Pub Date : 2017-06-02DOI: 10.1109/EAIS.2017.7954835
G. Andonovski, P. Angelov, S. Blažič, I. Škrjanc
This paper presents an autonomous Robust Evolving Cloud-based Controller (RECCo). The control algorithm is a fuzzy type with non-parametric (cloud-based) antecedent part and adaptive PID-R consequent part. The procedure starts with zero clouds (fuzzy rules) and the structure evolves during performing the process control. The PID-R parameters of the first cloud are initialized with zeros and furthermore, they are adapted on-line with a stable adaptation mechanism based on Lyapunov approach. The RECCo controller does not require any mathematical model of the controlled process but just basic information such as input and output range and the estimated value of the dominant time constant. Due to the problem space normalization the design parameters are fixed. The proposed controller with the same initial design parameters was tested on two different simulation examples. The experimental results show the convergence of the adaptive parameters and the effectiveness of the proposed algorithm.
{"title":"Robust Evolving Cloud-based Controller (RECCo)","authors":"G. Andonovski, P. Angelov, S. Blažič, I. Škrjanc","doi":"10.1109/EAIS.2017.7954835","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954835","url":null,"abstract":"This paper presents an autonomous Robust Evolving Cloud-based Controller (RECCo). The control algorithm is a fuzzy type with non-parametric (cloud-based) antecedent part and adaptive PID-R consequent part. The procedure starts with zero clouds (fuzzy rules) and the structure evolves during performing the process control. The PID-R parameters of the first cloud are initialized with zeros and furthermore, they are adapted on-line with a stable adaptation mechanism based on Lyapunov approach. The RECCo controller does not require any mathematical model of the controlled process but just basic information such as input and output range and the estimated value of the dominant time constant. Due to the problem space normalization the design parameters are fixed. The proposed controller with the same initial design parameters was tested on two different simulation examples. The experimental results show the convergence of the adaptive parameters and the effectiveness of the proposed algorithm.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117098017","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2017-05-31DOI: 10.1109/EAIS.2017.7954831
Xiaowei Gu, P. Angelov
In this paper, a new approach for autonomous anomaly detection is introduced within the Empirical Data Analytics (EDA) framework. This approach is fully data-driven and free from thresholds. Employing the nonparametric EDA estimators, the proposed approach can autonomously detect anomalies in an objective way based on the mutual distribution and ensemble properties of the data. The proposed approach firstly identifies the potential anomalies based on two EDA criteria, and then, partitions them into shape-free, non-parametric data clouds. Finally, it identifies the anomalies in regards to each data cloud (locally). Numerical examples based on synthetic and benchmark datasets demonstrate the validity and efficiency of the proposed approach.
{"title":"Autonomous anomaly detection","authors":"Xiaowei Gu, P. Angelov","doi":"10.1109/EAIS.2017.7954831","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954831","url":null,"abstract":"In this paper, a new approach for autonomous anomaly detection is introduced within the Empirical Data Analytics (EDA) framework. This approach is fully data-driven and free from thresholds. Employing the nonparametric EDA estimators, the proposed approach can autonomously detect anomalies in an objective way based on the mutual distribution and ensemble properties of the data. The proposed approach firstly identifies the potential anomalies based on two EDA criteria, and then, partitions them into shape-free, non-parametric data clouds. Finally, it identifies the anomalies in regards to each data cloud (locally). Numerical examples based on synthetic and benchmark datasets demonstrate the validity and efficiency of the proposed approach.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116872128","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2017-05-31DOI: 10.1109/EAIS.2017.7954832
P. Angelov, Xiaowei Gu
In this paper, a new type of 0-order multi-model classifier, called Autonomous Learning Multiple-Model (ALMMo-0), is proposed. The proposed classifier is non-iterative, feedforward and entirely data-driven. It automatically extracts the data clouds from the data per class and forms 0-order AnYa type fuzzy rule-based (FRB) sub-classifier for each class. The classification of new data is done using the “winner takes all” strategy according to the scores of confidence generated objectively based on the mutual distribution and ensemble properties of the data by the sub-classifiers. Numerical examples based on benchmark datasets demonstrate the high performance and computation-efficiency of the proposed classifier.
