Learn More
has a few major weaknesses that diminish its value as a good uninitiated to figure out why the techniques presented are source of reference for the subject matter concerned. called ‘neurofuzzy’ in the first place. The book seems unable to offer a technically coherent framework to address the adaptive neuro-fuzzy modeling and control problems. There is no(More)
This paper describes in detail a flexible approach to nonstationary time series analysis based on a Dynamic Harmonic Regression (DHR) model of the Unobserved Components (UC) type, formulated with a stochastic state space setting. The model is particularly useful for adaptive seasonal adjustment, signal extraction and interpolation over gaps, as well as(More)
Understanding spatio-temporal patterns in rainfall received beneath tropical forest is required for eco- hydrological modelling of soil-water status, river behaviour, soil erosion, nutrient loss and wet-canopy evaporation. As selective-logging of tropical forest leaves a very complex mosaic of canopy types, it is likely to add to the spatio-temporal(More)
Polycyclic aromatic hydrocarbon (PAH) air concentrations measured over the period 1992-2000 at the Canadian High Arctic station of Alert were subject to time-series analysis using dynamic harmonic regression (DHR). For most of the PAHs, the DHR model fit to the observed data was good, with DHR capable of interpolating over missing data points during periods(More)
An important aspect of flood risk management is the issuing of timely flood alerts. The spatial, as well as temporal, scale of these warnings is important. In many situations efficient risk management may be aided by the provision of local flood predictions at a high spatial resolution. Examples of such situations include issuing warnings for small groups(More)
The paper describes a software framework for implementing the main stages of the Data Based Mechanistic (DBM) modelling approach to the reduced order emulation (meta-modelling) of large dynamic system computer models, within the Matlab software environment. The framework exploits routines in the CAPTAIN Toolbox to identify and estimate transfer function(More)
Remote sensing applications are increasingly making use of sensors that generate demanding data flows (e.g. digital imaging devices). This paper presents the design of GridStix 1.5, a novel wireless sensor network platform that offers alternative high-performance and low-power 'personalities'. This two-tier approach allows GridStix 1.5 to provide power(More)