Recent advances in neuro-fuzzy system: A survey

Actuator fault detection and isolation on multi-rotor UAV using extreme learning neuro-fuzzy systems

Undetected partial actuator faults on multi-rotor UAVs can lead to system failures and uncontrolled crashes, necessitating the development of accurate and efficient fault detection and isolation (FDI) strategy. This paper proposes a hybrid FDI model for a quadrotor UAV that integrates an extreme learning neuro-fuzzy algorithm with a model-based extended Kalman filter (EKF). Three FDI models using Fuzzy-ELM, R-EL-ANFIS, and EL-ANFIS are compared based on training, validation performances, and sensitivity to weaker and shorter actuator faults. They are also tested online for linear and nonlinear incipient faults by measuring their isolation time delays and accuracies. The results show that the Fuzzy-ELM FDI model exhibits greater efficiency and sensitivity, while Fuzzy-ELM and R-EL-ANFIS FDI models demonstrate better performance than a conventional neuro-fuzzy algorithm, ANFIS.

Fatigue Life Modelling of Steel Suspension Coil Springs Based on Wavelet Vibration Features Using Neuro-Fuzzy Methods

This study proposed wavelet-based approaches to characterise random vibration road excitations for durability prediction of coil springs. Conventional strain-life approaches require long computational time, while the accuracy of the vibration fatigue methods is unsatisfactory. It is therefore a necessity to establish an accurate fatigue life prediction model based on vibrational features. Wavelet-based methods were applied to determine the low-frequency energy and multifractality of road excitations. Strain-life models were applied for fatigue life evaluation from strain histories. ANFIS modelling was subsequently adopted to associate the vibration features with the fatigue life of coil springs. Results showed that the proposed wavelet-based methods were effective to determine the signal energy and multifractality of vibration signals. The established vibration-based models showed good fatigue life conservativity with a data survivability of more than 90%. The highest Pearson coefficient of 0.955 associated with the lowest RMSE of 0.660 was obtained by the Morrow-based model. It is suggested that the low-frequency energy and multifractality of the vibration signals can be used as fatigue-related features in life predictions of coil springs under random loading. Finally, the proposed model is an acceptable fatigue life prediction method based on vibration features, and it can reduce the dependency on strain data measurement.

Developing Nonlinear Customer Preferences Models for Product Design Using Opining Mining and Multiobjective PSO-Based ANFIS Approach

Online customer reviews can clearly show the customer experience, and the improvement suggestions based on the experience, which are helpful to product optimization and design. However, the research on establishing a customer preference model based on online customer reviews is not ideal, and the following research problems are found in previous studies. Firstly, the product attribute is not involved in the modelling if the corresponding setting cannot be found in the product description. Secondly, the fuzziness of customers' emotions in online reviews and nonlinearity in the models were not appropriately considered. Thirdly, the adaptive neuro-fuzzy inference system (ANFIS) is an effective way to model customer preferences. However, if the number of inputs is large, the modelling process will be failed due to the complex structure and long computational time. To solve the above-given problems, this paper proposed multiobjective particle swarm optimization (PSO) based ANFIS and opinion mining, to build customer preference model by analyzing the content of online customer reviews. In the process of online review analysis, the opinion mining technology is used to conduct comprehensive analysis on customer preference and product information. According to the analysis of information, a new method for establishing customer preference model is proposed, that is, a multiobjective PSO based ANFIS. The results show that the introducing of multiobjective PSO method into ANFIS can effectively solve the defects of ANFIS itself. Taking hair dryer as a case study, it is found that the proposed approach performs better than fuzzy regression, fuzzy least-squares regression, and genetic programming based fuzzy regression in modelling customer preference.

