Mahboobeh Parsapoor

Chaotic Time Series Prediction Using Brain Emotional Learning Based Recurrent Fuzzy System (BELRFS)

In this paper, an architecture based on the anatomical structure of the emotional network in the brain of mammalians is applied as a prediction model for chaotic time series studies. The architecture is called Brain Emotional Learning-based Recurrent Fuzzy System (BELRFS), which stands for: Brain Emotional Learning-based Recurrent Fuzzy System. It adopts neuro-fuzzy adaptive networks to mimic the functionality of brain emotional learning. In particular, the model is investigated to predict space storms, since the phenomenon has been recognised as a threat to critical infrastructure in modern society. To evaluate the performance of BELRFS, three benchmark time series: Lorenz time series, sunspot number time series and Auroral Electrojet (AE) index. The obtained results of BELRFS are compared with Linear Neuro-Fuzzy (LNF) with the Locally Linear Model Tree algorithm (LoLiMoT). The results indicate that the suggested model outperforms most of data driven models in terms of prediction accuracy.

Brain Emotional Learning Based Fuzzy Inference System (BELFIS) for Solar Activity Forecasting

This paper presents a new architecture based on a brain emotional learning model that can be used in a wide varieties of AI applications such as prediction, identification and classification. The architecture is referred to as: Brain Emotional Learning Based Fuzzy Inference System (BELFIS) and it is developed from merging the idea of prior emotional models with fuzzy inference systems. The main aim of this model is presenting a desirable learning model for chaotic system prediction imitating the brain emotional network. In this research work, the model is used for predicting the solar activity, since it has been recognized as a threat to critical infrastructures in modern society. Specifically sunspot numbers are predicted by applying the proposed brain emotional learning model. The prediction results are compared with the outcomes of using other previous models like the locally linear model tree (LOLIMOT) and radial bias function (RBF) and adaptive neuro-fuzzy inference system (ANFIS).

Ant Colony Optimization for Channel Assignment Problem in a Clustered Mobile Ad Hoc Network

This paper presents an ant colony optimization (ACO) method as a method for channel assignment in a mobile ad hoc network (MANET), where achieving high spectral efficiency necessitates an efficient channel assignment. The suggested algorithm is intended for graph-coloring problems and it is specifically tweaked to the channel assignment problem in MANET with a clustered network topology. A multi-objective function is designed to make a tradeoff between maximizing spectral utilization and minimizing interference. We compare the convergence behavior and performance of ACO-based method with obtained results from a grouping genetic algorithm (GGA).

Neuro-fuzzy models, BELRFS and LoLiMoT, for prediction of chaotic time series

This paper suggests a novel learning model for prediction of chaotic time series, brain emotional learning-based recurrent fuzzy system (BELRFS). The prediction model is inspired by the emotional learning system of the mammal brain. BELRFS is applied for predicting Lorenz and Ikeda time series and the results are compared with the results from a prediction model based on local linear neuro-fuzzy models with linear model tree algorithm (LoLiMoT).

A Brain Emotional Learning-based Prediction Model for the prediction of geomagnetic storms

This paper introduces a new type of brain emotional learning inspired models (BELIMs). The suggested model is utilized as a suitable model for predicting geomagnetic storms. The model is known as BELPM which is an acronym for Brain Emotional Learning-based Prediction Model. The structure of the suggested model consists of four main parts and mimics the corresponding regions of the neural structure underlying fear conditioning. The functions of these parts are implemented by assigning adaptive networks to the different parts. The learning algorithm of BELPM is based on the steepest descent (SD) and the least square estimator (LSE). In this paper, BELPM is employed to predict geomagnetic storms using the Disturbance Storm Time (Dst) index. To evaluate the performance of BELPM, the obtained results have been compared with the results of the adaptive neuro-fuzzy inference system (ANFIS).

Prediction of solar cycle 24

Accurate prediction of solar activity as one aspect of space weather phenomena is essential to decrease the damage from these activities on the ground based communication, power grids, etc. Recently, the connectionist models of the brain such as neural networks and neuro-fuzzy methods have been proposed to forecast space weather phenomena; however, they have not been able to predict solar activity accurately. That has been a motivation for the development of the connectionist model of the brain; this paper aims to apply a connectionist model of the brain to accurately forecasting solar activity, in particular, solar cycle 24. The neuro-fuzzy method has been referred to as the brain emotional learning-based recurrent fuzzy system (BELRFS). BELRFS is tested for prediction of solar cycle 24, and the obtained results are compared with well-known neuro-fuzzy methods and neural networks as well as with physical-based methods.

Neuro-fuzzy Models for Geomagnetic Storms Prediction Using the Auroral Electrojet Index

This study presents comparative results obtained from employing four different neuro-fuzzy models to predict geomagnetic storms. Two of this neuro-fuzzy models can be classified as Brain Emotional Learning Inspired Models (BELIMs) These two models are BELFIS (Brain Emotional Learning Based Fuzzy Inference System) and BELRFS (Brain Emotional Learning Recurrent Fuzzy System). The two other models are Adaptive Neuro-Fuzzy Inference System (ANFIS) and Locally Linear Model Tree (LoLiMoT) learning algorithm, two powerful neuro-fuzzy models to accurately predict a nonlinear system. These models are compared for their ability to predict geomagnetic storms using the AE index.

An Imperialist Competitive Algorithm for Interference-Aware Cluster-Heads Selection in Ad Hoc Networks

This paper presents the results of applying a new clustering algorithm in ad hoc networks. This algorithm is a centralized method and is designed on the basis of an imperialist competitive algorithm (ICA). This algorithm aims to find a minimum number of cluster-heads while satisfying two constraints, the connectivity and interference. This work is a part of an ongoing research to develop a distributed interference aware cluster-based channel allocation method. As a matter of fact, the results of the centralized method are required to provide an upper level for the performance of the distributed version. The suggested method is evaluated for several scenarios and compares the obtained results with the reported results of ant colony optimization-based methods.

A centralized channel assignment algorithm for clustered ad hoc networks

This paper presents new channel assignment algorithm for a clustered ad hoc network. The suggested method is based on a graph-theoretic model and seeks a solution for the channel assignment problem in a clustered ad hoc network. The method is based on a new meta-heuristic algorithm that is referred to as imperialist competitive algorithm (ICA). It provides a scheme for allocating the available channels to the cluster heads, maximizing spectrum efficiency and minimizing co-channel interference. The suggested method is tested for several scenarios and its performance is compared with a genetic algorithm based scheme.