Power series for noise attenuation in linear regression parameter estimation

Reprint of: Development and Evaluation of a Global Malnutrition Composite Score.

Development and Evaluation of a Global Malnutrition Composite Score

In this paper, we explicitly model a discriminative-ambiguous setup by two jointly learned parametric nonlinear transforms. The idea is to use one nonlinear transform for ambiguization and the other one for discrimination, and also to address a privacy-utility setup that is conditioned on ambiguization and discrimination priors, respectively, together with minimum information loss prior. The generic coupled representation is composed by linear combination using the two nonlinear transforms. The model parameters are learned by minimizing the empirical log likelihood of the model, where we propose an efficient solution using block coordinate descend alternating algorithm. The proposed model has low computational complexity and high recognition accuracy for the authorized parties while having low recognition accuracy for the unauthorized parties. We validate the potential of the proposed approach by numerical experiments.

Quality of life impact of early radiation treatment for breast cancer

Speech signals are often produced or received in the presence of noise, which is known to degrade the performance of a speech recognition system. In this paper, a perception- and PDE-based nonlinear transformation was developed to process spoken words in noisy environment. Our goal is to distinguish essential speech features and suppress noise so that the processed words are better recognized by a computer software. The nonlinear transformation was made on the spectrogram (short-term Fourier spectra) of speech signals, which reveals the signal energy distribution in time and frequency. The transformation reduces noise through time adaptation (reducing temporally slowly varying portions of spectra) and enhances spectral peaks (formants) by evolving a focusing quadratic fourth-order PDE. Short-term spectra of speech signals were initially divided into three (low, mid and high) frequency bands based on the critical bandwidth of human audition. An algorithm was developed to trace the upper and lower intensity envelopes of signal in each band. The difference between the upper and lower envelopes reflects the signal-to-noise (SNR) ratio of each band. Constant, low SNR signals in each band were adaptively decreased to reduce noise. Then evolution of the focusing PDE was used to enhance the spectral peaks, and further reduce noise interference. Numerical results on noisy spoken words indicated that the transformed spectral pattern of the spoken words was insensitive to noise for SNR ranging from 0 to 20 dB (decibel). The spectral distances between noisy words and original words decreased after the transformation. A numerical experiment was performed on 11 spoken words at SNR D 5 dB. A noisy word is recognized numerically by computing the closest L 2 spectral distance from the clean template. The experiment reached a recognition rate as high as 100%. Analyses on the properties of the transformation are provided. © 2001 Elsevier Science

Variables Associated With Nurse-Reported Quality Improvement Participation

Modern statistical studies often encounter regression models with high dimensions in which the number of features p is greater than the sample size n. Although the theory of linear models is well–established for the traditional assumption p < n, making valid statistical inference in high dimensional cases is a considerable challenge. With recent advances in technologies, the problem appears in many biological, medical, social, industrial, and economic studies. As known, the LASSO method is a popular technique for variable selection/estimation in high dimensional sparse linear models. Here, we show that the prediction performance of the LASSO method can be improved by eliminating the structured noises through a mixed–integer programming approach. As a result of our analysis, a modified variable selection/estimation scheme is proposed for a high dimensional regression model which can be considered as an alternative of the LASSO method. Some numerical experiments are made on the classical riboflavin production and some simulated data sets to shed light on the practical performance of the suggested method.

This paper presents an economic production quantity model taking into consideration of 1) investment in setup cost reduction; 2) investment in quality improvement; 3) investment in both setup cost reduction and quality improvement, to determine simultaneously optimal production run length and inspection schedules in a deteriorating production process. A simple and accurate algorithm is presented to locate the optimal production run length and then to find the optimal setup cost and optimal process quality simultaneously. A numerical experiment is carried out to compare the results of: 1) number of inspection schedules; 2) optimal production run length; 3) average cost per unit time of the proposed model to an existing model without considering the setup cost reduction and quality improvement. The results show that investment in setup cost reduction will result in reduction in primarily the optimal production run length means small lot size, while the investment in quality improvement results in number of inspections undertaken to be unity during each production run. The results show that investments in setup cost reduction and process quality improvements of a production process achieve some of the characteristics of JIT system such as small lot sizes and high quality.

Reduction in Anxiety and Depression Scores Associated with Improvement in Quality of Life in Patients with Inflammatory Bowel Disease

Blood culture contamination is a common problem in the emergency department (ED) that leads to unnecessary patient morbidity and health care costs. The study objective was to develop and evaluate the effectiveness of a quality improvement (QI) intervention for reducing blood culture contamination in an ED. The authors developed a QI intervention to reduce blood culture contamination in the ED and then evaluated its effectiveness in a prospective interrupted times series study. The QI intervention involved changing the technique of blood culture specimen collection from the traditional clean procedure to a new sterile procedure, with standardized use of sterile gloves and a new materials kit containing a 2% chlorhexidine skin antisepsis device, a sterile fenestrated drape, a sterile needle, and a procedural checklist. The intervention was implemented in a university-affiliated ED and its effect on blood culture contamination evaluated by comparing the biweekly percentages of blood cultures contaminated during a 48-week baseline period (clean technique) and 48-week intervention period (sterile technique), using segmented regression analysis with adjustment for secular trends and first-order autocorrelation. The goal was to achieve and maintain a contamination rate below 3%. During the baseline period, 321 of 7,389 (4.3%) cultures were contaminated, compared to 111 of 6,590 (1.7%) during the intervention period (p < 0.001). In the segmented regression model, the intervention was associated with an immediate 2.9% (95% confidence interval [CI] = 2.2% to 3.2%) absolute reduction in contamination. The contamination rate was maintained below 3% during each biweekly interval throughout the intervention period. A QI assessment of ED blood culture contamination led to development of a targeted intervention to convert the process of blood culture collection from a clean to a fully sterile procedure. Implementation of this intervention led to an immediate and sustained reduction of contamination in an ED with a high baseline contamination rate.