{"title":"Autonomous learning multi-model classifier of 0-Order (ALMMo-0)","authors":"P. Angelov, Xiaowei Gu","doi":"10.1109/EAIS.2017.7954832","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954832","url":null,"abstract":"In this paper, a new type of 0-order multi-model classifier, called Autonomous Learning Multiple-Model (ALMMo-0), is proposed. The proposed classifier is non-iterative, feedforward and entirely data-driven. It automatically extracts the data clouds from the data per class and forms 0-order AnYa type fuzzy rule-based (FRB) sub-classifier for each class. The classification of new data is done using the “winner takes all” strategy according to the scores of confidence generated objectively based on the mutual distribution and ensemble properties of the data by the sub-classifiers. Numerical examples based on benchmark datasets demonstrate the high performance and computation-efficiency of the proposed classifier.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127557838","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2017-05-01DOI: 10.1109/EAIS.2017.7954819
Mihael Sok, Eva Svegl, I. Grabec
A sensory-neural network for automatic diagnosing of diseases is described. The network gathers information using the patient's answers to a questionnaire. Specific questions correspond to sensors that react when patients acknowledge symptoms. The signals from the sensors stimulate neurons in which the characteristics of the disease are stored in terms of synaptic weights assigned to indicators of symptoms. The response of a neuron is determined by the weighted sum of input stimuli. The disease corresponding to the most excited neuron represents the result of diagnosis. Its reliability is assessed by the likelihood defined as the relative excitation of the neuron with respect to all others. The performance of the network is demonstrated through characteristic examples of diagnosis.
{"title":"A sensory-neural network for medical diagnosis","authors":"Mihael Sok, Eva Svegl, I. Grabec","doi":"10.1109/EAIS.2017.7954819","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954819","url":null,"abstract":"A sensory-neural network for automatic diagnosing of diseases is described. The network gathers information using the patient's answers to a questionnaire. Specific questions correspond to sensors that react when patients acknowledge symptoms. The signals from the sensors stimulate neurons in which the characteristics of the disease are stored in terms of synaptic weights assigned to indicators of symptoms. The response of a neuron is determined by the weighted sum of input stimuli. The disease corresponding to the most excited neuron represents the result of diagnosis. Its reliability is assessed by the likelihood defined as the relative excitation of the neuron with respect to all others. The performance of the network is demonstrated through characteristic examples of diagnosis.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125170589","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2017-05-01DOI: 10.1109/EAIS.2017.7954836
E. Lughofer, Mahardhika Pratama, I. Škrjanc
We propose an incremental rule splitting concept for generalized fuzzy rules in evolving fuzzy regression models in order to properly react on gradual drifts and to compensate inappropriate settings of rule evolution parameters; both occurrences may lead to oversized rules with untypically large local errors, which also usually affects the global model error. The generalized rules are directly defined in the multi-dimensional feature space through a kernel function, and thus allowing any rotated orientation of their shapes. Our splitting condition is based 1.) on the local error of rules measured in terms of a weighted contribution to the whole model error and 2.) on the size of the rules measured in terms of its volume. Thereby, we use the concept of statistical process control for automatic thresholding, in order to omit two extra parameters. The splitting technique relies on the eigendecompisition of the rule covariance matrix by adequately manipulating the largest eigenvector and eigenvalues in order to retrieve the new centers and contours of the two split rules. Thus, splitting is performed along the main principal component direction of a rule. The splitting concepts are integrated in the generalized smart evolving learning engine (Gen-Smart-EFS) and successfully tested on two real-world application scenarios, engine test benches and rolling mills, the latter including a real-occurring gradual drift (whose position in the data is known). Results show clearly improved error trend lines over time when splitting is applied: reduction of the error by about one third (rolling mills) and one half (engine test benches). In case of rolling mills, three rule splits right after the gradual drift starts were essential for this significant improvement.
{"title":"Incremental rule splitting in generalized evolving fuzzy regression models","authors":"E. Lughofer, Mahardhika Pratama, I. Škrjanc","doi":"10.1109/EAIS.2017.7954836","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954836","url":null,"abstract":"We propose an incremental rule splitting concept for generalized fuzzy rules in evolving fuzzy regression models in order to properly react on gradual drifts and to compensate inappropriate settings of rule evolution parameters; both occurrences may lead to oversized rules with untypically large local errors, which also usually affects the global model error. The generalized rules are directly defined in the multi-dimensional feature space through a kernel function, and thus allowing any rotated orientation of their shapes. Our splitting condition is based 1.) on the local error of rules measured in terms of a weighted contribution to the whole model error and 2.) on the size of the rules measured in terms of its volume. Thereby, we use the concept of statistical process control for automatic thresholding, in order to omit two extra parameters. The splitting technique relies on the eigendecompisition of the rule covariance matrix by adequately manipulating the largest eigenvector and eigenvalues in order to retrieve the new centers and contours of the two split rules. Thus, splitting is performed along the main principal component direction of a rule. The splitting concepts are integrated in the generalized smart evolving learning engine (Gen-Smart-EFS) and successfully tested on two real-world application scenarios, engine test benches and rolling mills, the latter including a real-occurring gradual drift (whose position in the data is known). Results show clearly improved error trend lines over time when splitting is applied: reduction of the error by about one third (rolling mills) and one half (engine test benches). In case of rolling mills, three rule splits right after the gradual drift starts were essential for this significant improvement.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128435647","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2017-05-01DOI: 10.1109/EAIS.2017.7954833
Rim Ben Ali, H. Schulte, A. Mami
This paper presents a dynamic modeling and simulation of a small wind turbine connected to a Permanent Magnet Synchronous Generator PMSG, which is controlled by a Zero d-axis current (ZDC) control scheme. The model of the wind turbine system is developed using basic mathematical equations and it is carried out using the Matlab /Simulink environment. The simulation results demonstrate the effectiveness of the proposed mathematical model of the small wind turbine to determine its dynamic behaviors.
