Patrick E. McSharry

A dynamical model for generating synthetic electrocardiogram signals

A dynamical model based on three coupled ordinary differential equations is introduced which is capable of generating realistic synthetic electrocardiogram (ECG) signals. The operator can specify the mean and standard deviation of the heart rate, the morphology of the PQRST cycle, and the power spectrum of the RR tachogram. In particular, both respiratory sinus arrhythmia at the high frequencies (HFs) and Mayer waves at the low frequencies (LFs) together with the LF/HF ratio are incorporated in the model. Much of the beat-to-beat variation in morphology and timing of the human ECG, including QT dispersion and R-peak amplitude modulation are shown to result. This model may be employed to assess biomedical signal processing techniques which are used to compute clinical statistics from the ECG.

Suitability of Dysphonia Measurements for Telemonitoring of Parkinson's Disease

In this paper, we present an assessment of the practical value of existing traditional and nonstandard measures for discriminating healthy people from people with Parkinsons disease (PD) by detecting dysphonia. We introduce a new measure of dysphonia, pitch period entropy (PPE), which is robust to many uncontrollable confounding effects including noisy acoustic environments and normal, healthy variations in voice frequency. We collected sustained phonations from 31 people, 23 with PD. We then selected ten highly uncorrelated measures, and an exhaustive search of all possible combinations of these measures finds four that in combination lead to overall correct classification performance of 91.4%, using a kernel support vector machine. In conclusion, we find that nonstandard methods in combination with traditional harmonics-to-noise ratios are best able to separate healthy from PD subjects. The selected nonstandard methods are robust to many uncontrollable variations in acoustic environment and individual subjects, and are thus well suited to telemonitoring applications.

Short-Term Load Forecasting Methods: An Evaluation Based on European Data

This paper uses intraday electricity demand data from ten European countries as the basis of an empirical comparison of univariate methods for prediction up to a day-ahead. A notable feature of the time series is the presence of both an in-traweek and an intraday seasonal cycle. The forecasting methods considered in the study include: ARIMA modeling, periodic AR modeling, an extension for double seasonality of Holt-Winters exponential smoothing, a recently proposed alternative exponential smoothing formulation, and a method based on the principal component analysis (PCA) of the daily demand profiles. Our results show a similar ranking of methods across the 10 load series. The results were disappointing for the new alternative exponential smoothing method and for the periodic AR model. The ARIMA and PCA methods performed well, but the method that consistently performed the best was the double seasonal Holt-Winters exponential smoothing method.

Wind Power Density Forecasting Using Ensemble Predictions and Time Series Models

Wind power is an increasingly used form of renewable energy. The uncertainty in wind generation is very large due to the inherent variability in wind speed, and this needs to be understood by operators of power systems and wind farms. To assist with the management of this risk, this paper investigates methods for predicting the probability density function of generated wind power from one to ten days ahead at five U.K. wind farm locations. These density forecasts provide a description of the expected future value and the associated uncertainty. We construct density forecasts from weather ensemble predictions, which are a relatively new type of weather forecast generated from atmospheric models. We also consider density forecasting from statistical time series models. The best results for wind power density prediction and point forecasting were produced by an approach that involves calibration and smoothing of the ensemble-based wind power density.

Novel Speech Signal Processing Algorithms for High-Accuracy Classification of Parkinson's Disease

There has been considerable recent research into the connection between Parkinsons disease (PD) and speech impairment. Recently, a wide range of speech signal processing algorithms (dysphonia measures) aiming to predict PD symptom severity using speech signals have been introduced. In this paper, we test how accurately these novel algorithms can be used to discriminate PD subjects from healthy controls. In total, we compute 132 dysphonia measures from sustained vowels. Then, we select four parsimonious subsets of these dysphonia measures using four feature selection algorithms, and map these feature subsets to a binary classification response using two statistical classifiers: random forests and support vector machines. We use an existing database consisting of 263 samples from 43 subjects, and demonstrate that these new dysphonia measures can outperform state-of-the-art results, reaching almost 99% overall classification accuracy using only ten dysphonia features. We find that some of the recently proposed dysphonia measures complement existing algorithms in maximizing the ability of the classifiers to discriminate healthy controls from PD subjects. We see these results as an important step toward noninvasive diagnostic decision support in PD.