Many power management algorithms demand accurate and fine-grained runtime estimations of dynamic core power. In the absence of fine-grained power sensors, model-based estimations are needed. Such power models commonly approximate the switching activity of logic gates using performance counters while assuming a linear performance counter/power relation at a fixed frequency and voltage. It has been shown that this relation cannot be captured accurately enough with purely linear models and that well-established nonlinear modeling techniques, e.g., polynomial modeling, easily overfit the underlying performance/power relations. Although neural-network-based modeling has shown to accurately capture nonlinear relations, it has a large training and inference overhead which is too high for fine-grained models on core-level and estimation rates in the range of 1-10 kHz. We propose a methodology for nonlinear transformation of specific performance counters to increase power modeling accuracy at constant frequency and voltage with a relatively low overhead for both model generation and run-time application over a linear model. Furthermore, we use least-angle regression (LARS) to determine a ranking of the performance counter inputs for use in linear and nonlinear modeling and show that the transformed performance counters are better suited for power modeling. The generated dynamic power model consisting of a nonlinear transformation block and a linear regression block reduces relative estimation error on average by 4% and in worst-case scenarios by 7% compared to state-of-the-art fine-grained linear power models. Compared to a state-of-the-art polynomial regression model our proposed approach reduces the relative estimation error by 10% in worst-case scenarios.

This article considers a semiparametric varying-coefficient partially linear regression model. The semiparametric varying-coefficient partially linear regression model which is a generalization of the partially linear regression model and varying-coefficient regression model that allows one to explore the possibly nonlinear effect of a certain covariate on the response variable. A sieve M-estimation method is proposed and the asymptotic properties of the proposed estimators are discussed. Our main object is to estimate the nonparametric component and the unknown parameters simultaneously. It is easier to compute and the required computation burden is much less than the existing two-stage estimation method. Furthermore, the sieve M-estimation is robust in the presence of outliers if we choose appropriate ρ(·). Under some mild conditions, the estimators are shown to be strongly consistent; the convergence rate of the estimator for the unknown nonparametric component is obtained and the estimator for the unknown parameter is shown to be asymptotically normally distributed. Numerical experiments are carried out to investigate the performance of the proposed method.

To examine perceptions of facilitators and barriers to quality measurement and improvement in palliative care programs and differences by professional and leadership roles. We surveyed team members in diverse US and Canadian palliative care programs using a validated survey addressing teamwork and communication and constructs for educational support and training, leadership, infrastructure, and prioritization for quality measurement and improvement. We defined key facilitators as constructs rated ≥4 (agree) and key barriers as those ≤3 (disagree) on 1 to 5 scales. We conducted multivariable linear regressions for associations between key facilitators and barriers and (1) professional and (2) leadership roles, controlling for key program and respondent factors and clustering by program. We surveyed 103 respondents in 11 programs; 45.6% were physicians and 50% had leadership roles. Key facilitators across sites included teamwork, communication, the implementation climate (or environment), and program focus on quality improvement. Key barriers included educational support and incentives, particularly for quality measurement, and quality improvement infrastructure such as strategies, systems, and skilled staff. In multivariable analyses, perceptions did not differ by leadership role, but physicians and nurse practitioners/nurses/physician assistants rated most constructs statistically significantly more negatively than other team members, especially for quality improvement (6 of the 7 key constructs). Although participants rated quality improvement focus and environment highly, key barriers included lack of infrastructure, especially for quality measurement. Building on these facilitators and measuring and addressing these barriers might help programs enhance palliative care quality initiatives' acceptability, particularly for physicians and nurses.

This study examined the performance of the kernel model over the linear regression model for a real-life application in Nigeria. The linear regression and kernel regression model was used to assess the impact of the volume of money supply in Nigeria on industrial growth in Nigeria. The source of data for this study was the secondary source of data collection. Findings showed that there exist a weak positive coefficient of determination measure between volume of money supply and industrial growth which implies that the volume of money weakly explains the total amount of variation in industrial growth using the linear regression model while the kernel model found a strong positive coefficient of determination value which implies that the kernel model was adequate and far better than the linear model for estimating industrial growth in Nigeria. Also, it was found that volume of the money supply does not impact significantly on industrial growth in Nigeria using the linear model while it was found that volume of money impacts significantly on industrial growth using the kernel model. Further findings showed that the residual standard error value for the smoothed model is relatively more efficient than that of the linear model which was attributed to the performance of the kernel regression model.

Feature transformations are commonly used in speech recognition to account for distribution mismatches between the source and target domains also referred to as covariate shift. Linear affine or piecewise linear transformations are typically considered. In this paper, we present deep neural network DNN based nonlinear feature transformations estimated under the maximum likelihood criterion. We use the hidden Markov model HMM to model speech feature sequences and features in each HMM state assume a Gaussian mixture model GMM distribution. The network is pre-trained close to a linear transformation followed by a fine-tuning using the gradient descent algorithm. Due to the nonlinearity, the gradients and the partition functions of GMM-HMM state distributions are evaluated using the Monte Carlo MC method based on importance sampling. In addition, a deep stacked architecture is proposed to hierarchically build a DNN as a series of sub-networks with each representing a nonlinear transformation itself, which can be learned using a block-wise learning strategy. Applications of the proposed nonlinear transformations in speaker/environment adaptation and acoustic modeling in large vocabulary continuous speech recognition tasks show its superior performance over the widely-used constrained maximum likelihood linear regression CMLLR.