{"title":"Modeling and simulation of a small wind turbine system based on PMSG generator","authors":"Rim Ben Ali, H. Schulte, A. Mami","doi":"10.1109/EAIS.2017.7954833","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954833","url":null,"abstract":"This paper presents a dynamic modeling and simulation of a small wind turbine connected to a Permanent Magnet Synchronous Generator PMSG, which is controlled by a Zero d-axis current (ZDC) control scheme. The model of the wind turbine system is developed using basic mathematical equations and it is carried out using the Matlab /Simulink environment. The simulation results demonstrate the effectiveness of the proposed mathematical model of the small wind turbine to determine its dynamic behaviors.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124565611","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2017-05-01DOI: 10.1109/EAIS.2017.7954820
M. Stepancic, J. Kocijan
The on-line identification of nonlinear dynamical system with the regularised nonparametric regression approach is considered. The model structure is a nonlinear finite impulse response (NFIR) based on a Gaussian process (GP). The online estimation of the tuning parameters of the GP model leads to an Evolving Gaussian process whose structure adapts to the current dynamics of the measured system. The GP regression is a kernel method which requires storing the past measurements. The kernel-based on-line system identification is implementable only with a constraint on the amount of data stored. The on-line identification method combines together the forgetting factor for discounting old data and the moving window which neglects the highly discounted data. As a consequence, the online-identification problem may be ill-posed due to the discounted data. A regularisation approach is introduced for the estimation of the tuning parameters in order to avoid the ill-posed identification problem. The performance of the online identification method is demonstrated with an illustrative example.
{"title":"On-line identification with regularised Evolving Gaussian process","authors":"M. Stepancic, J. Kocijan","doi":"10.1109/EAIS.2017.7954820","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954820","url":null,"abstract":"The on-line identification of nonlinear dynamical system with the regularised nonparametric regression approach is considered. The model structure is a nonlinear finite impulse response (NFIR) based on a Gaussian process (GP). The online estimation of the tuning parameters of the GP model leads to an Evolving Gaussian process whose structure adapts to the current dynamics of the measured system. The GP regression is a kernel method which requires storing the past measurements. The kernel-based on-line system identification is implementable only with a constraint on the amount of data stored. The on-line identification method combines together the forgetting factor for discounting old data and the moving window which neglects the highly discounted data. As a consequence, the online-identification problem may be ill-posed due to the discounted data. A regularisation approach is introduced for the estimation of the tuning parameters in order to avoid the ill-posed identification problem. The performance of the online identification method is demonstrated with an illustrative example.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121847697","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2017-05-01DOI: 10.1109/EAIS.2017.7954823
M. Pătrașcu, Adrian Patrascu, I. Beres
Reinterpretation and adaptation of knowledge from technical sciences into the field of sports science is at the forefront of advancing performance. Evolution-based systems open the possibility of evaluating different stress factors that appear in the systematization of training programs. Our aim was to demonstrate that it is possible to use evolution mechanisms to model the real life influence of stress factors during performance training of athletes, a concept that can be generalized to other supervised or controlled systems. We have recorded data from one former basketball player to be used in the development of a simulation model based on immune genetic algorithms. We developed an immunization scheme that is parameterized for simulating different training outcomes based on type of athlete. Results confirm one of these cases is possibly the most similar to real life situations. Thus, we obtained an evolution model that aligns with the generative experiment as proof of concept for the evaluation of performance under stress. In particular, for sports science, we may have found a way to analyze training programs before their execution and to spot weaknesses in them.