Using neural networks and a fuzzy inference system to evaluate the risk of wildfires and the pinpointing of firefighting stations in forests on the northern slopes of the Zagros Mountains, Iran (case study: Shimbar national wildlife preserve)

Predicting potential fire hazard zones in natural areas is one of the means of mitigating and managing fires. The current research focuses on the prioritizing of elements which contribute to the spread of fire and the special zoning of potentially dangerous areas in addition to the pinpointing of locations for the establishment of fire stations in forested areas in the Shimbar national reserve based on historical data spanning 2001 to 2018. The study utilizes elements (physiological, vegetation cover, meteorological, anthropological factors) contributing to wildfires as inputs into an artificial neural network and the development of a fuzzy inference system in order to produce fire zoning maps for the region under study. The map is divided into five sectors, i.e., minimum, low, moderate, high, and maximum risk of fire. The validation of the fire zoning map was evaluated at 0.83 and the RMSE error was 0.75. The results obtained show that 20% of the area under study is within the average risk category, 11% is within the high-risk category, and 10% is within the very high-risk category of a potential fire hazard. The most important variables were distance from a flowing source, i.e., river or stream, the land formation type, elevation, and the minimum temperature. The identification of suitable locations for firefighting stations was carried out by merging the fuzzy inference system model and Arc GIS, and the results obtained defined 16 possible locations. It was concluded that the application of hybrid models when dealing with the aforementioned variables is effective when seeking to determine locations for the establishment of firefighting stations and rural safety services; moreover, such hybrid models are highly efficacious for determining of fire hazard zones. It is proposed that hybrid models be applied on a large scale for the prevention, control, and management of fires throughout the country.

Disease categorization with clinical data using optimized bat algorithm and fuzzy value

Diagnosis of human disease is a more difficult and complex process since it requires the consideration of various factors and symptoms to make a decision. Generally, the classification of diseases with fuzzy values is the most interesting topic because of accurate results. In this paper, we design a Bat-based Random Forest (BbRF) framework to enhance the performance of categorizing diseases with fuzzy values which also protect the privacy of the developed scheme. It involves pre-processing, attributes selection, fuzzy value generation, and classification. Additionally, the developed framework is implemented in Python tool and patient disease datasets are used for implementation. Moreover, pre-processing remove the error and noise, attributes are selected based on the duration of diseases. Finally, classify the patient disease based on the generated fuzzy value. To prove the efficiency of the developed framework, attained results are compared with other existing techniques in terms of accuracy, sensitivity, specificity, F-measure, and precision.

ReNFuzz-LF: A Recurrent Neurofuzzy System for Short-Term Load Forecasting

A neurofuzzy system is proposed for short-term electric load forecasting. The fuzzy rule base of ReNFuzz-LF consists of rules with dynamic consequent parts that are small-scale recurrent neural networks with one hidden layer, whose neurons have local output feedback. The particular representation maintains the local learning nature of the typical static fuzzy model, since the dynamic consequent parts of the fuzzy rules can be considered as subsystems operating at the subspaces defined by the fuzzy premise parts, and they are interconnected through the defuzzification part. The Greek power system is examined, and hourly based predictions are extracted for the whole year. The recurrent nature of the forecaster leads to the use of a minimal set of inputs, since the temporal relations of the electric load time-series are identified without any prior knowledge of the appropriate past load values being necessary. An extensive simulation analysis is conducted, and the forecaster’s performance is evaluated using appropriate metrics (APE, RMSE, forecast error duration curve). ReNFuzz-LF performs efficiently, attaining an average percentage error of 1.35% and an average yearly absolute error of 86.3 MW. Finally, the performance of the proposed forecaster is compared to a series of Computational Intelligence based models, such that the learning characteristics of ReNFuzz-LF are highlighted.

Study on Resistant Hierarchical Fuzzy Neural Networks

Novel resistant hierarchical fuzzy neural networks are proposed in this study and their deep learning problems are investigated. These fuzzy neural networks can be used to model complex controlled plants and can also be used as fuzzy controllers. In general, real-world data are usually contaminated by outliers. These outliers may have undesirable or unpredictable influences on the final learning machines. The correlations between the target and each of the predictors are utilized to partition input variables into groups so that each group becomes the input variables of a fuzzy system in each level of the hierarchical fuzzy neural network. In order to enhance the resistance of the learning machines, we use the least trimmed squared error as the cost function. To test the resistance of learning machines to adverse effects of outliers, we add at the output node some noise from three different types of distributions, namely, normal, Laplace, and uniform distributions. Real-world datasets are used to compare the performances of the proposed resistant hierarchical fuzzy neural networks, resistant densely connected artificial neural networks, and densely connected artificial neural networks without noise.

