Adaptation of Fuzzy Inference System Using Neural Learning

Adaptive Neuro-Fuzzy Inference System guided objective function parameter optimization for inverse treatment planning

Intensity-Modulated Radiation Therapy requires the manual adjustment to numerous treatment plan parameters (TPPs) through a trial-and-error process to deliver precise radiation doses to the target while minimizing exposure to surrounding healthy tissues. The goal is to achieve a dose distribution that adheres to a prescribed plan tailored to each patient. Developing an automated approach to optimize patient-specific prescriptions is valuable in scenarios where trade-off selection is uncertain and varies among patients. This study presents a proof-of-concept artificial intelligence (AI) system based on an Adaptive Neuro-Fuzzy Inference System (ANFIS) to guide IMRT planning and achieve optimal, patient-specific prescriptions in aligned with a radiation oncologist's treatment objectives. We developed an in-house ANFIS-AI system utilizing Prescription Dose (PD) constraints to guide the optimization process toward achievable prescriptions. Mimicking human planning behavior, the AI system adjusts TPPs, represented as dose-volume constraints, to meet the prescribed dose goals. This process is informed by a Fuzzy Inference System (FIS) that incorporates prior knowledge from experienced planners, captured through "if-then" rules based on routine planning adjustments. The innovative aspect of our research lies in employing ANFIS's adaptive network to fine-tune the FIS components (membership functions and rule strengths), thereby enhancing the accuracy of the system. Once calibrated, the AI system modifies TPPs for each patient, progressing through acceptable prescription levels, from restrictive to clinically allowable. The system evaluates dosimetric parameters and compares dose distributions, dose-volume histograms, and dosimetric statistics between the conventional FIS and ANFIS. Results demonstrate that ANFIS consistently met dosimetric goals, outperforming FIS with a 0.7% improvement in mean dose conformity for the planning target volume (PTV) and a 28% reduction in mean dose exposure for organs at risk (OARs) in a C-Shape phantom. In a mock prostate phantom, ANFIS reduced the mean dose by 17.4% for the rectum and by 14.1% for the bladder. These findings highlight ANFIS's potential for efficient, accurate IMRT planning and its integration into clinical workflows.

Digitalization enhancement in the pharmaceutical supply network using a supply chain risk management approach

One major issue in pharmaceutical supply chain management is the supply shortage, and determining the root causes of medicine shortages necessitates an in-depth investigation. The concept of risk management is proposed in this study to identify significant risk factors in the pharmaceutical supply chain. Fuzzy failure mode and effect analysis and data envelopment analysis were used to evaluate the risks of the pharmaceutical supply chain. Based on a case study on the Malaysian pharmaceutical supply chain, it reveals that the pharmacy node is the riskiest link. The unavailability of medicine due to unexpected demand, as well as the scarcity of specialty or substitute drugs, pose the most significant risk factors. These risks could be mitigated by digital technology. We propose an appropriate digital technology platform consisting of big data analytics and blockchain technologies to undertake these challenges of supply shortage. By addressing risk factors through the implementation of a digitalized supply chain, organizations can fortify their supply networks, fostering resilience and efficiency, and thereby playing a pivotal role in advancing the Pharma 4.0 era.

Evaluation of sisal fiber and aluminum waste concrete blend for sustainable construction using adaptive neuro-fuzzy inference system

This research study presents evaluation of aluminum waste-sisal fiber concrete's mechanical properties using adaptive neuro-fuzzy inference system (ANFIS) to achieve sustainable and eco-efficient engineering works. The deployment of artificial intelligence (AI) tools enables the optimization of building materials combined with admixtures to create durable engineering designs and eliminate the drawbacks encountered in trial-and-error or empirical method. The features of the cement-AW blend's setting time were evaluated in the laboratory and the results revealed that 0-50% of aluminum-waste (AW) inclusion increased both the initial and final setting time from 51-165 min and 585-795 min respectively. The blended concrete mix's flexural strength tests also show that 10% sisal-fiber (SF) substitution results in a maximum flexural strength of 11.6N/mm², while 50% replacement results in a minimum flexural strength of 4.11N/mm². Moreover, compressive strength test results show that SF and AW replacements of 0.08% and 0.1%, respectively, resulted in peak outcome of 24.97N/mm², while replacements of 0.5% and 0.45% resulted in a minimum response of 17.02N/mm². The ANFIS-model was developed using 91 datasets obtained from the experimental findings on varying replacements of cement and fine-aggregates with AW and SF respectively ranging from 0 to 50%. The ANFIS computation toolbox in MATLAB software was adopted for the model simulation, testing, training and validation of the response function using hybrid method of optimization and grid partition method of FIS at 100 Epochs. The compressive strength behavior is the target response, and the mixture variations of cement-AW and fine aggregates-SF combinations were used as the independent variables. The ANFIS-model performance assessment results obtained using loss function criteria demonstrates MAE of 0.1318, RMSE of 0.412, and coefficient of determination value of 99.57% which indicates a good relationship between the predicted and actual results while multiple linear regression (MLR) model presents a coefficient of determination of 82.46%.