{"title":"An immune evolution mechanism for the study of stress factors in supervised and controlled systems","authors":"M. Pătrașcu, Adrian Patrascu, I. Beres","doi":"10.1109/EAIS.2017.7954823","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954823","url":null,"abstract":"Reinterpretation and adaptation of knowledge from technical sciences into the field of sports science is at the forefront of advancing performance. Evolution-based systems open the possibility of evaluating different stress factors that appear in the systematization of training programs. Our aim was to demonstrate that it is possible to use evolution mechanisms to model the real life influence of stress factors during performance training of athletes, a concept that can be generalized to other supervised or controlled systems. We have recorded data from one former basketball player to be used in the development of a simulation model based on immune genetic algorithms. We developed an immunization scheme that is parameterized for simulating different training outcomes based on type of athlete. Results confirm one of these cases is possibly the most similar to real life situations. Thus, we obtained an evolution model that aligns with the generative experiment as proof of concept for the evaluation of performance under stress. In particular, for sports science, we may have found a way to analyze training programs before their execution and to spot weaknesses in them.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123227197","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2017-05-01DOI: 10.1109/EAIS.2017.7954830
S. Blažič, Andrej Zdešar
A fuzzy model reference adaptive control approach is proposed in the paper where the antecedent part of fuzzy rules evolves with the measured data. The consequent part consists of a controller with integral nature and the adaptation scheme is a direct one. The proposed algorithm is capable of controlling a plant with poorly known and/or time-varying nonlinearity which is an advantage over approaches with fixed antecedent part. It is intended for control of a large class of nonlinear plant models with the dominant dynamics of the first order. Such plants occur quite often in process industries. It is shown in the paper that the approach is also suitable for controlling an under-damped mechanical system.
{"title":"An implementation of an evolving fuzzy controller","authors":"S. Blažič, Andrej Zdešar","doi":"10.1109/EAIS.2017.7954830","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954830","url":null,"abstract":"A fuzzy model reference adaptive control approach is proposed in the paper where the antecedent part of fuzzy rules evolves with the measured data. The consequent part consists of a controller with integral nature and the adaptation scheme is a direct one. The proposed algorithm is capable of controlling a plant with poorly known and/or time-varying nonlinearity which is an advantage over approaches with fixed antecedent part. It is intended for control of a large class of nonlinear plant models with the dominant dynamics of the first order. Such plants occur quite often in process industries. It is shown in the paper that the approach is also suitable for controlling an under-damped mechanical system.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129401688","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2017-05-01DOI: 10.1109/EAIS.2017.7954826
Leandro Maciel, R. Vieira, Alisson Porto, F. Gomide, R. Ballini
Financial interval time series (ITS) describe the evolution of the maximum and minimum prices of an asset throughout time. These price ranges are related to the concept of volatility. Hence, their accurate forecasts play a key role in risk management, derivatives pricing and asset allocation, as well as supplements the information extracted by the time series of the closing price values. This paper addresses evolving fuzzy systems and financial ITS forecasting considering as the empirical application the main index of the Brazilian stock market, the IBOVESPA. An evolving participatory learning fuzzy model, named ePL-KRLS, is proposed. The model extends traditional ePL approach by considering Kernel functions to the identification of rule consequents parameters as well as a metaheuristic algorithm to automatically set model control parameters. One step ahead interval forecasts is compared against linear and nonlinear time series benchmark methods and with the state of the art evolving fuzzy models in terms of traditional accuracy metrics and quality measures designed for ITS. The results provide evidence for the predictability of of IBOVESPA ITS and significant forecast contribution of ePL-KRLS.
{"title":"Evolving participatory learning fuzzy modeling for financial interval time series forecasting","authors":"Leandro Maciel, R. Vieira, Alisson Porto, F. Gomide, R. Ballini","doi":"10.1109/EAIS.2017.7954826","DOIUrl":"https://doi.org/10.1109/EAIS.2017.7954826","url":null,"abstract":"Financial interval time series (ITS) describe the evolution of the maximum and minimum prices of an asset throughout time. These price ranges are related to the concept of volatility. Hence, their accurate forecasts play a key role in risk management, derivatives pricing and asset allocation, as well as supplements the information extracted by the time series of the closing price values. This paper addresses evolving fuzzy systems and financial ITS forecasting considering as the empirical application the main index of the Brazilian stock market, the IBOVESPA. An evolving participatory learning fuzzy model, named ePL-KRLS, is proposed. The model extends traditional ePL approach by considering Kernel functions to the identification of rule consequents parameters as well as a metaheuristic algorithm to automatically set model control parameters. One step ahead interval forecasts is compared against linear and nonlinear time series benchmark methods and with the state of the art evolving fuzzy models in terms of traditional accuracy metrics and quality measures designed for ITS. The results provide evidence for the predictability of of IBOVESPA ITS and significant forecast contribution of ePL-KRLS.","PeriodicalId":286312,"journal":{"name":"2017 Evolving and Adaptive Intelligent Systems (EAIS)","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126516210","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：