Fuzzy Inference System Model Research Articles

Fuzzy-logic modelling for quality prediction of smoked Tilapia (Oreochromis niloticus) fish

Statistical modelling of hot smoke processing of pre-treated Tilapia (Oreochromis niloticus) fish was reportedly inaccurate, making it difficult to design, predict, and reproduce the finished product’s quality; hence, accurate modelling of this process is a gap in study. This study filled this gap and extended the literature by investigating the accuracy of artificial intelligent based model for the same process. Fuzzy inference system (FIS) model was developed using the already presented dataset in the literature from where inaccurate statistical models were reportedly derived. The dataset is on the effect of smoke temperature (80, 90 and 100OC) and smoke time (2.00, 2.50 and 3.00 h) on the gross energy value (GEV) (Kcal/g) and the overall acceptability (OA) properties of brined pre-solar dried and brined non-dried Tilapia (Oreochromis niloticus) fish. The efficiency of FIS membership function types (pimf, trimf and gbellmf) on the accuracy of the developed FIS model was also investigated. Coefficient of determination, root mean square error, individual percentage error and model accuracy were used to discern the model accuracy. Results showed that FIS had a modelling accuracy (𝑅2 value) between 0.9873 and 0.9999 as against 0.1072 and 0.5800 reported for the statistical model. The results suggested that FIS model outperformed the statistical model of Tilapia (Oreochromis niloticus) smoke processing and it is recommended for process/product design, control, and standardization.

Comparative Analysis with Neural Network and Neuro-Fuzzy Model of Net Energy Consumption for Turkey

Energy, which is one of the main determinants of the economy, is an important production factor for all countries. As a developing country, Turkey is a country that increases its energy demand day by day. It is very important to make reliable energy consumption forecasts for the future in today's world where there is an energy crisis. In this work; the artificial neural networks (ANN) and adaptive-network-based fuzzy inference system (ANFIS) models were used to examine the effects of imports, exports, economic growth (Gross Domestic Product) and population on net energy consumption of Turkey. The reliability of the ANN and ANFIS models was determined using several statistical indicators. In the ANN model; R2, MAPE, and cov values were found as 0.997397669, 0.78259322, and 5.3228538, respectively. In the ANFIS model; R2, MAPE, and cov values were found as 0.997845364, 0.70709233, and 4.84339908, respectively. The obtained results from the ANN are compared with the ANFIS, in which the same data sets are used. The ANFIS model is a little better than ANN model. Using the weights obtained from the trained network, a new formula for determining net energy consumption is proposed. The results obtained, it is showing that both models can be successfully used to forecast energy consumption.

Prediction of NOx emissions from gas turbines of a combined cycle power plant using an ANFIS model optimized by GA

Combined cycle power plants, which combine gas and steam turbines, have negative impacts on surrounding populations and structures. Control of NOx emissions is an important issue for these gas-fired power plants. Accurate estimation of NOx emissions is critical for developing incinerators and reducing the environmental impact of existing plants. The objective of this study is to model ANFISGA and estimate NOx emissions from a natural gas-fired combined cycle power plant using emission monitoring system (PEMS) data. First, Adaptive Neuro Fuzzy Inference System (ANFIS) models were developed using fuzzy C-Means (FCM). Then, the parameters were optimized using a genetic algorithm (GA) to reduce the error. The proposed ANFISGA system was created, trained, and tested with PEMS datasets. The developed models were compared using several statistical performance criteria, including correlation coefficient (R2), mean squared error (MSE), error mean (EM), root mean square error (RMSE), standard deviation of error (STD), and mean absolute percentage error (MAPE). The obtained results show that the coefficient of determination varies between 0.79933 and 0.90363 for the data separated into test and training data with different rates. The minimum values of the criteria MSE, RMSE, EM, STD, and MAPE were found to be 24.8379, 4.9838, 3.4625e-05, 4.9839, and 5.1660, respectively, for the training data. The minimum values of these criteria for the test data were 26.5961, 5.1571, 0.065696, 5.157, and 5.3695, respectively. The collected results show that the proposed ANFISGA models have high potential for NOx prediction. Thus, the results show that GA has a great impact on the performance of ANFIS training and significantly improves the predictive accuracy of the model.