Rheological Behavior of SAE50 Oil-SnO2-CeO2 Hybrid Nanofluid: Experimental Investigation and Modeling Utilizing Response Surface Method and Machine Learning Techniques

In this study, for the first time, the effects of temperature and nanopowder volume fraction (NPSVF) on the viscosity and the rheological behavior of SAE50-SnO2-CeO2 hybrid nanofluid have been studied experimentally. Nanofluids in NPSVFs of 0.25% to 1.5% have been made by a two-step method. Experiments have been performed at temperatures of 25 to 67 °C and shear rates (SRs) of 1333 to 2932.6 s-1. The results revealed that for base fluid and nanofluid, shear stress increases with increasing SR and decreasing temperature. By increasing the temperature to about 42 °C at a NPSVF of 1.5%, about 89.36% reduction in viscosity is observed. The viscosity increases with increasing NPSVF about 37.18% at 25 °C. In all states, a non-Newtonian pseudo-plastic behavior has been observed for the base fluid and nanofluid. The highest relative viscosity occurs for NPSVF = 1.5%, temperature = 25 °C and SR = 2932.6 s-1, which increases the viscosity by 37.18% compared to the base fluid. The sensitivity analysis indicated that the highest sensitivity is related to temperature and the lowest sensitivity is related to SR. Response surface method, curve fitting method, adaptive neuro-fuzzy inference system and Gaussian process regression (GPR) have been used to predict the dynamic viscosity. Based on the results, all four models can predict the dynamic viscosity. However, the GPR model has better performance than the other models.

Comparison of Different Approaches to the Creation of a Mathematical Model of Melt Temperature in an LD Converter

In the steel production process in the LD converter, it is important to have information about the melt temperature. The temperature and chemical composition of the steel are important parameters in this process in terms of its completion. During the process, continuous measurement of the melt temperature and thus also information about the end of the process are missing. This paper describes three approaches to creating a mathematical model of melt temperature. The first approach is a regression model, which predicts an immeasurable melt temperature based on other directly measured process variables. The second approach to creating a mathematical model is based on the machine learning method. Simple and efficient learning algorithms characterize the machine learning methods. We used support vector regression (SVR) method and the adaptive neuro-fuzzy inference system (ANFIS) to create a mathematical model of the melt temperature. The third approach is the deterministic approach, which is based on the decomposition of the process and its heat balance. The mathematical models that were compiled based on the mentioned approaches were verified and compared to real process data.

Prediction of Arsenic Removal from Contaminated Water Using Artificial Neural Network Model

Arsenic is a deleterious heavy metal that is usually removed from polluted water based on adsorption processes. The latest mode of modeling such a process is to implement artificial intelligence (AI). In the current work, a new artificial neural network (ANN) model was developed to predict the adsorption efficiency of arsenate (As(III)) from contaminated water by analyzing different architectures of an adaptive network-based fuzzy inference system (ANFIS). The database for the current study consisted of the experimental data of the adsorption of As(III) by different adsorbents/biosorbents. The data were randomly divided into two sets: 70% for the training phase and 30% for the testing phase. Four statistical evaluation metrics, namely, mean square error (MSE), root-mean-square error (RMSE), Pearson’s correlation coefficient (R%), and the determination coefficient (R2) were used for the analysis. The best performing ANFIS model was characterized with the average values of 97.72%, 0.9333, 0.137, and 0.274 of R%, R2, MSE, and RMSE, respectively. In addition, a parametric investigation revealed that the most dominating parameters on the adsorption process efficiency were in the following order: pH, As initial concentration, contact time, adsorbent dosage, inoculum size, and temperature. The results of the current study would be useful in the adsorption process scale-up and optimization.