A Hybrid Shuffled Frog Leaping Algorithm and Its Performance Assessment in Multi-Dimensional Symmetric Function

Ensemble learning of swarm intelligence evolutionary algorithm of artificial neural network (ANN) is one of the core research directions in the field of artificial intelligence (AI). As a representative member of swarm intelligence evolutionary algorithm, shuffled frog leaping algorithm (SFLA) has the advantages of simple structure, easy implementation, short operation time, and strong global optimization ability. However, SFLA is susceptible to fall into local optimas in the face of complex and multi-dimensional symmetric function optimization, which leads to the decline of convergence accuracy. This paper proposes an improved shuffled frog leaping algorithm of threshold oscillation based on simulated annealing (SA-TO-SFLA). In this algorithm, the threshold oscillation strategy and simulated annealing strategy are introduced into the SFLA, which makes the local search behavior more diversified and the ability to escape from the local optimas stronger. By using multi-dimensional symmetric function such as drop-wave function, Schaffer function N.2, Rastrigin function, and Griewank function, two groups (i: SFLA, SA-SFLA, TO-SFLA, and SA-TO-SFLA; ii: SFLA, ISFLA, MSFLA, DSFLA, and SA-TO-SFLA) of comparative experiments are designed to analyze the convergence accuracy and convergence time. The results show that the threshold oscillation strategy has strong robustness. Moreover, compared with SFLA, the convergence accuracy of SA-TO-SFLA algorithm is significantly improved, and the median of convergence time is greatly reduced as a whole. The convergence accuracy of SFLA algorithm on these four test functions are 90%, 100%, 78%, and 92.5%, respectively, and the median of convergence time is 63.67 s, 59.71 s, 12.93 s, and 8.74 s, respectively; The convergence accuracy of SA-TO-SFLA algorithm on these four test functions is 99%, 100%, 100%, and 97.5%, respectively, and the median of convergence time is 48.64 s, 32.07 s, 24.06 s, and 3.04 s, respectively.

Multi-gene genetic programming to building up fuzzy rule-base in Neo-Fuzzy-Neuron networks

This paper introduces a new approach to build the rule-base on Neo-Fuzzy-Neuron (NFN) Networks. The NFN is a Neuro-Fuzzy network composed by a set of n decoupled zero-order Takagi-Sugeno models, one for each input variable, each one containing m rules. Employing Multi-Gene Genetic Programming (MG-GP) to create and adjust Gaussian membership functions and a Gradient-based method to update the network parameters, the proposed model is dubbed NFN-MG-GP. In the proposed model, each individual of MG-GP represents a complete rule-base of NFN. The rule-base is adjusted by genetic operators (Crossover, Reproduction, Mutation), and the consequent parameters are updated by a predetermined number of Gradient method epochs, every generation. The algorithm uses Elitism to ensure that the best rule-base is not lost between generations. The performance of the NFN-MG-GP is evaluated using instances of time series forecasting and non-linear system identification problems. Computational experiments and comparisons against state-of-the-art alternative models show that the proposed algorithms are efficient and competitive. Furthermore, experimental results show that it is possible to obtain models with good accuracy applying Multi-Gene Genetic Programming to construct the rule-base on NFN Networks.

Analysis of AM Parameters on Surface Roughness Obtained in PLA Parts Printed with FFF Technology

Fused filament fabrication (FFF) 3D printing technology allows very complex parts to be obtained at a relatively low cost and in reduced manufacturing times. In the present work, the effect of main 3D printing parameters on roughness obtained in curved surfaces is addressed. Polylactic acid (PLA) hemispherical cups were printed with a shape similar to that of the acetabular part of the hip prostheses. Different experiments were performed according to a factorial design of experiments, with nozzle diameter, temperature, layer height, print speed and extrusion multiplier as variables. Different roughness parameters were measured—Ra, Rz, Rku, Rsk—both on the outer surface and on the inner surface of the parts. Arithmetical mean roughness value Ra and greatest height of the roughness profile Rz are usually employed to compare the surface finish among different manufacturing processes. However, they do not provide information about the shape of the roughness profile. For this purpose, in the present work kurtosis Rku and skewness Rsk were used. If the height distribution in a roughness profile follows a normal law, the Rku parameter will take a value of 3. If the profile distribution is symmetrical, the Rsk parameter will take a value of 0. Adaptive neural fuzzy inference system (ANFIS) models were obtained for each response. Such models are often employed to model different manufacturing processes, but their use has not yet been extended to 3D printing processes. All roughness parameters studied depended mainly on layer height, followed by nozzle diameter. In the present work, as a general trend, Rsk was close to but lower than 0, while Rku was slightly lower than 3. This corresponds to slightly higher valleys than peaks, with a rounded height distribution to some degree.

