN and ANFIS based soft computing models vis-Ã -vis existingstatistical model (ENM) in order to find the actual noise status in an opencast coal mine. From thepresent study, it could be concluded that ANFIS gave better prediction results than the ANN (MLPand BP) methods. This paper also includes the preparation of a noise map in the opencast mine usingtwo methods (Kriging and IDW) and ArcGIS.

A novel comparative model proposal for hospital pneumatic systems

Hospital pneumatic systems are of great importance in terms of efficient use of resources and early diagnosis of diseases. Each sample is transported at a constant and specific speed/pressure in existing pneumatic systems. This study aims to provide pressure/speed control according to the type of material being transported, prevent deterioration during transportation, and make a positive contribution to system users. We have proposed a new control model for hospital pneumatic systems. This hospital pneumatic system is modeled using fuzzy logic, adaptive neural network-based fuzzy inference system (ANFIS), and artificial neural network (ANN) methods. The parameters affecting the performance of the system are determined by experimental trials carried out. Fuzzy logic, ANFIS, and ANN models are compared using the regression coefficient (R2), Mean Absolute Deviation (MAD), and Root Mean Square Error (RMSE) metrics to determine the most suitable modeling method for the system. As a result of this study, carrying the pressure value (velocity) ideal for the sample type will increase the energy efficiency of the system.

A new innovative method for model efficiency performance

Abstract In every aspect of scientific research, model predictions need calibration and validation as their representativity of the record measurement. In the literature, there are a myriad of formulations, empirical expressions, algorithms and software for model efficiency assessment. In general, model predictions are curve fitting procedures with a set of assumptions that are not cared for sensitively in many studies, but only a single value comparison between the measurements and predictions is taken into consideration, and then the researcher makes the decision as for the model efficiency. Among the classical statistical efficiency formulations, the most widely used ones are bias (BI), mean square error (MSE), correlation coefficient (CC) and Nash-Sutcliffe efficiency (NSE) procedures, all of which are embedded within the visual inspection and numerical analysis (VINAM) square graph as measurements versus predictions scatter diagram. The VINAM provides a set of verbal interpretations and then numerical improvements embracing all the previous statistical efficiency formulations. The fundamental criterion in the VINAM is 1:1 (45°) main diagonal along which all visual, science philosophical, logical, rational and mathematical procedures boil down for model validation. The application of the VINAM approach is presented for artificial neural network (ANN) and adaptive network-based fuzzy inference system (ANFIS) model predictions.

Prediction of soil-bearing capacity on forest roads by statistical approaches.

The soil-bearing capacity is one of the important criteria in dimensioning the superstructure. In Turkey, predictability of California Bearing Ratio values, which may be used in the planning and dimensioning of forest roads, of which about 26% lacks the superstructure, by using soil mechanical properties (cost and time efficient parameters that are easier to determine) is investigated. Simple linear regression, multiple linear regression, artificial neural networks and adaptive network-based fuzzy inference system methods were utilized. Two hundred sixty-four California Bearing Ratio values obtained from the project carried out on the forest roads of Bartin Forest Operation Directorate were used in both the production of training-test data and the creation of models. Statistical performance of the models was assessed by means of parameters such as root-mean-square error, mean absolute error and R2. The obtained results show that the bearing capacity values predicted by artificial neural networks and adaptive network based fuzzy inference system models display significantly better performance than the simple linear regression and multiple linear regression models. While the highest prediction capacity belongs to adaptive network based fuzzy inference system (0.969-0.991), it is followed by artificial neural networks (R2 = 0.796-0.974), multiple linear regression (R2 = 0.796) and simple linear regression (R2 = 0.554). What makes the algorithms superior than the traditional statistical models is the fact that they have many processing neurons, each with local connections, and thus have higher error tolerance. On the other hand, for the forest and rural roads, which play an important role in rural development of the forest peasants, to be able to operate all-seasons, superstructure should be immediately built in order to minimize the wear on the roads.

Biogas maximization using data-driven modelling with uncertainty analysis and genetic algorithm for municipal wastewater anaerobic digestion

