PubMed Health⌕ Search
Home/Archive/PubMed

PubMed · 14692635

A new algorithm for online structure and parameter adaptation of RBF networks.

Alex Alexandridis·Haralambos Sarimveis·George Bafas
View original source ↗Export metadata ↓

Abstract

This paper deals with the problem of online adaptation of radial basis function (RBF) neural networks. A new adaptive training method is presented, which is able to modify both the structure of the network (the number of nodes in the hidden layer) and the output weights, as the algorithm proceeds. These adaptation capabilities make the algorithm suitable for modeling dynamical time varying systems, where not only the dynamics but also the operating region changes with time. Therefore, the important issue of extrapolation is faced successfully, but at the same time the algorithm takes care of the size of the network, by deleting the hidden node centers that remain inactive for a long time. The selection of the network centers is based on a fuzzy partition of the input space, which defines a number of fuzzy subspaces. The algorithm considers the centers of the fuzzy subspaces as candidates for becoming hidden node centers and makes the selections, so that at least one center is close enough to each input example. The proposed technique is illustrated through the application to time varying dynamical systems and is compared to other adaptive training methods.

Explore related subjects

Adaptation, BiologicalAlgorithmsNeural Networks, ComputerJournal ArticleResearch Support, Non-U.S. Gov't

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Alex Alexandridis, Haralambos Sarimveis, George Bafas. 2003. A new algorithm for online structure and parameter adaptation of RBF networks.. https://doi.org/10.1016/s0893-6080(03)00052-2

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2015

Evolutionary genomics of Culex pipiens: global and local adaptations associated with climate, life-history traits and anthropogenic factors.

We present the first genome-wide study of recent evolution in Culex pipiens species complex focusing on the genomic extent, functional targets and likely causes of global and local adaptations. We resequenced pooled samples of six populations of C. pipiens and two populations of the outgroup Culex torrentium. We used principal component analysis to systematically study differential natural selection across populations and developed a phylogenetic scanning method to analyse admixture without haplotype data. We found evidence for the prominent role of geographical distribution in shaping population structure and specifying patterns of genomic selection. Multiple adaptive events, involving genes implicated with autogeny, diapause and insecticide resistance were limited to specific populations. We estimate that about 5-20% of the genes (including several histone genes) and almost half of the annotated pathways were undergoing selective sweeps in each population. The high occurrence of sweeps in non-genic regions and in chromatin remodelling genes indicated the adaptive importance of gene expression changes. We hypothesize that global adaptive processes in the C. pipiens complex are potentially associated with South to North range expansion, requiring adjustments in chromatin conformation. Strong local signature of adaptation and emergence of hybrid bridge vectors necessitate genomic assessment of populations before specifying control agents.

Adaptation, Biological↗
PubMed2007

Plasticity in animal personality traits: does prior experience alter the degree of boldness?

Theoreticians predict that animal 'personality' traits may be maladaptive if fixed throughout different contexts, so the present study aimed to test whether these traits are fixed or plastic. Rainbow trout (Onchorhyncus mykiss) were given emboldening or negative experiences in the forms of watching bold or shy individuals responding to novelty or winning or losing fights to examine whether prior experience affected boldness. Bold individuals that lost fights or watched shy demonstrators became more shy by increasing their latency to approach a novel object, whereas shy observers that watched bold demonstrators remained cautious and did not modify their responses to novelty. Shy winners became bolder and decreased their latency to approach a novel object, but shy losers also displayed this shift. In comparison, control groups showed no change in behaviour. Bold fishes given negative experiences reduced their boldness which may be an adaptive response; however, shy fishes may base their strategic decisions upon self-assessment of their relative competitive ability and increase their boldness in situations where getting to resources more quickly ensures they outcompete better competitors.

Adaptation, Biological↗
PubMed2007

Even a stopped clock tells the right time twice a day: circadian timekeeping in Drosophila.

"Even a stopped clock tells the right time twice a day, and for once I'm inclined to believe Withnail is right. We are indeed drifting into the arena of the unwell... What we need is harmony. Fresh air. Stuff like that" "Bruce Robinson (1986, ref. 1)". Although a stopped Drosophila clock probably does not tell the right time even once a day, recent findings have demonstrated that accurate circadian time-keeping is dependent on harmony between groups of clock neurons within the brain. Furthermore, when harmony between the environment and the endogenous clock is lost, as during jet lag, we definitely feel unwell. In this review, we provide an overview of the current understanding of circadian rhythms in Drosophila, focussing on recent discoveries that demonstrate how approximately 100 neurons within the Drosophila brain control the behaviour of the whole fly, and how these rhythms respond to the environment.

Adaptation, Biological↗