Utilization of adaptive neuro-fuzzy interference system and functional network in prediction of total organic carbon content

Abstract

This paper presents the application of two artificial intelligence (AI) approaches in the prediction of total organic carbon content (TOC) in Devonian Duvernay shale. To develop and test the models, around 1250 data points from three wells were used. Each point comprises TOC value with corresponding spectral and conventional well logs. The tested AI techniques are adaptive neuro-fuzzy interference system (ANFIS) and functional network (FN) which their predictions are compared to existing empirical correlations. Out of these two methods, ANFIS yielded the best outcomes with 0.98, 0.90, and 0.95 correlation coefficients (R) in training, testing, and validation respectively, and the average errors ranged between 7 and 18%. In contrast, the empirical correlations resulted in R values less than 0.85 and average errors greater than 20%. Out of eight inputs, gamma ray was found to have the most significant impact on TOC prediction. In comparison to the experimental procedures, AI-based models produces continuous TOC profiles with good prediction accuracy. The intelligent models are developed from preexisting data which saves time and costs.

Article highlights

Machine Learning Models for Equivalent Circulating Density Prediction from Drilling Data

Equivalent circulating density (ECD) is considered a critical parameter during the drilling operation, as it could lead to severe problems related to the well control such as fracturing the drilled formation and circulation loss. The conventional way to determine the ECD is either by carrying out the downhole tool measurements or by using mathematical models. The downhole measurement is costly and has some limitations with the practical operations, while the mathematical models do not provide a high level of accuracy. Determination of the ECD should have a high level of accuracy, and therefore, the objective of this study is to employ machine learning techniques such as artificial neural networks (ANNs) and adaptive network-based fuzzy inference systems (ANFISs) to predict the ECD from only the drilling data with a high accuracy level. The study utilized drilling data from a horizontal drilling section that includes drilling parameters (penetration rate, rotating speed, torque, weight on bit, pumping rate, and pressure of standpipe). The models were built and tested from a data set that has 3570 data points, and another data set of 1130 measurements was employed for validating the models. The accuracy of the models was determined by key performance indices, which are the coefficient of correlation (R) and the average absolute percentage error (AAPE). The results showed the strong prediction capability for ECD from the two models through training, testing, and validation processes with R greater than 0.98 and a very low error of 0.3% for the ANN model, while ANFIS recorded R of 0.96 and AAPE of 0.7, and hence, the two models showed great performance for ECD estimation application. Also, the study introduces a newly developed equation for ECD determination from drilling data in real time.

Generating pseudo-computerized tomography (P-CT) scan images from magnetic resonance imaging (MRI) images using machine learning algorithms based on fuzzy theory for radiotherapy treatment planning

PURPOSE

The substitution of computerized tomography (CT) with magnetic resonance imaging (MRI) has been investigated for external radiotherapy treatment planning. The present study aims to use pseudo-CT (P-CT) images created by MRI images to calculate the dose distribution for facilitating the treatment planning process.

METHODS

In this work, following image segmentation with a fuzzy clustering algorithm, an adaptive neuro-fuzzy algorithm was utilized to design the Hounsfield unit (HU) conversion model based on the features vector of MRI images. The model was generated on the set of extracted features from the gray-level co-occurrence matrices and the gray-level run-length matrices for 14 arbitrarily selected patients with brain malady. The performance of the algorithm was investigated on blind datasets through dose-volume histogram and isodose curve evaluations, using the RayPlan treatment planning system (TPS), along with the gamma analysis and statistical indices.

RESULTS

In the proposed approach, the mean absolute error within the range of 45.4 HU was found among the test data. Also, the relative dose difference between the planning target volume region of the CT and the P-CT was 0.5%, and the best gamma pass rate for 3%/3 mm was 97.2%.