Anaerobic digestion processes create biogases that can be useful sources of energy. The development of data-driven models of anaerobic digestion processes via operating parameters can lead to increased biogas production rates, resulting in greater energy production, through process modification and optimization. This study assessed processed and unprocessed input operating parameter variables for the development of regression models with transparent structures (‘white-box’ models) to: (1) estimate biogas production rates from municipal wastewater treatment plant (MWTP) anaerobic digestors; (2) compare their performances to artificial neural network (ANN) and adaptive network-based fuzzy inference system (ANFIS) models with opaque structures (‘black-box’ models) using Monte Carlo Simulation for uncertainty analysis; and (3) integrate the models with a genetic algorithm (GA) to optimize operating parameters for maximization of MWTP biogas production rates. The input variables were anaerobic digestion operating parameters from a MWTP including volatile fatty acids, total/fixed/volatile solids, pH, and inflow rate, which were processed via correlation tests and principal component analysis. Overall, the results indicated that the processed data did not improve regression model performances. Additionally, the developed non-linear regression model with the unprocessed inputs had the best performance based on values including R = 0.81, RMSE = 0.95, and IA = 0.89. However, this model was less accurate, but interestingly had less uncertainty, as compared to ANN and ANFIS models which indicates the compromise between model accuracy and uncertainty. Thus, all three models were coupled with GA optimization with maximum biogas production rate estimates of 22.0, 23.1, and 28.6 m3/min for ANN, ANFIS, and non-linear regression models, respectively.

Effective modelling of hydrogen and energy recovery in microbial electrolysis cell by artificial neural network and adaptive network-based fuzzy inference system

This study aims to analyze and model cathodic H2 recovery (rcat), coulombic efficiency (CE) with inputs of voltage, electrical conductivity (EC) and anode potential, and H2 production rate and total energy recovery with inputs of rcat and CE in a microbial electrolysis cell using artificial neural network (ANN) and adaptive network-based fuzzy inference system (ANFIS) procedures. Both ANN and ANFIS models demonstrated great goodness of fit for rcat, CE, H2 production rate and total energy recovery prediction with high R2 values. The sum square error values for rcat (0.0017), CE (0.0163), H2 production rate (0.1062) and total energy recovery (0.0136) in ANN models were slightly higher than those in ANFIS models at 0.0005, 0.0091, 0.1247 and 0.0148 respectively. Sensitivity analysis by ANN models demonstrated that voltage, EC, rcat and rcat were the most effective factors for rcat, CE, H2 production rate and total energy recovery, respectively.

Data Modelling To Determine Room Rate with Adaptive Network Based Fuzzy Inference System And Particle Swarm Optimization

Determination of room rate in a hotel isinfluenced by two factors, namely internal andexternal. From an external perspective, PT. PIMhas eight competitor hotels that affect its roomrate. The Hotel Manager analyzes eachcompetitor's room rate changes to staycompetitive. Human analysis has severalshortcomings: subjectivity, fatigue andinconsistency. Then we need a decision supportor decision companion machine to determine theroom rate. ANFIS-PSO is a hybrid algorithmfrom the Adaptive neural network based fuzzyinference system (ANFIS) by utilizing ParticleSwarm Optimization (PSO) optimization.Traditional ANFIS is Gradient Decent (GD) asan algorithm for parameter optimization (model).This often happens to be stuck to get optimallocal results, to overcome this PSO is used as asolution. The results obtained from the ANFIS- PSO training contained a difference of Rp.3173,187 or a percentage of 1.34%. From themodeling obtained applied to the hotel PT.PIM,with the result of an increase in revenue of Rp.17,493,548. The conclusion obtained is thatANFIS-PSO can help managers to determine theroom rate by modeling data obtained from theANFIS-PSO method. Suggestion for thedevelopment of this research is that ANFIS-PSOhas a complex complexity of training algorithmsbecause there is a combination of twoalgorithms, so to make it better a differentalgorithm design is needed.

Fuzzy wavelet neural networks applied as inferential sensors of neonatal incubator dynamics

Newborns with health complications have great difficulty in regulating the body temperature due to distinct factors, which include the high metabolism rate and low weight. In this context, neonatal incubators help maintaining good health conditions because they provide a thermally-neutral environment, which is adequate to ensure the least energy expenditure by the newborn. In the last decades, artificial neural networks (ANNs) have been established as one of the main tools for the identification of nonlinear systems. Among the various approaches used in the identification process, the fuzzy wavelet neural network (FWNN) can be regarded as a prominent technique, consisting of the combination of wavelet neural network (WNN) and adaptive network-based fuzzy inference system (ANFIS). This work proposes the use of FWNN to infer the temperature and humidity values inside the incubator in order to certify the equipment operation. Results obtained with the analyzed neural system have shown the generalization and inference capacities of FWNNs, thus allowing their application to practical tasks aiming to increase the efficiency of incubators.

Compressive strength of earth block masonry: Estimation based on neural networks and adaptive network-based fuzzy inference system

Estimating the compressive strength of earth block masonry is an essential aspect of structural design. Artificial neural networks and an adaptive network-based fuzzy inference system were utilized in this study to predict the compressive strength of earth block masonry per three parameters: height-to-thickness ratio, and compressive strength of blocks, and compressive strength of mortars. Seventy-two datasets were collected from experiments and references to train and test two respective models. The prediction results are compared against empirical calculation results to validate the proposed technique for determining earth block masonry compressive strength.