Linear and non-linear methods for automatic seizure detection in scalp electro-encephalogram recordings.

The electro-encephalogram is a time-varying signal that measures electrical activity in the brain. A conceptually intuitive non-linear technique, multidimensional probability evolution (MDPE), is introduced. It is based on the time evolution of the probability density function within a multi-dimensional state space. A synthetic recording is employed to illustrate why MDPE is capable of detecting changes in the underlying dynamics that are invisible to linear statistics. If a nonlinear statistic cannot outperform a simple linear statistic such as variance, then there is no reason to advocate its use. Both variance and MDPE were able to detect the seizure in each of the ten scalp EEG recordings investigated. Although MDPE produced fewer false positives, there is no firm evidence to suggest that MDPE, or any other non-linear statistic considered, outperforms variance-based methods at identifying seizures.

Probabilistic forecasts of the magnitude and timing of peak electricity demand

Adequate capacity planning requires accurate forecasts of the future magnitude and timing of peak electricity demand. Electricity demand is affected by the day of the week, seasonal variations, holiday periods, feast days, and the weather. A model that provides probabilistic forecasts of both magnitude and timing for lead times of one year is presented. This model is capable of capturing the main sources of variation in demand and uses simulated weather time series, including temperature, wind speed, and luminosity, for producing probabilistic forecasts of future peak demand. Having access to such probabilistic forecasts provides a means of assessing the uncertainty in the forecasts and can lead to improved decision making and better risk management.

Nonlinear, Biophysically-Informed Speech Pathology Detection

This paper reports a simple nonlinear approach to online acoustic speech pathology detection for automatic screening purposes. Straightforward linear preprocessing followed by two nonlinear measures, based parsimoniously upon the biophysics of speech production, combined with subsequent linear classification, achieves an overall normal/pathological detection performance of 91.4%, and over 99% with rejection of 15% ambiguous cases. This compares favourably with more complex, computationally intensive methods based on a large number of linear and other measures. This demonstrates that nonlinear approaches to speech pathology detection, informed by biophysics, can be both simple and robust, and are amenable to implementation as online algorithms

Characterizing artefact in the normal human 24-hour RR time series to aid identification and artificial replication of circadian variations in human beat to beat heart rate using a simple threshold

The authors present an investigation into the incidences of ectopy and artefact as a function of time of day, heart rate and state changes for 19 normal subjects. State changes are defined to be a statistically significant change in mean or variance over a window of a few minutes. Artefact incidence is shown to be significantly correlated with state change and heart rate in normal humans, whereas ectopy exhibits no significant relationship. Artefact is therefore shown to be a source of information which can aid identification of activity or state changes and facilitate abnormality detection inpatient populations. Timing thresholds are proposed which differentiate between artefact, ectopy and sinus beats. A classification system based upon the frequency of artefact occurrences in relation to state changes is presented which correctly separates 78% of the the real (normal) and artificial RR interval time series in event 2 of the CinC Challenge 2002 (entry number 38).

Approaches for multi-step density forecasts with application to aggregated wind power

The generation of multi-step density forecasts for non-Gaussian data mostly relies on Monte Carlo simulations which are computationally intensive. Using aggregated wind power in Ireland, we study two approaches of multi-step density forecasts which can be obtained from simple iterations so that intensive computations are avoided. In the first approach, we apply a logistic transformation to normalize the data approximately and describe the transformed data using ARIMA-GARCH models so that multi-step forecasts can be iterated easily. In the second approach, we describe the forecast densities by truncated normal distributions which are governed by two parameters, namely, the conditional mean and conditional variance. We apply exponential smoothing methods to forecast the two parameters simultaneously. Since the underlying model of exponential smoothing is Gaussian, we are able to obtain multi-step forecasts of the parameters by simple iterations and thus generate forecast densities as truncated normal distributions. We generate forecasts for wind power from 15 minutes to 24 hours ahead. Results show that the first approach generates superior forecasts and slightly outperforms the second approach under various proper scores. Nevertheless, the second approach is computationally more efficient and gives more robust results under different lengths of training data. It also provides an attractive alternative approach since one is allowed to choose a particular parametric density for the forecasts, and is valuable when there are no obvious transformations to normalize the data.