Abstract
The existing uncertainties during the operation of processes could strongly affect the performance of forecasting systems, control strategies and fault detection systems when they are not considered in the design. Because of that, the study of uncertainty quantification has gained more attention among the researchers during past decades. From this field of study, the prediction intervals arise as one of the techniques most used in literature to represent the effect of uncertainty over the future process behavior. Thus, researchers have focused on developing prediction intervals based on the use of fuzzy systems and neural networks, thanks to their usefulness for represent a wide range of processes as universal approximators. In this work, a review of the state-of-the-art of methodologies for prediction interval modelling based on fuzzy systems and neural networks is presented. The main characteristics of each method for prediction interval construction are presented and some recommendations are given for selecting the most appropriate method for specific applications. To illustrate the advantages of these methodologies, a comparative analysis of selected methods of prediction intervals is presented, using a benchmark series and real data from solar power generation of a microgrid.
Highlights
The use of fuzzy logic systems (FLS) and neural networks (NNs) has proliferated in the literature for the modeling of systems and time series
Where is a zero mean random variable called data noise that is responsible for introducing uncertainty modeling into predictive models in the form of aleatory uncertainty, whose origin can be traced to the exclusion of complicated variables from the model which cannot be determined with sufficient precision, or due to the presence of inherently stochastic processes in the observed system or data obtention procedure
In this review we focus on methods based on computational intelligence that directly construct the prediction interval (PI) for dynamical systems, in particular those based on fuzzy models and neural networks
Summary
The use of fuzzy logic systems (FLS) and neural networks (NNs) has proliferated in the literature for the modeling of systems and time series. In order to fully understand the information provided by Prediction Intervals, it is important to study what are the possible causes of uncertainty in dynamical system modeling, and on what phenomena they can originate. Where is a zero mean random variable called data noise that is responsible for introducing uncertainty modeling into predictive models in the form of aleatory uncertainty, whose origin can be traced to the exclusion of complicated variables from the model which cannot be determined with sufficient precision, or due to the presence of inherently stochastic processes in the observed system or data obtention procedure Based on this formulation, predictive models attempt to produce an estimate f(x) of the data generating function in order to calculate predictions of the expected value of the system. It is important to note that since total uncertainty can come from many diverse sources, the expression can be highly complex and difficult to quantify, which is why interval modeling has been proposed as a solution to this problem
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