Estimating tread width values for stair design appropriated to people’s comfort and safety needs with adaptive network-based fuzzy inference system

In this study, it is aimed to develop an adaptive network-based fuzzy inference system (ANFIS) model that can estimate tread width values based on the physical characteristics of people and suitable for comfort and safety needs. The input values were obtained by measuring the height, step, and shoe sole lengths of the sample group of 200 people. For the tread width value to be used as output value, a prototype stair model in which different step sizes can be experienced was used. The tread width value obtained by using the test data in the developed ANFIS model was compared with the tread width value obtained from the experimental study. It has been concluded that the ANFIS model developed as a result of the comparison can be used as an efficient tool in estimating the value of stair tread width, which can meet people’s physical comfort needs.

An improved adaptive neuro fuzzy inference system model using conjoined metaheuristic algorithms for electrical conductivity prediction

Precise prediction of water quality parameters plays a significant role in making an early alert of water pollution and making better decisions for the management of water resources. As one of the influential indicative parameters, electrical conductivity (EC) has a crucial role in calculating the proportion of mineralization. In this study, the integration of an adaptive hybrid of differential evolution and particle swarm optimization (A-DEPSO) with adaptive neuro fuzzy inference system (ANFIS) model is adopted for EC prediction. The A-DEPSO method uses unique mutation and crossover processes to correspondingly boost global and local search mechanisms. It also uses a refreshing operator to prevent the solution from being caught inside the local optimal solutions. This study uses A-DEPSO optimizer for ANFIS training phase to eliminate defects and predict accurately the EC water quality parameter every month at the Maroon River in the southwest of Iran. Accordingly, the recorded dataset originated from the Tange-Takab station from 1980 to 2016 was operated to develop the ANFIS-A-DEPSO model. Besides, the wavelet analysis was jointed to the proposed algorithm in which the original time series of EC was disintegrated into the sub-time series through two mother wavelets to boost the prediction certainty. In the following, the comparison between statistical metrics of the standalone ANFIS, least-square support vector machine (LSSVM), multivariate adaptive regression spline (MARS), generalized regression neural network (GRNN), wavelet-LSSVM (WLSSVM), wavelet-MARS (W-MARS), wavelet-ANFIS (W-ANFIS) and wavelet-GRNN (W-GRNN) models was implemented. As a result, it was apparent that not only was the W-ANFIS-A-DEPSO model able to rise remarkably the EC prediction certainty, but W-ANFIS-A-DEPSO (R = 0.988, RMSE = 53.841, and PI = 0.485) also had the edge over other models with Dmey mother in terms of EC prediction. Moreover, the W-ANFIS-A-DEPSO can improve the RMSE compared to the standalone ANFIS-DEPSO model, accounting for 80%. Hence, this model can create a closer approximation of EC value through W-ANFIS-A-DEPSO model, which is likely to act as a promising procedure to simulate the prediction of EC data.

Solar Radiation Forecasting by Pearson Correlation Using LSTM Neural Network and ANFIS Method: Application in the West-Central Jordan

Solar energy is one of the most important renewable energies, with many advantages over other sources. Many parameters affect the electricity generation from solar plants. This paper aims to study the influence of these parameters on predicting solar radiation and electric energy produced in the Salt-Jordan region (Middle East) using long short-term memory (LSTM) and Adaptive Network-based Fuzzy Inference System (ANFIS) models. The data relating to 24 meteorological parameters for nearly the past five years were downloaded from the MeteoBleu database. The results show that the influence of parameters on solar radiation varies according to the season. The forecasting using ANFIS provides better results when the parameter correlation with solar radiation is high (i.e., Pearson Correlation Coefficient PCC between 0.95 and 1). In comparison, the LSTM neural network shows better results when correlation is low (PCC in the range 0.5–0.8). The obtained RMSE varies from 0.04 to 0.8 depending on the season and used parameters; new meteorological parameters influencing solar radiation are also investigated.