On the Application of a Design of Experiments along with an ANFIS and a Desirability Function to Model Response Variables

In manufacturing engineering, it is common to use both symmetrical and asymmetrical factorial designs along with regression techniques to model technological response variables, since the in-advance prediction of their behavior is of great importance to determine the levels of variation that lead to optimal response values to be obtained. For this purpose, regression techniques based on the response surface method combined with a desirability function for multi-objective optimization are commonly employed, since it is usual to find manufacturing processes that require simultaneous optimization of several variables, which exhibit in many cases an opposite behavior. However, these regression models are sometimes not accurate enough to predict the behavior of these response variables, especially when they have significant non-linearities. To deal with this drawback, soft computing techniques are very effective in overcoming the limitations of conventional regression models. This present study is focused on the employment of a symmetrical design of experiments along with a new desirability function, which is proposed in this study, and with soft computing techniques based on fuzzy logic. It will be shown that more accurate results than those obtained from regression techniques are obtained. Moreover, this new desirability function is analyzed in this study.

Development of adaptive neuro-fuzzy inference system model for predict trihalomethane formation potential in distribution network simulation test

Trihalomethanes (THMs), which is one of the major classes of DBP known to be highly cytotoxic and genotoxic, were formed and modeled under controlled conditions by laboratory-scale distribution network simulation test. The formation potentials of THM depending on the parameters such as natural organic matter, bromide, chlorine, pH, and contact time were determined. Subsequently, the Adaptive Neuro-Fuzzy Inference System (ANFIS) model was developed using these parameters as inputs and THM formation potentials as output, and the correlation coefficient was 0.9817. In the range of the inputs, the ANFIS model representing the simulation test results were compared with THM formations of an actual distribution network system in dry and wet seasons. As a result, the predictions of the ANFIS model were little affected by the unidentified factors that were not used in model training but are known to affect THM formations in real waters and gave more consistent results than the EPA model.

Multivariate data-based optimization of membrane adsorption process for wastewater treatment: Multi-layer perceptron adaptive neural network versus adaptive neural fuzzy inference system

Application of machine-learning methods to assess the batch adsorption of malachite green (MG) dye on chitosan/polyvinyl alcohol/zeolite imidazolate frameworks membrane adsorbents (CPZ) was investigated in this study. Our previous research results proved the suitability of the CPZ membranes for wastewater decoloring. In the current work, the residence time was combined with the other operational variables i.e., pH, initial dye concentration, and adsorbent dose (AD), to obtain the possible interactions involved in nonequilibrium adsorption. Two well-known soft-computing approaches, multi-layer perceptron adaptive neural network (MLP-ANN) and adaptive neural fuzzy inference system (ANFIS), were selected among different machine learning alternatives and then, comprehensively compared with each other considering reliability and accuracy for a 60 number of runs. The ANFIS structure with nine centers of clusters could predict the adsorption performance better than the ANN approach. Root mean square error (RMSE) and R-square were obtained 0.01822 and 0.9958 for the test data, respectively. The interpretability test resulted a linear trend predicted by the model and disclosed that the maximum value of the removal efficiency (99.5%) could be obtained when the amount of the inputs set to the upper limit. Lastly, the sensitivity analysis uncovered that the residence time has a decisive effect (relevancy factor > 80%) on the removal efficiency. According to the results, ANFIS is an effective and reliable tool to optimize and intensify the membrane adsorption process.

Nonlinear system modeling using the takagi-sugeno fuzzy model and long-short term memory cells

 The data driven black-box or gray-box models like neural networks and fuzzy systems have some disadvantages, such as the high and uncertain dimensions and complex learning process. In this paper, we combine the Takagi-Sugeno fuzzy model with long-short term memory cells to overcome these disadvantages. This novel model takes the advantages of the interpretability of the fuzzy system and the good approximation ability of the long-short term memory cell. We propose a fast and stable learning algorithm for this model. Comparisons with others similar black-box and grey-box models are made, in order to observe the advantages of the proposal.

