scholarly journals When Choosing the Best Subset Is Not the Best Choice

2021 ◽  
Author(s):  
Moritz Hanke ◽  
Louis Dijkstra ◽  
Ronja Foraita ◽  
Vanessa Didelez
Keyword(s):  
Variable Selection ◽  
Signal To Noise Ratio ◽  
Computation Time ◽  
Mixed Integer ◽  
Integer Optimization ◽  
Signal To Noise ◽  
Wide Range ◽  
Noise Ratio ◽  
Low Dimensional ◽  
Measure Of Performance

Abstract Background: Variable selection in linear regression settings is a much discussed problem. Best subset selection (BSS) is often considered as an intuitively appealing ‘gold standard’, with its use being restricted mainly by its N P-hard nature. Instead, alternatives such as the least absolute shrinkage and selection operator (Lasso) or the elastic net (Enet) have become methods of choice in high-dimensional settings. A recent proposal represents BSS as a mixed integer optimization problem so that much larger problems have become feasible in reasonable computation time. This has been exploited to study the prediction performance of BSS and its competitors. Here, we present an extensive simulation study assessing, instead, the variable selection performance of BSS compared to forward stepwise selection (FSS), Lasso and Enet. The analysis considers a wide range of settings that are challenging with regard to dimensionality, signal-to-noise ratio and correlations between relevant and irrelevant direct predictors. As measure of performance we used the best possible F1 score for each method so as to ensure a fair comparison irrespective of any criterion for choosing the tuning parameters.Results: Somewhat surprisingly, it was only in settings where the signal-to-noise ratio was high and the variables were (nearly) uncorrelated that BSS reliably outperformed the other methods. This was the case even in low dimensional settings where the number of observations exceeded the number of variables by a factor of ten. Further, the FSS approach performed nearly identically to BSS. Conclusion: Our results shed a new light on the usual presumption of BSS being, in principle, the best choice for variable selection. More attention needs to be payed to the data generating process when considering variable selection methods. Especially for correlated variables, convex alternatives like Enet are not only faster but also appear to be more accurate in practical settings.

Efficient Response Amplification of Electrothermally Driven Resonators Using Magnets

2021 ◽  
Author(s):  
Ghanimah Abuhaimed ◽  
Nizar Jaber ◽  
Nouha Alcheikh ◽  
Mohammad I. Younis
Keyword(s):  
Parametric Excitation ◽  
Energy Demand ◽  
Signal To Noise Ratio ◽  
Ambient Pressure ◽  
Thermal Relaxation ◽  
Primary Resonance ◽  
Signal To Noise ◽  
Electrothermal Actuation ◽  
Wide Range ◽  
Noise Ratio

Abstract Micro/Nano-electromechanical systems, MEMS/NEMS-based resonators are presently an important part of a wide range of applications. However, many of these devices suffer from the low signal-to-noise ratio and the need for a large driving force. Different principles were proposed to enhance the sensitivity and improve their signal-to-noise ratios (SNR), such as bifurcations, jumps and higher-order excitation. However, these methods require special designs and high actuation voltages, which are not always available in the standard function generators and power supplies. Also, it increases the devices’ overall cost and power requirements. Furthermore, parametric excitation is explored as an option to amplify the signal at a lower cost and energy demand. However, this type of excitation requires specific geometrical settings, in addition to very low damping conditions. Electrothermal actuation is investigated to achieve excitation of primary resonance, which can be used for parametric excitation. This type of excitation is desirable due to its simplicity, robustness and ability to create large internal forces at low voltages. However, the time response is limited by the thermal relaxation time. In this work, we demonstrate the use of electromagnetic actuation to significantly amplify the response of electrothermally actuated clamped-clamped resonators at first mode (primary) resonance. At ambient pressure, experimental data show 18 times amplification of the response amplitude compared with electrothermal actuation only. The method is based on introducing a permanent magnetic field to induce an out-of-plane Lorentz-force. The results show the great potential of this technique being used for a variety of sensing and signal processing applications, especially, where a large signal-to-noise ratio is required while using low operational voltages.