Using an adaptive network‐based fuzzy inference system model to predict the loss ratio of petroleum insurance in Egypt

AbstractInsurance companies and those interested in developing insurance services seek to use modern mathematical and statistical methods to study further and analyze all the company's corporate internal and external performance indicators. Loss ratio is a vital indicator used to measure performance and predict future losses in insurance companies. Many pivotal processors, such as underwriting and pricing depending on it. Therefore, accurate predictions assist insurance companies in making decisions properly. Thus, this paper aims to use the adaptive network‐based fuzzy inference system (ANFIS) and autoregressive integrated moving average (ARIMA) models in forecasting the loss ratio of petroleum insurance in Misr Insurance Holding Company from 1995 to 2019. We applied many ANFIS models according to ANFIS properties and used the first 21 years (1995–2015), making up the training data set, which represents 85% of the data, as well as the past 4 years (2016–2019). Which are used for the testing stage and represent 15% of the data. Our finding concluded that ANFIS models give more accurate results than ARIMA models in predicting the loss ratio during the investigation by comparing results using predictive accuracy measures.

Watermain's failure index modeling via Monte Carlo simulation and fuzzy inference system

The aging and degradation of water supply systems are deemed serious problems that cause pipeline failure and breakages. The risk of watermains failure assessment is one of the key strategies that can pinpoint pipes at risk and maintain their sustainability. This research paper showcases a novel method for deterioration modeling in conjunction with quantifying the water network’s failure index (FI). The methodology builds on various algorithms, computational intelligence, and interactions between numerous factors. It involves developing two intelligent models; the first is the Monte Carlo Simulation Model (MCSM) that is designed to estimate the deterioration indices (DIs) of watermains through intricate iterative simulations. The produced indices are then streamlined and channeled to the fuzzy engine to develop the second model, namely, the fuzzy inference system model (FISM). After designing the model’s configuration, the DIs are mapped to 84 fuzzy-if-then-rules that are in turn mapped to output values; finally, the fuzzy consolidator generates one crisp number that represents the watermain’s FI. The developed method is implemented on the water system of the City of El Pedregal in Peru. The results indicate a moderate risk of failure status, which corresponds to 62.3 FI. Moreover, the efficacy of this method is verified against the multiple linear regression (MLR) method and proved to be sound. This research provides insights for infrastructure managers in the aspects of when to intervene, what to maintain, replace, or rehabilitate, and how to focus their constrained funding on the most deserving assets.

Motion comfort analysis of tight-fitting sportswear from multi-dimensions using intelligence systems

Focusing on the “human–sport-clothing” system, this paper analyzed the influence of combinations of different tights and sport states on human body parts and overall body comfort from multiple dimensions. However, the motion state and some fabric parameters are non-numerical parameters, which could not be used for model analysis directly. In addition, there are too many numerical fabric parameters whose relationships are complicated, and it is difficult for general models to deal with these relationships, resulting in low accuracy of the comfort prediction model. Moreover, when using the artificial neural network to study comfort, it has some difficulties in expressing comfort and low prediction accuracy. To solve these problems, One-Hot was used to encode non-numerical parameters, and then intelligent algorithms were adopted to deal with these complex fabric parameters. Finally, a comfort prediction model was established in combination with an adaptive fuzzy reasoning system. The results showed that different fabric combinations and motion states had significant effects on local comfort (comfort of specific human body parts) and global comfort (whole body comfort). Moreover, the prediction model with non-numerical parameters has higher accuracy than the model without non-numerical parameters, which indicated that the prediction accuracy of the model had been improved after the introduction of One-Hot coding, so the non-numerical parameters cannot be ignored. The particle swarm optimization algorithm-cuckoo search algorithm-adaptive network-based fuzzy inference system hybrid model was superior to the particle swarm optimization algorithm-adaptive network-based fuzzy inference system and cuckoo search algorithm-adaptive network-based fuzzy inference system model in predicting local comfort and global comfort.