A Proposal of an Adaptive Neuro-Fuzzy Inference System for Modeling Experimental Data in Manufacturing Engineering

In Manufacturing Engineering there is a need to be able to model the behavior of technological variables versus input parameters in order to predict their behavior in advance, so that it is possible to determine the levels of variation that lead to optimal values of the response variables to be obtained. In recent years, it has been a common practice to rely on regression techniques to carry out the above-mentioned task. However, such models are sometimes not accurate enough to predict the behavior of these response variables, especially when they have significant non-linearities. In this present study a comparative analysis between the precision of different techniques based on conventional regression and soft computing is initially carried out. Specifically, regression techniques, based on the response surface model, as well as the use of artificial neural networks and fuzzy inference systems along with adaptive neuro-fuzzy inference systems will be employed to predict the behavior of the aforementioned technological variables. It will be shown that when there are difficulties in predicting the response parameters by using regression models, soft computing models are highly effective, being much more efficient than conventional regression models. In addition, a new method is proposed in this study that consists of using an iterative process to obtain a fuzzy inference system from a design of experiments and then using an adaptive neuro-fuzzy inference system for tuning the constants of the membership functions. As will be shown, with this method it is possible to obtain improved results in the validation metrics. The means of selecting the membership functions to develop this model from the design of experiments is discussed in this present study in order to obtain an initial solution, which will be then tuned by using an adaptive neuro-fuzzy inference system, to predict the behavior of the response variables. Moreover, the obtained results will also be compared.

An intelligent system based on optimized ANFIS and association rules for power transformer fault diagnosis

This research work put forward an intelligent method for diagnosis and classification of power transformers faults based on the instructive Dissolved Gas Analysis Method (DGAM) attributes and machine learning algorithms. In the proposed method, 14 attributes obtained through DGAM are utilized as the initial and unprocessed inputs of Adaptive Neuro-Fuzzy Inference System (ANFIS). In this method, attribute selection and improved learning algorithm are utilized to enhance fault detection and recognition precision. In the propounded fault detection and classification method, the most instructive attributes obtained by DGAM are selected by association rules learning technique (ARLT). Using efficient enlightening attributes and eliminating tautological attributes lead to higher accuracy and superior operation. Furthermore, appropriate training of ANFIS has significant effect on its precision and robustness. Therefore, Black Widow Optimization Algorithm (BWOA) is applied to train the ANFIS. Having excellent exploration and extraction capability, fast convergence speed and simplicity is the main reason for choosing the BWOA as the learning algorithm. Two industrial datasets are utilized to test and evaluate the performance of the put forward method. The results show that the propounded diagnosis system has high accuracy, robust performance and short run time. Selecting the most educative attributes of DGAM, training ANFIS optimally, improving the robustness of ANFIS and increasing the classification accuracy are the main contribution of this paper in the field of power transformer fault detection and classification.

Utilizing an Adaptive Neuro-Fuzzy Inference System (ANFIS) for Overcrowding Level Risk Assessment in Railway Stations

The railway network plays a significant role (both economically and socially) in assisting the reduction of urban traffic congestion. It also accelerates the decarbonization in cities, societies and built environments. To ensure the safe and secure operation of stations and capture the real-time risk status, it is imperative to consider a dynamic and smart method for managing risk factors in stations. In this research, a framework to develop an intelligent system for managing risk is suggested. The adaptive neuro-fuzzy inference system (ANFIS) is proposed as a powerful, intelligently selected model to improve risk management and manage uncertainties in risk variables. The objective of this study is twofold. First, we review current methods applied to predict the risk level in the flow. Second, we develop smart risk assessment and management measures (or indicators) to improve our understanding of the safety of railway stations in real-time. Two parameters are selected as input for the risk level relating to overcrowding: the transfer efficiency and retention rate of the platform. This study is the world’s first to establish the hybrid artificial intelligence (AI) model, which has the potency to manage risk uncertainties and learns through artificial neural networks (ANNs) by integrated training processes. The prediction result shows very high accuracy in predicting the risk level performance, and proves the AI model capabilities to learn, to make predictions, and to capture risk level values in real time. Such risk information is extremely critical for decision making processes in managing safety and risks, especially when uncertain disruptions incur (e.g., COVID-19, disasters, etc.). The novel insights stemmed from this study will lead to more effective and efficient risk management for single and clustered railway station facilities towards safer, smarter, and more resilient transportation systems.