Gray Level Transforms and Lesion Detectability in Echographic Images

1988 ◽  
Vol 10 (3) ◽  
pp. 171-195 ◽  
Author(s):  
J.M. Thijssen ◽  
B.J. Oosterveld ◽  
R.F. Wagner
Keyword(s):  
Order Statistics ◽  
Signal To Noise Ratio ◽  
Ideal Observer ◽  
Sigmoid Function ◽  
Gray Level ◽  
Power Functions ◽  
Signal To Noise ◽  
Amplitude Data ◽  
Wide Range ◽  
Noise Ratio

In search of the optimal display of echographic information for the detection of focal lesions, a systematic study was performed considering a wide range of gray level transforms (i.e., lookup tables). This range comprised power functions of the echo envelope signal (1/8 ≤ n ≤ 8), power functions of the logarithmic transform and a sigmoid function. The implications of the transforms on the first order statistics (histogram, “point signal-to-noise ratio” SNRp) and on the second order statistics (autocorrelation function) could be derived both analytically, and from the analysis of simulated and experimentally obtained echograms of homogeneously scattering tissue models. These results were employed to estimate the lesion signal-to-noise ratio SNRQ, which specifies the detectability of a lesion by an ideal observer. It was found, both theoretically and practically, that the intensity display corresponds to the optimal transform (i.e., n=2) for a low contrast lesion. When the data were first logarithmically compressed, the lesion SNR appeared to increase with increasing power (1/8 ≤ n ≤ 8). A logarithmic transform followed by a sigmoid compression did not produce much improvement. These effects of gray level transforms on the SNRQ were shown to be relatively small, with the exception of powers n > 2 when applied to linear (i.e. amplitude) data. In the case of high lesion contrast, the sequence of log compression, followed by a square law produced the optimum SNRQ. This sequence is equivalent to the processing within echographic equipment, where the TV monitor has a gamma of the order of 2.

scholarly journals A non-invasive, quantitative study of broadband spectral responses in human visual cortex

2017 ◽  
Author(s):  
Eline R. Kupers ◽  
Helena X. Wang ◽  
Kaoru Amano ◽  
Kendrick N. Kay ◽  
David J. Heeger ◽  
...  
Keyword(s):  
Human Brain ◽  
Neuronal Activity ◽  
Signal To Noise Ratio ◽  
Spike Rate ◽  
Signal To Noise ◽  
Electrical Fields ◽  
Non Invasive ◽  
Rate Dependent ◽  
Wide Range ◽  
Noise Ratio