Automated reasoning of vehicle brake-force: a fuzzy inference system model

Automatic regulation of vehicle braking system is very crucial to control vehicle speed, keep it on its track, and avoid any collision. A traditional binary control for brake system is not appropriate as the on-off conditions induce unpleasant feelings to the passengers. Control engineering technique may solve this problem but does not reflect artificial intelligence in terms of human reasoning behaviour. Although, a few study on fuzzy-based automated brake-force reasoning has been conducted showing the superiority over all the techniques, comparisons among various defuzzification methods are rarely found in the literature. Thus, the objective of this study is to design a fuzzy-based architecture for automated brake-force reasoning system through comparative analysis among various defuzzification approaches. The model takes two inputs as linguistic variables, the speed and the vehicular distance, and the output of the system is the required brake-force. The Mamdani fuzzy inference system (MFIS) is applied in designing the fuzzy architecture for automated brake-force prediction. The output fuzzy set is defuzzified with five defuzzification methods and analysed through observations. The results manifest that out of the five defuzzification approaches, the centroid and bisector methods demonstrate most satisfactory outputs and could be chosen for real system application.

Automated Reasoning of Vehicle Brake-Force: A Fuzzy Inference System Model

Automated Reasoning of Vehicle Brake-Force: A Fuzzy Inference System Model

An Adaptive Fuzzy Multi-Criteria Model for Sustainability Assessment of Sugarcane Agroindustry Supply Chain

Supply chain sustainability assessment is key to maintaining and improving the performance of agroindustry supply chains, particularly in sustainable agroindustry development. The assessment of agro-industrial supply chain performance is a complex and dynamic process. Hence, there is a need for an adaptive fuzzy multi-criteria sustainability assessment model as an alternative method of analysis and improvement. This study aimed to design an adaptive fuzzy multi-criteria sustainability assessment and improvement model for the sugarcane agroindustry supply chain. In this study: (1) a fuzzy inference system (FIS) was developed to assess the performance of the sustainability dimensions. This study proposed 24 indicators for four dimensions: economic, social, environmental, and resource. (2) An adaptive neuro-fuzzy inference system (ANFIS) is designed to aggregate the overall supply chain sustainability performance. (3) The proposed fuzzy multi-criteria assessment model was compared with the common multidimensional scaling (MDS) and linear models. This study proved that the proposed synthesis of the FIS and ANFIS models is powerful and adaptive for evaluating supply chain sustainability and providing accurate results. (4) The strategies to improve sustainability performance were developed using the cosine amplitude method (CAM). The proposed model determined that the overall supply chain sustainability value was 68.58%, which was almost sustainable. Several strategies have been suggested to improve sustainability performance, including maintaining the sugarcane supply by strengthening the partnership program and improving the mill’s overall recovery, followed by factory revitalization or new factory investment.

Cloud spot price prediction approach using adaptive neural fuzzy inference system with chaos theory

The dynamic pricing of cloud computing is a major challenge for cloud users all over the world. This challenge was first addressed by Amazon under the name of Amazon spot instance market. Cloud users can bid for a spot instance using this market and obtain the requested spot if their bids exceed a dynamically changing spot price. Amazon publicises the spot price but does not reveal how it is determined. In this paper, we perform chaotic time series analysis over the spot price trace. We also develop a chaos based adaptive neural fuzzy inference system (ANFIS) model based on phase-space vectors obtained during the phase of chaotic analysis. Next, we study the effect of chaos existence on the prediction accuracy of the spot price by comparing the proposed chaos-ANFIS model with the baseline ANFIS model (non-chaotic approach). Evaluation results show that the proposed chaos-ANFIS model yields better predictions of spot price compared to the baseline ANFIS model in terms of root mean square error (RMSE) and mean absolute percentage error (MAPE).

Application of fuzzy inference system (FIS) for assessment and predication of compressive asset of concrete containing fly ash

Fuzzy Inference System (FIS) for forecasting the 90 days compressive asset of cements containing low and high limes cinders have been modified by using MATLAB2016a. For motivation behind to modified this model is 160 specimens and 48 different mixes found form the literatures. Water, cover of the day, crushed stone-I (3–7 mm), Portland cement, Fly Ash (FA) replacement ratio has been used as input parameter in the FIS modelling system. The output is the boundary which is compressive strength of cement. The preparation and testing results advocate that the FIS modelling system have the durable potentially predict 90 days compressive strength of concretes fly ash and maybe used for some other purpose also. This FIS system provide the reasonable results with the comparison of previous studies.