CONCLUSION

The proposed method provides a satisfactory average error rate for the generation of P-CT data in the different parts of the brain region from a collection of MRI series. Also, dosimetric parameters evaluation shows good agreement between reference CT and related P-CT images.

Proposing an ensemble learning model based on neural network and fuzzy system for keratoconus diagnosis based on Pentacam measurements

PURPOSE

The present study was done to evaluate efficiency of an ensemble learning structure for automatic keratoconus diagnosis and to categorize eyes into four different groups based on a combination of 19 parameters obtained from Pentacam measurements.

METHODS

Pentacam data from 450 eyes were enrolled in the study. Eyes were separated into training, validation, and testing sets. An ensemble system was used to analyze corneal measurements and categorize the eyes into four groups. The ensemble system was trained to consider indices from both anterior and posterior corneal surfaces. Efficiency of the ensemble system was evaluated and compared in each group.

RESULTS

The best accuracy was achieved by the ensemble system with both multilayer perceptron and neuro-fuzzy system classifiers alongside the Naïve Bayes combination method. The accuracy achieved in KC versus N distinction task was equal to 98.2% with 99.1% of sensitivity and 96.2% of specificity for KC detection. The global accuracy was equal to 98.2% for classification of 4 groups, with an average sensitivity of 98.5% and specificity of 99.4%.

CONCLUSION

In this study, authority of an ensemble learning system to work out intricate problems was presented. Despite using fewer parameters, herein, comparable or, in some cases, better results were obtained than methods reported in the literature. The proposed method demonstrated very good accuracy in discriminating between normal eyes and different stages of keratoconus eyes. In some cases, it was not possible to directly compare our results with the literature, due to differences in definitions of KC group as well as differences in selection of items and parameters.

Real-time prediction of Poisson's ratio from drilling parameters using machine learning tools

Rock elastic properties such as Poisson’s ratio influence wellbore stability, in-situ stresses estimation, drilling performance, and hydraulic fracturing design. Conventionally, Poisson’s ratio estimation requires either laboratory experiments or derived from sonic logs, the main concerns of these methods are the data and samples availability, costs, and time-consumption. In this paper, an alternative real-time technique utilizing drilling parameters and machine learning was presented. The main added value of this approach is that the drilling parameters are more likely to be available and could be collected in real-time during drilling operation without additional cost. These parameters include weight on bit, penetration rate, pump rate, standpipe pressure, and torque. Two machine learning algorithms were used, artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS). To train and test the models, 2905 data points from one well were used, while 2912 data points from a different well were used for model validation. The lithology of both wells contains carbonate, sandstone, and shale. Optimization on different tuning parameters in the algorithm was conducted to ensure the best prediction was achieved. A good match between the actual and predicted Poisson’s ratio was achieved in both methods with correlation coefficients between 0.98 and 0.99 using ANN and between 0.97 and 0.98 using ANFIS. The average absolute percentage error values were between 1 and 2% in ANN predictions and around 2% when ANFIS was used. Based on these results, the employment of drilling data and machine learning is a strong tool for real-time prediction of geomechanical properties without additional cost.

Analysis and Prediction of COVID-19 Pandemic in Bangladesh by Using ANFIS and LSTM Network

The dangerously contagious virus named "COVID-19" has struck the world strong and has locked down billions of people in their homes to stop the further spread. All the researchers and scientists in various fields are continually developing a vaccine and prevention methods to aid the world from this challenging situation. However, a reliable prediction of the epidemic may help control this contiguous disease until the cure is available. The machine learning techniques are one of the frontiers in predicting this outbreak's future trend and behavior. Our research is focused on finding a suitable machine learning algorithm that can predict the COVID-19 daily new cases with higher accuracy. This research has used the adaptive neuro-fuzzy inference system (ANFIS) and the long short-term memory (LSTM) to foresee the newly infected cases in Bangladesh. We have compared both the experiments' results, and it can be forenamed that LSTM has shown more satisfactory results. Upon study and testing on several models, we have shown that LSTM works better on a scenario-based model for Bangladesh with mean absolute percentage error (MAPE)-4.51, root-mean-square error (RMSE)-6.55, and correlation coefficient-0.75. This study is expected to shed light on COVID-19 prediction models for researchers working with machine learning techniques and avoid proven failures, especially for small imprecise datasets.