Modelling and prediction of antibacterial activity of knitted fabrics made from silver nanocomposite fibres using soft computing approaches

Antibacterial activity of knitted fabrics has been modelled and predicted by using two soft computing approaches, namely artificial neural network (ANN) and adaptive network-based fuzzy inference system (ANFIS). Four parameters, namely proportion of polyester–silver nanocomposite fibres in yarn, yarn count (diameter), machine gauge and type of fabric (100% polyester or 50:50 polyester–cotton), were used as input parameters for predicting antibacterial activity of knitted fabrics. For each of the input parameters, two fuzzy sets (low and high) were considered to reduce the complexity of ANFIS model. The sixteen linguistic fuzzy rules trained by ANFIS were able to explain the relationship between input parameters and antibacterial activity. A comparison between ANN and ANFIS models has also been presented. Both the models predicted the antibacterial activity of knitted fabrics with very good prediction accuracy in the training and testing data sets with coefficient of determination greater than 0.92 and mean absolute prediction error less than 5%. The robustness of the prediction results against data partitioning between training and testing sets has also been investigated. It is found that prediction accuracy of both the models was quite robust with ANFIS showing better performance with lesser number of training data.

A neuro-fuzzy risk prediction methodology for falling from scaffold

Fall from height is one of the most significant safety issues in the construction industry, due to the high number of fatal injuries. Scaffolds are a leading cause and have one of the highest injury rates. Therefore, it is crucial to introduce preventive measures and strategies. This study introduces a hybrid approach that merges an Adaptive Neural Network-based Fuzzy Inference System (ANFIS) and a safety inspection checklist to identify risk factors and predict the risk of falling from scaffold on construction sites. Our findings indicate that platform, joints, ladders, personal protective equipment and guardrails are the most important factors. The approach can identify and assess key conditions and situations that have the greatest impact on fall risk. The hybrid ANFIS–checklist model is found to outperform the regression method in predicting fall risk. Experts can use also this approach in other safety areas to identify and predict workplace risk.

Establishing the waist as the better location for attaching a single accelerometer to estimate center of pressure trajectories

BackgroundIn this study, we seek to replace conventional force platforms with a single accelerometer for measuring Center of Pressure trajectories, in order to achieve portability and convenience without sacrificing accuracy. MethodsWe measure the actual Anterior/Posterior and Medial/Lateral Center of Pressure trajectories of ten healthy young subjects using a force platform, and compare them with estimated measurements derived from accelerometer signals collected from three body locations (upper trunk, waist, and lower thigh) using three machine learning algorithms (Neural Network, Genetic Algorithm, and Adaptive Network-based Fuzzy Inference System). Error ratios and correlation coefficients corresponding to body locations were compared via one-way repeated-measures ANOVA. The ratios and coefficients corresponding to the three algorithms were also compared using the same approach. FindingsEstimated Anterior/Posterior trajectories indicated that measurements collected from the waist provided the lowest margins of error (8.1–8.4% v. 12.1–13.4%, P ≤ .001) and the highest correlation (.95 v. .82–.86, P ≤ .032). Estimated Medial/Lateral trajectories indicated that measurements collected from both the waist and thigh, as compared to the upper trunk, provided lower margins of error (7.0–7.3% v. 8.5–10.8%). In general, the waist is the better accelerometer attachment location. InterpretationThe results of our study corroborate our deduction that the high correlation between Center of Pressure and body's Center of Mass provides the rationale to place the single accelerometer close to the waist for Center of Pressure estimations. This study also supports the feasibility of using one single accelerometer programmed with algorithms for similar clinical applications.

Optimization of sugarcane bagasse pretreatment using alkaline hydrogen peroxide through ANN and ANFIS modelling

The present study compares the optimization using Artificial Neural Networks (ANN) and Adaptive Network-based Fuzzy Inference System (ANFIS) in the sugarcane bagasse delignification process using Alkaline Hydrogen Peroxide (AHP). Two variables were assessed experimentally: temperature (25–45 °C) and hydrogen peroxide concentration (1.5–7.5%(w/v)). The Klason Method was used to measure the amount of insoluble lignin, the High Performance Liquid Chromatography (HPLC) was used to determine the glucose and xylose concentrations and the Fourier Transform Infrared Spectroscopy (FT-IR) was applied to identify oxidized lignin structure in the samples. The analytical results were used for training and testing of ANN and ANFIS models. The statistical quality of the models was significant due to the low values of the errors indices (RMSE) and determination coefficient R2 between experimental and calculated values.