AbstractCurrently, non-invasive methods for studying the human brain do not reliably measure signals that depend on the rate of action potentials (spikes) in a neural population, independent of other responses such as hemodynamic coupling (functional magnetic resonance imaging) and subthreshold neuronal synchrony (oscillations and event-related potentials). In contrast, invasive methods - animal microelectrode recordings and human intracortical recordings (electrocorticography, or ECoG) - have recently measured broadband power elevation spanning 50-200 Hz in electrical fields generated by neuronal activity as a proxy for the locally averaged spike rates. Here, we sought to detect and quantify stimulus-related broadband responses using a non-invasive method - magnetoencephalography (MEG) - in individual subjects. Because extracranial measurements like MEG have multiple global noise sources and a relatively low signal-to-noise ratio, we developed an automated denoising technique, adapted from Kay et al, 2013 (1), that helps reveal the broadband signal of interest. Subjects viewed 12-Hz contrast-reversing patterns in the left, right, or bilateral visual field. Sensor time series were separated into an evoked component (12-Hz amplitude) and a broadband component (60–150 Hz, excluding stimulus harmonics). In all subjects, denoised broadband responses were reliably measured in sensors over occipital cortex. The spatial pattern of the broadband measure depended on the stimulus, with greater broadband power in sensors contralateral to the stimulus. Because we obtain reliable broadband estimates with relatively short experiments (~20 minutes), with a sufficient signal-to-noise-ratio to distinguish responses to different stimuli, we conclude that MEG broadband signals, denoised with our method, offer a practical, non-invasive means for characterizing spike-rate-dependent neural activity for a wide range of scientific questions about human brain function.Author SummaryNeuronal activity causes perturbations in nearby electrical fields. These perturbations can be measured non-invasively in the living human brain using electro- and magneto-encephalography (EEG and MEG). These two techniques have generally emphasized two kinds of measurements: oscillations and event-related responses, both of which reflect synchronous activity from large populations of neurons. A third type of signal, a stimulus-related increase in power spanning a wide range of frequencies (‘broadband’), is routinely measured in invasive recordings in animals and pre-surgical patients with implanted electrodes, but not with MEG and EEG. This broadband response is of great interest because unlike oscillations and event-related responses, it is correlated with neuronal spike rates. Here we report quantitative, spatially specific measurements of broadband fields in individual human subjects using MEG. These results demonstrate that a spike- rate-dependent measure of brain activity can be obtained non-invasively from the living human brain, and is suitable for investigating a wide range of questions about spiking activity in the human brain.

A new power-based method to determine the first arrival information of an acoustic emission wave

2018 ◽  
Vol 18 (5-6) ◽  
pp. 1620-1632 ◽  
Author(s):  
Avik Kumar Das ◽  
Christopher KY Leung
Keyword(s):  
Acoustic Emission ◽  
Arrival Time ◽  
Signal To Noise Ratio ◽  
Source Location ◽  
Signal To Noise ◽  
Power Characteristics ◽  
Wide Range ◽  
Noise Ratio ◽  
A New Technique ◽  
Wave Arrival

Acoustic emission is a powerful experimental structural health monitoring technique for determining the location of cracks formed in a member. Pinpointing wave arrival time is essential for accurate source location. Conventional arrival detection technique’s accuracy deteriorates rapidly in low signal to noise ratio (5–40 dB) region, thus unsuitable for source location due to this inaccuracy. A new technique to pinpoint the arrival time based on the power of the wave is proposed. We have designed an adaptive filter based on the power characteristics of acoustic emission wave. After filtration of the acoustic emission wave, sliding window is employed to accurately identify the region of wave arrival based on the change in transmitted power. The results from various experimental and numerical arrival time detection experiments consistently show that the proposed methodology is stable and accurate for a wide range of signal to noise ratio values (5–100 dB). Particularly, in signal to noise ratio region (5–40 dB), the method is significantly more accurate as compared to the other methods described in the literature. The method was then employed to study the localized damage progression in a steel fiber–reinforced beam under four-point bending. The results suggest that calculated source location using the new method is consistent with that from visual inspection of the member at failure and more accurate than the localization results from existing method.

Chinese speech intelligibility of children in noisy and reverberant environments

2017 ◽  
Vol 27 (10) ◽  
pp. 1357-1363
Author(s):  
Jianxin Peng ◽  
Shengju Wu
Keyword(s):  
Speech Intelligibility ◽  
Signal To Noise Ratio ◽  
Reverberation Time ◽  
Critical Factors ◽  
Older Children ◽  
Signal To Noise ◽  
Wide Range ◽  
Reverberant Environments ◽  
Noise Ratio ◽  
Intelligibility Score

Reverberation time and signal-to-noise ratio in classrooms are critical factors to speech intelligibility. In this study, the combined effect of reverberation time and signal-to-noise ratio on Chinese speech intelligibility of children was investigated in 28 elementary school classrooms in China. The results show that Chinese speech intelligibility scores increase with an increase of signal-to-noise ratio and the age of children, and decrease with an increase of reverberation time in classrooms. Younger children require higher signal-to-noise ratio and shorter reverberation time than older children to understand the speech. The A-weighted signal-to-noise ratio combined with a wide range of reverberation time can be used to predict speech intelligibility score and serve as a criterion for classroom design for elementary schools.