Differential diagnosis of asthma and restrictive respiratory diseases by combining forced oscillation measurements, machine learning and neuro-fuzzy classifiers

To design machine learning classifiers to facilitate the clinical use and increase the accuracy of the forced oscillation technique (FOT) in the differential diagnosis of patients with asthma and restrictive respiratory diseases. FOT and spirometric exams were performed in 97 individuals, including controls (n = 20), asthmatic patients (n = 38), and restrictive (n = 39) patients. The first experiment of this study showed that the best FOT parameter was the resonance frequency, providing moderate accuracy (AUC = 0.87). In the second experiment, a neuro-fuzzy classifier and different supervised machine learning techniques were investigated, including k-nearest neighbors, random forests, AdaBoost with decision trees, and support vector machines with a radial basis kernel. All classifiers achieved high accuracy (AUC ≥ 0.9) in the differentiation between patient groups. In the third and fourth experiments, the use of different feature selection techniques allowed us to achieve high accuracy with only three FOT parameters. In addition, the neuro-fuzzy classifier also provided rules to explain the classification. Neuro-fuzzy and machine learning classifiers can aid in the differential diagnosis of patients with asthma and restrictive respiratory diseases. They can assist clinicians as a support system providing accurate diagnostic options.

Prediction of Anger Expression of Individuals with Psychiatric Disorders using the Developed Computational Codes based on the Various Soft Computing Algorithms

Anger is defined as a psychobiological emotional state that consists of feelings varying in intensity from mild irritation or annoyance to intense fury and rage. Dysfunction in anger regulation is marker of most psychiatric disorders. The most important point about anger regulation by the individuals is how to express anger and control it. The purpose of the present study is to predict the anger expression from the anger experience in individuals with psychiatric disorder for assessment of how to express and control the anger. To this end, the number of 3,000 subjects of individuals with clinical disorders had filled in the State-Trait Anger Expression Inventory-II (STAXI-II). After removing the uncertain diagnoses (900 subjects), the number of 2,100 data was considered in the analysis. Then, the computational codes based on three soft computing algorithms, including Radial Basis Function (RBF), Adaptive Neuro-Fuzzy Inference System (ANFIS) and Decision Tree (DT) were developed to predict the scales of anger expression of the individuals with psychiatric disorders. The scales of anger experience were used as input data of the developed computational codes. Comparison between the results obtained from the DT, RBF and ANFIS algorithms show that all the developed soft computing algorithms forecast the anger expression scales with an acceptable accuracy. However, the accuracy of the DT algorithm is better than the other algorithms.

Application of ANFIS for the Selection of Optimal Wire-EDM Parameters While Machining Ti-6Al-4V Alloy and Multi-Parametric Optimization Using GRA Method

The applications of artificial intelligence (AI) are becoming more popular and relevant research have been conducted in every field of engineering and science by using these AI techniques. Therefore, this research aims to examine the influence of wire electric-discharge machining (WEDM) parameters on performance parameters to improve the productivity with a higher surface finish of titanium alloy (Ti-6Al-4V) by using the artificial intelligent technique. In this experimental analysis, the Adaptive Network Based fuzzy Inference System (ANFIS) model has been highly-developed and the multi-parametric optimization has been done to find the optimal solution for the machining of the titanium superalloy. The peak current (Ip), taper angle, pulse on time (Ton), pulse of time (Toff) and the dielectric fluid flow rate were selected as operation constraints to conduct experimental trials. The surface roughness (SR) and MRR were considered as output responses. The influence on machining performance has been analyzed by an ANFIS model and the developed model was validated with the full factorial regression models. The developed models showed the minimum mean percentage error and the optimized parameters by the GRA method showed the considerable improvement in the process.

Prediction of formation force during single-point incremental sheet metal forming using artificial intelligence techniques

Single-point incremental forming (SPIF) is a technology that allows incremental manufacturing of complex parts from a flat sheet using simple tools; further, this technology is flexible and economical. Measuring the forming force using this technology helps in preventing failures, determining the optimal processes, and implementing on-line control. In this paper, an experimental study using SPIF is described. This study focuses on the influence of four different process parameters, namely, step size, tool diameter, sheet thickness, and feed rate, on the maximum forming force. For an efficient force predictive model based on an adaptive neuro-fuzzy inference system (ANFIS), an artificial neural network (ANN) and a regressions model were applied. The predicted forces exhibited relatively good agreement with the experimental results. The results indicate that the performance of the ANFIS model realizes the full potential of the ANN model.

