Kyungduk Ko

Wavelet Methods for the Detection of Anomalies and their Application to Network Traffic Analysis

Here we develop an integrated tool for the online detection of network anomalies. We consider statistical change point detection algorithms, for both local changes in the variance and for the detection of jumps, and propose modified versions of these algorithms based on moving window techniques. We investigate performances on simulated data and on network traffic data with several superimposed attacks. All detection methods are based on wavelet packet transforms. Copyright c

Bayesian Wavelet-Based Methods for the Detection of Multiple Changes of the Long Memory Parameter

Long memory processes are widely used in many scientific fields, such as economics, physics, and engineering. Change point detection problems have received considerable attention in the literature because of their wide range of possible applications. Here we describe a wavelet-based Bayesian procedure for the estimation and location of multiple change points in the long memory parameter of Gaussian autoregressive fractionally integrated moving average models (ARFIMA(p,d,q)), with unknown autoregressive and moving average parameters. Our methodology allows the number of change points to be unknown. The reversible jump Markov chain Monte Carlo algorithm is used for posterior inference. The method also produces estimates of all model parameters. Performances are evaluated on simulated data and on the benchmark Nile river dataset

A Wavelet-Based Bayesian Approach to Regression Models with Long Memory Errors and Its Application to fMRI Data

This article considers linear regression models with long memory errors. These models have been proven useful for application in many areas, such as medical imaging, signal processing, and econometrics. Wavelets, being self-similar, have a strong connection to long memory data. Here we employ discrete wavelet transforms as whitening filters to simplify the dense variance-covariance matrix of the data. We then adopt a Bayesian approach for the estimation of the model parameters. Our inferential procedure uses exact wavelet coefficients variances and leads to accurate estimates of the model parameters. We explore performances on simulated data and present an application to an fMRI data set. In the application we produce posterior probability maps (PPMs) that aid interpretation by identifying voxels that are likely activated with a given confidence.

Prescribed Fire Effects on Activity and Movement of Cattle in Mesic Sagebrush Steppe

ABSTRACT Prescribed fire has long been used worldwide for livestock and wildlife management. Prescribed fire effects on activity budgets and movement path characteristics of grazing animals, however, remain largely unknown. We evaluated whether prescribed-fire treatments in mesic sagebrush steppe affect cattle behaviors, which could influence foraging efficiency and, ultimately, impact animal productivity.Mature, lactating beef cows grazing within two study areas located in the Owyhee Mountains of southwestern Idaho, United States were tracked with Global Positioning System collars for 2 yr before and 5 yr after fall prescribed fire. Tracking data were then classified into stationary, foraging, and traveling activity types on the basis of movement distance and velocity thresholds. After fire, cattle in burned sites tended to increase their foraging activity budgets, bout durations, bout counts, and path lengths relative to prefire and cattle in unburned sites. Fire did not affect steepness of cattle foraging paths. Prescribed fire in mesic sagebrush steppe can be used to create opportunities for cattle to improve foraging efficiency by altering their activity budgets and movement path characteristics. Any consequent improvements in foraging efficiency could, in turn, promote increased rates of weight gain, better body condition, enhanced reproductive success, and ultimately, more pounds of beef for market.

Wavelet deconvolution in a periodic setting using cross-validation

The wavelet deconvolution method WaveD using band-limited wavelets offers both theoretical and computational advantages over traditional compactly supported wavelets. The translation-invariant WaveD with a fast algorithm improves further. The twofold cross-validation method for choosing the threshold parameter and the finest resolution level in WaveD is introduced. The algorithms performance is compared with the fixed constant tuning and the default tuning in WaveD.