Hybrid modelling of biological systems using fuzzy continuous Petri nets

Integrated modelling of biological systems is challenged by composing components with sufficient kinetic data and components with insufficient kinetic data or components built only using experts’ experience and knowledge. Fuzzy continuous Petri nets (FCPNs) combine continuous Petri nets with fuzzy inference systems, and thus offer an hybrid uncertain/certain approach to integrated modelling of such biological systems with uncertainties. In this paper, we give a formal definition and a corresponding simulation algorithm of FCPNs, and briefly introduce the FCPN tool that we have developed for implementing FCPNs. We then present a methodology and workflow utilizing FCPNs to achieve hybrid (uncertain/certain) modelling of biological systems illustrated with a case study of the Mercaptopurine metabolic pathway. We hope this research will promote the wider application of FCPNs and address the uncertain/certain integrated modelling challenge in the systems biology area.

Fuzzy inference system for modeling the environmental risk map of air pollutants in Taiwan

This study aimed to improve the uncertainty in spatial data of risk assessment through a Fuzzy inference system (FIS) as a way to conduct an environmental risk map of air pollution in Taiwan. In modeling, the feature inputs of FIS included the geographic coordinates and time, while the outputs are the pollutant concentrations. The outputs are supplements to the concentration contour on the map in comparison with Kriging interpolation. In our model, the FIS was designed using the official open data of air pollutants, including Pb and PM_2.5 that were collected from the monitoring stations in mid-southern Taiwan. The model involved data filtration and imputation in the preliminary scheme to extract the historical data for analysis. We used the data of Pb (2001-2013) and PM_2.5 (2006-2013) for the training process, and then used the data from 2014 to 2015 for validation. Our model was able to compute the smaller errors of inferred and measured values of Pb and PM_2.5 than the conventional method. The approach was applied to deduce the exposure of PM_2.5 distributed over the Taiwan Island in accordance with the governmental open data of seventy-three stations during 2006-2016 in order to produce our risk map. The designed model upon Fuzzy inference accesses potential risks of spatiotemporal exposures in the unmeasured locations with feasibility and adaptability for environmental management.

Hybrid Model for Credit Risk Prediction: An Application of Neural Network Approaches

Credit risk prediction is essential for banks and financial institutions as it helps them to evade any inappropriate assessments that can lead to wasted opportunities or monetary losses. In recent times, the hybrid prediction model, a combination of traditional and modern artificial intelligence (AI) methods that provides better prediction capacity than the use of single techniques, has been introduced. Similarly, using conventional and topical artificial intelligence (AI) technologies, researchers have recommended hybrid models which amalgamate logistic regression (LR) with multilayer perceptron (MLP). To investigate the efficiency and viability of the proposed hybrid models, we compared 16 hybrid models created by combining logistic regression (LR), discriminant analysis (DA), and decision trees (DT) with four types of neural network (NN): adaptive neurofuzzy inference systems (ANFISs), deep neural networks (DNNs), radial basis function networks (RBFs) and multilayer perceptrons (MLPs). The experimental outcome, investigation, and statistical examination express the capacity of the planned hybrid model to develop a credit risk prediction technique different from all other approaches, as indicated by ten different performance measures. The classifier was authenticated on five real-world credit scoring data sets.

Fuzzy Credit Risk Scoring Rules using FRvarPSO

There is consensus that the best way for reducing insolvency situations in financial institutions is through good risk management, which involves a good client selection process. In the market, there are methodologies for credit scoring, each analyzing a large number of microeconomic and/or macroeconomic variables selected mostly depending on the type of credit to be granted. Since these variables are heterogeneous, the review process carried out by credit analysts takes time. The objective of this article is to propose a solution for this problem by applying fuzzy logic to the creation of classification rules for credit granting. To achieve this, linguistic variables were used to help the analyst interpret the information available from the credit officer. The method proposed here combines the use of fuzzy logic with a neural network and a variable population optimization technique to obtain fuzzy classification rules. It was tested with three databases from financial entities in Ecuador — one credit and savings cooperative and two banks that grant various types of credits. To measure its performance, three benchmarks were used: accuracy, number of classification rules generated, and antecedent length. The results obtained indicate that the hybrid model that is proposed performs better than its previous versions due to the addition of fuzzy logic. At the end of the article, our conclusions are discussed and future research lines are suggested.