Automatic and heuristic complete design for ANFIS classifier

There is a variety of fuzzy classifiers, one of which is Adaptive Neuro-Fuzzy Inference system (ANFIS) classifier. One of the main challenges in designing such data classifiers is selection of effective and appropriate type and location of membership functions and its training method to reduce the classification error. In this paper, a new technique (based on intelligent methods) is presented and implemented to select and locate the membership functions and simultaneous training using a new method based on Inclined Planes System Optimization (IPO) to minimize errors of an ANFIS classifier for the first time. The presented method is evaluated for classification of data sets with different reference classes and different length feature vectors, which have acceptable complexity. According to the results of the research, the presented method has a higher level of accuracy and efficiency in selecting the type and location of membership functions (based on intelligent methods) and simultaneous training with IPO, compared to other methods, such as particle swarm optimization, genetic algorithm, differential evolution, and ACOR algorithms.

Towards Utilization of Neurofuzzy Systems for Taxonomic Identification Using Psittacines as a Case Study

Demonstration of the neurofuzzy application to the task of psittacine (parrot) taxonomic identification is presented in this paper. In this work, NEFCLASS-J neurofuzzy system is utilized for classification of parrot data for 141 and 183 groupings, using 68 feature points or qualities. The reported results display classification accuracies of above 95%, which is strongly tied to the setting of certain parameters of the neurofuzzy system. Rule base sizes were in the range of 1,750 to 1,950 rules.

INTELLIGENT MULTIAGENT COORDINATION BASED ON REINFORCEMENT HIERARCHICAL NEURO-FUZZY MODELS

This paper presents the research and development of two hybrid neuro-fuzzy models for the hierarchical coordination of multiple intelligent agents. The main objective of the models is to have multiple agents interact intelligently with each other in complex systems. We developed two new models of coordination for intelligent multiagent systems, which integrates the Reinforcement Learning Hierarchical Neuro-Fuzzy model with two proposed coordination mechanisms: the MultiAgent Reinforcement Learning Hierarchical Neuro-Fuzzy with a market-driven coordination mechanism (MA-RL-HNFP-MD) and the MultiAgent Reinforcement Learning Hierarchical Neuro-Fuzzy with graph coordination (MA-RL-HNFP-CG). In order to evaluate the proposed models and verify the contribution of the proposed coordination mechanisms, two multiagent benchmark applications were developed: the pursuit game and the robot soccer simulation. The results obtained demonstrated that the proposed coordination mechanisms greatly improve the performance of the multiagent system when compared with other strategies.

Performance of a neuro-fuzzy model in predicting weight changes of chronic schizophrenic patients exposed to antipsychotics

Artificial intelligence has become a possible solution to resolve the problem of loss of information when complexity of a disease increases. Obesity phenotypes are observable clinical features of drug-naive schizophrenic patients. In addition, atypical antipsychotic medications may cause these unwanted effects. Here we examined the performance of neuro-fuzzy modeling (NFM) in predicting weight changes in chronic schizophrenic patients exposed to antipsychotics. Two hundred and twenty inpatients meeting DSMIV diagnosis of schizophrenia, treated with antipsychotics, either typical or atypical, for more than 2 years, were recruited. All subjects were assessed in the same study period between mid-November 2003 and mid-April 2004. The baseline and first visit's physical data including weight, height and circumference were used in this study. Clinical information (Clinical Global Impression and Life Style Survey) and genotype data of five single nucleotide polymorphisms were also included as predictors. The subjects were randomly assigned into the first group (105 subjects) and second group (115 subjects), and NFM was performed by using the FuzzyTECH 5.54 software package, with a network-type structure constructed in the rule block. A complete learned model trained from merged data of the first and second groups demonstrates that, at a prediction error of 5, 93% subjects with weight gain were identified. Our study suggests that NFM is a feasible prediction tool for obesity in schizophrenic patients exposed to antipsychotics, with further improvements required.