scholarly journals Influence of Analog-To-Digital Conversion and Decimation Parameters on the Signal-To-Noise Ratio in the Radar Path

2021 ◽  
Vol 24 (6) ◽  
pp. 38-50
Author(s):  
S. R. Heister ◽  
V. V. Kirichenko
Keyword(s):  
Signal To Noise Ratio ◽  
Sampling Rate ◽  
Quantization Noise ◽  
Least Significant Bit ◽  
Signal To Noise ◽  
Analog To Digital ◽  
Noise Interference ◽  
Analog To Digital Conversion ◽  
Wide Range ◽  
Noise Ratio

Introduction. The digital representation of received radar signals has provided a wide range of opportunities for their processing. However, the used hardware and software impose some limits on the number of bits and sampling rate of the signal at all conversion and processing stages. These limitations lead to a decrease in the signal-to-interference ratio due to quantization noise introduced by powerful components comprising the received signal (interfering reflections; active noise interference), as well as the attenuation of a low-power reflected signal represented by a limited number of bits. In practice, the amplitude of interfering reflections can exceed that of the signal reflected from the target by a factor of thousands.Aim. In this connection, it is essential to take into account the effect of quantization noise on the signal-tointerference ratio.Materials and methods. The article presents expressions for calculating the power and power spectral density (PSD) of quantization noise, which take into account the value of the least significant bit of an analog-to-digital converter (ADC) and the signal sampling rate. These expressions are verified by simulating 4-, 8- and 16-bit ADCs in the Mathcad environment.Results. Expressions are derived for calculating the quantization noise PSD of interfering reflections, which allows the PSD to be taken into account in the signal-to-interference ratio at the output of the processing chain. In addition, a comparison of decimation options (by discarding and averaging samples) is performed drawing on the estimates of the noise PSD and the signal-to-noise ratio.Conclusion. Recommendations regarding the ADC bit depth and sampling rate for the radar receiver are presented.

scholarly journals A Comparison of Environment Classification Among Premium Hearing Instruments

2021 ◽  
Vol 25 ◽  
pp. 233121652098096
Author(s):  
Anusha Yellamsetty ◽  
Erol J. Ozmeral ◽  
Robert A. Budinsky ◽  
David A. Eddins
Keyword(s):  
Signal Processing ◽  
Hearing Aids ◽  
Signal To Noise Ratio ◽  
Human Subjects ◽  
Signal To Noise ◽  
Environmental Classification ◽  
Hearing Instruments ◽  
Wide Range ◽  
Noise Ratio ◽  
Acoustic Environments

Hearing aids classify acoustic environments into multiple, generic classes for the purposes of guiding signal processing. Information about environmental classification is made available to the clinician for fitting, counseling, and troubleshooting purposes. The goal of this study was to better inform scientists and clinicians about the nature of that information by comparing the classification schemes among five premium hearing instruments in a wide range of acoustic scenes including those that vary in signal-to-noise ratio and overall level (dB SPL). Twenty-eight acoustic scenes representing various prototypical environments were presented to five premium devices mounted on an acoustic manikin. Classification measures were recorded from the brand-specific fitting software then recategorized to generic labels to conceal the device company, including (a) Speech in Quiet, (b) Speech in Noise, (c) Noise, and (d) Music. Twelve normal-hearing listeners also classified each scene. The results revealed a variety of similarities and differences among the five devices and the human subjects. Where some devices were highly dependent on input overall level, others were influenced markedly by signal-to-noise ratio. Differences between human and hearing aid classification were evident for several speech and music scenes. Environmental classification is the heart of the signal processing strategy for any given device, providing key input to subsequent decision-making. Comprehensive assessment of environmental classification is essential when considering the cost of signal processing errors, the potential impact for typical wearers, and the information that is available for use by clinicians. The magnitude of differences among devices is remarkable and to be noted.