A Fuzzy Inference System Model for Fluid Resuscitation of Burn Patients

A Fuzzy Inference System Model for Fluid Resuscitation of Burn Patients

Determination of ride comfort thresholds based on international roughness index for asphalt concrete pavements

ABSTRACT The study is aimed to mathematically model the relationship between the amount of whole-body vibration exposed in a passenger car type vehicle and the ride speed of the vehicle in the same section and the International Roughness Index (IRI), which is used as a pavement performance indicator. Vibration measurements were analysed according to the evaluation method determined in the ISO 2631 standard, and frequency-weighted root-mean-square acceleration (aw) values were obtained in the vertical direction. Mathematical relationships between real measurement data performed in 1114 road sections were investigated using regression analysis, artificial neural networks (ANN) and Adaptive Neural Fuzzy Inference System (ANFIS) modelling techniques. The models' performance levels were determined using comparison metrics, and it was determined that ANFIS was the model with the best goodness of fit, with the regression coefficient 0.955, mean absolute error (MAE) 0.013, and root mean squared error (RMSE) 0.020. The threshold values affecting the ride comfort of IRI and the ride comfort values corresponding to the IRI limit values recommended by the Federal Highway Administration (FHWA) were determined through the ANFIS mathematical model. Then, a sensitivity analysis was conducted to determine the effects of a certain increase in the IRI value on discomfort.

An intelligent approach based on hybrid of principal component analysis-fuzzy inference system-adaptive neuro-fuzzy inference system for the simultaneous spectrophotometric determination of sofosbuvir and velpatasvir as antiviral drugs in pharmaceutical formulation and urine sample

In this study, a simple, fast, accurate, and inexpensive chemometrics-assisted spectrophotometric method was developed for the simultaneous determination of sofosbuvir (SOF) and velpatasvir (VEL) as antiviral drugs. Fuzzy inference system (FIS) and adaptive neuro-fuzzy inference system (ANFIS) along with principal component analysis (PCA) were used for the analysis of SOF and VEL in binary mixtures, pharmaceutical formulation, and urine sample. Root mean square error (RMSE) of FIS model was 0.1518 and 0.075 for SOF and VEL, respectively. Also, the average testing error related to the ANFIS method was achieved 2.23 ​× ​10−5 and 5.58 ​× ​10−6 for SOF and VEL, respectively. The range of mean recovery percentage was between 99.76 and 100.41 for both methods. Relative standard deviation (RSD) of the real sample was lower than 0.06% for FIS and ANFIS for both components. Analysis of variance (ANOVA) was applied to the results of the real sample that there was no significant difference between the suggested approaches. PCA-FIS and PCA-ANFIS as efficient and accurate methods can be useful for the simultaneous estimation of binary and ternary mixtures in quality control of pharmaceutical formulations.

ANFIS-based accurate modeling of silica gel adsorption cooling cycle

Using an eco-friendly system to provide cooling power for buildings is a great challenge. The adsorption cooling (AC) cycle is a promising thermally driven technology to provide green cooling energy for buildings. It uses environmentally friendly working fluids and could be powered by low-grade thermal energy. Despite these advantages, it suffers from low performance compared to electrical-driven cooling systems. Achieving high performance of the AC cycle relies basically on the state of operation that needs to be optimized to get the compromise between the specific cooling power (SCP) and the coefficient of performance (COP). So, the key objective of the paper is to model the outputs (SCP and COP) of the cooling cycle in terms of its operating conditions using artificial intelligence. The evaporator temperature, chilled water cycle time, chilled water mass flow rate are the main operating parameters (inputs) of the system that are optimized. An accurate Adaptive Network-based Fuzzy Inference System (ANFIS) model of COP and SCP is created based on experimental datasets. ANFIS can accurately grasp the trend of the data and hence produce an accurate model even if the data describes a highly nonlinear system. To verify the performance superiority of the fuzzy model, the attained findings are compared with ANOVA. The ANOVA uses the linear regression method to obtain the mathematical model of the system. For the COP model and during the training phase, the coefficient of determination is increased from 0.7477 by using ANOVA to 0.9997 by using ANFIS (33.7% increase). Also, during the testing phase, the coefficient of determination is increased from 0.6522 to 0.9725 (32.9% increase). Regarding the SCP model, the coefficient of determination is increased by 27.66% and 46.6% for training and testing, respectively. The RMSE values are found to be 0.0026 and 0.0437, respectively, for training and testing for modeling the ANFIS model of COP. These values prove the superiority of fuzzy modeling compared with ANOVA. Also, the RMSE values are decreased by 97.37% and 80.91% using ANFIS compared with ANOVA, respectively, for COP and SCP models. Therefore, the obtained results demonstrated the excellence of the ANFIS model of implementing the AC cycle in comparison with ANOVA.

Mutating fuzzy logic model with various rigorous meta-heuristic algorithms for soil moisture content estimation

Soil moisture content (SMC) prediction can contribute to diverse geo-science engineering applications such as plantation, crops production, and several irrigation activities. Although, there have been several methodologies introduced for the SMS estimation, methods are still associated with challenges and limitations (e.g., time-consuming, high cost, and the need for large data sets). In the current research, the motivation was inspired to develop new data-intelligence models that are capable to estimate SMC accurately and reliably. Several hybridized adaptive neuro fuzzy inference system (NF) models were developed. The feasibility of four meta-heuristic algorithms including the gray wolf optimization (GWO), Bee algorithm (BA), Firefly algorithm (FA), and imperialistic competitive algorithm (ICA), were used to improve the predictability performance of the NF model. The performances of the hybrid models were tested at two different climate characteristics (i.e., humid and arid) and four different soil depths (5, 10, 20, and 50 cm). The dataset used in this study were consisted of ten years with daily scale. Practical inputs predictors for the prediction matrix were identified using the mutual information method (MI) and correlation coefficient, which included maximum and minimum temperature, solar radiation, and soil temperature. K-Fold technique and various indices were used to evaluate hybrid models. Despite the promising results of all developed models, the NF-FA is superior to the others in both climates. The average results of all depths (in humid climate) using correlation coefficient (r), mean absolute percentage error (MAPE), and scattering index (SI) were 0.924, 6.089%, and 0.074, respectively. While in the arid climate, NF-FA model attained 0.903, 14.439%, and 0.176 in r, MAPE, and SI, respectively. The evaluation of different depths and climates showed that NF-FA yielded the best accuracy within humid climates and depths of 20 and 10 cm had the lowest estimation error in humid and arid climates. This study introduced the FA algorithm as an efficient tool in NF development to estimate SMC by readily available inputs without knowing soil physical parameters.

Modeling and optimization of EDM machined AZ-91 Mg alloy using ANFIS-VIKOR method

In this paper, machining of AZ-91 magnesium alloy was performed in EDM using different tool electrode (Cu, CuW and graphite). To perform experiments, Taguchi L18 design of experiments was used to reduce experimental runs. EDM process parameters viz. polarity, current (Ip), pulse-on-time (Ton), pulse-off-time (Toff), and tool electrode material were considered in experimental design to measure the responses (MRR, Ra). Multi-input-single-output adaptive neuro fuzzy inference system (ANFIS) model was developed to predict responses, and predicted results were found in good agreement with the experimental results. Maximum MRR (0.089 g/min) was found at positive polarity, Ip-5 A, Ton-50 μs, Toff - 30 μs, and Cu tool, and minimum Ra (0.08 μm) was at parameters positive polarity, Ip-4 A, Ton-30 μs, Toff -20 μs, and Cu tool. Relative errors between experiential results and ANFIS predicted results were 1.17 % for MRR, and 2.20 % for Ra. Multi response optimization ANFIS-VIKOR method was successfully developed and gave compromise solution for MRR and Ra corresponding to minimum ANFIS-VIKOR index (Qi). A factor level analysis was performed to evaluate optimal factor combination for ANFIS-VIKOR index, and it shows that current (Ip) have a significant effect.