scholarly journals Preparation of Palm Oil Ash Nanoparticles: Taguchi Optimization Method by Particle Size Distribution and Morphological Studies

2020 ◽  
Vol 10 (3) ◽  
pp. 985
Author(s):  
Samsul Rizal ◽  
C. K. Abdullah ◽  
N. G. Olaiya ◽  
N. A. Sri Aprilia ◽  
Ikramullah Zein ◽  
...  
Keyword(s):  
Palm Oil ◽  
Milling Time ◽  
Signal To Noise Ratio ◽  
Optimization Method ◽  
Parameters Optimization ◽  
Signal To Noise ◽  
Taguchi Optimization ◽  
Morphological Studies ◽  
Wide Range ◽  
Noise Ratio

The interest in the utilization of palm oil ash is high, mainly due to their renewable material, opportunity to enhance the properties and possibility to use in a wide range of applications. Palm oil ash is the by-product of the palm oil mill boilers and locally available in the form of micro-size particles. In this research, optimization of the milling process was designed using the Taguchi method to find the most influencing parameters for the preparation of palm oil ash (POA) nanoparticles using a ball milling technique. The experiment was applied using a L9 orthogonal array and signal to noise ratio to investigate the performance of parameters, which are milling time, milling speed, and balls size. The results from signal to noise ratio reveal that to produce POA nanoparticles in optimum parameters, the size of balls shows the highest significant effect on the production of POA nanoparticles, followed by milling time and speed. The results of the parameters optimization experiment were validated by a confirmation test of milling machine operations.

scholarly journals Obtaining reliable localizations with Time Reverse Imaging: limits to array design, velocity models and signal-to-noise ratios

2018 ◽  
Author(s):  
Claudia Werner ◽  
Erik H. Saenger
Keyword(s):  
Signal To Noise Ratio ◽  
Scale Up ◽  
Velocity Model ◽  
Source Location ◽  
Seismic Events ◽  
Complex Velocity ◽  
Signal To Noise ◽  
Velocity Models ◽  
Wide Range ◽  
Noise Ratio

Abstract. Time Reverse Imaging (TRI) is evolving into a standard technique for localizing and characterizing seismic events. In recent years, TRI has been applied to a wide range of applications from the lab scale over the field scale up to the global scale. No identification of events and their onset times is necessary when localizing events with TRI. Therefore, it is especially suited for localizing quasi-simultaneous events and events with a low signal-to-noise ratio. However, in contrast to more regularly applied localization methods, the prerequisites for applying TRI are not sufficiently known. To investigate the significance of station distributions, complex velocity models and signal-to-noise ratios for the localization quality, numerous simulations were performed using a finite difference code to propagate elastic waves through three-dimensional models. Synthetic seismograms were reversed in time and re-inserted into the model. The time-reversed wavefield backpropagates through the model and, in theory, focuses at the source location. This focusing was visualized using imaging conditions. Additionally, artificial focusing spots were removed with an illumination map specific to the setup. Successful localizations were sorted into four categories depending on their reliability. Consequently, individual simulation setups could be evaluated by their ability to produce reliable localizations. Optimal inter-station distances, minimum apertures, relations between array and source location, heterogeneities of inter-station distances and total number of stations were investigated for different source depth as well as source types. Additionally, the quality of the localization was analysed when using a complex velocity model or a low signal-to-noise ratio. Finally, an array in Southern California was investigated for its ability to localize seismic events in specific target depths while using the actual velocity model for that region. In addition, the success rate with recorded data was estimated. Knowledge about the prerequisites for using TRI enables the estimation of success rates for a given problem. Furthermore, it reduces the time needed for adjusting stations to achieve more reliable localizations and provides a foundation for designing arrays for applying TRI.

Sign in / Sign up

Export Citation Format

Share Document