scholarly journals Anomaly Detection in Multichannel Data Using Sparse Representation in RADWT Frames

Mathematics ◽  
2021 ◽  
Vol 9 (11) ◽  
pp. 1288
Author(s):  
Daniela De Canditiis ◽  
Italia De Feis

We introduce a new methodology for anomaly detection (AD) in multichannel fast oscillating signals based on nonparametric penalized regression. Assuming the signals share similar shapes and characteristics, the estimation procedures are based on the use of the Rational-Dilation Wavelet Transform (RADWT), equipped with a tunable Q-factor able to provide sparse representations of functions with different oscillations persistence. Under the standard hypothesis of Gaussian additive noise, we model the signals by the RADWT and the anomalies as additive in each signal. Then we perform AD imposing a double penalty on the multiple regression model we obtained, promoting group sparsity both on the regression coefficients and on the anomalies. The first constraint preserves a common structure on the underlying signal components; the second one aims to identify the presence/absence of anomalies. Numerical experiments show the performance of the proposed method in different synthetic scenarios as well as in a real case.

Author(s):  
Jianwen Xu ◽  
Hu Yang

In this paper, the preliminary test almost unbiased ridge estimators of the regression coefficients based on the conflicting Wald (W), Likelihood ratio (LR) and Lagrangian multiplier (LM) tests in a multiple regression model with multivariate Student-t errors are introduced when it is suspected that the regression coefficients may be restricted to a subspace. The bias and quadratic risks of the proposed estimators are derived and compared. Sufficient conditions on the departure parameter ∆ and the ridge parameter k are derived for the proposed estimators to be superior to the almost unbiased ridge estimator, restricted almost unbiased ridge estimator and preliminary test estimator. Furthermore, some graphical results are provided to illustrate theoretical results.


2016 ◽  
Vol 15 (1) ◽  
Author(s):  
Antonio Boada

This paper, a practical application, proven through actual data, how the Bayesian Dynamic Linear Model Order 1 can be applied directly to the random waste from a multiple regression model Classic Static, thus creating an interesting addition is exposed for predictive statistical models. This Bayesian component generates a retro factor that feeds on waste (difference between predictions and actual historical values), adjusted according to the most recent historical information, all of them automatically and without the need to continually adjusts the Multiple Regression coefficients, generating an increase in the strength and stability of such models for prediction automated tools companies. This article provides a case of how Bayesian statistics can be an excellent complement to the techniques of classical frequentist statistics.RESUMEN Mediante este artículo, se expone una aplicación práctica, comprobada a través de datos reales, de cómo el Modelo Lineal Dinámico Bayesiano de Orden 1, puede ser aplicado directamente sobre los residuos aleatorios provenientes de un Modelo Clásico de Regresión Múltiple Estático, generando así un complemento interesante para los modelos estadísticos predictivos. Este componente bayesiano, genera un factor que se retro alimenta de los residuos (diferencia entre las predicciones y los valores históricos reales), ajustándose según la información histórica más reciente, todo ellos de forma automatizada y sin necesidad de ajustar continuamente los coeficientes de Regresión Múltiple, lo que genera un incremento en la robustez y estabilidad de dichos modelos para herramientas automatizadas de predicción en empresas. Este artículo establece un caso de cómo la estadística bayesiana puede ser un excelente complemento para las técnicas de las estadística clásica frecuentista.RESUMO Através deste artigo, uma aplicação prática, comprovada através de dados reais, como o modelo linear dinâmico Bayesian Ordem 1 pode ser aplicado diretamente sobre os resíduos aleatória de um modelo clássico de regressão múltipla estático, gerando assim um suplemento exposta interessante para os modelos estatísticos preditivos. Este componente Bayesian gera um fator que retro alimenta de resíduos (diferença entre as previsões e os valores históricos reais), ajustado de acordo com as últimas informações históricas, todas elas automaticamente e sem a necessidade de ajustar continuamente os coeficientes de regressão múltipla , gerando um aumento na força e estabilidade de tais modelos para ferramentas de previsão automatizada empresas. Este artigo fornece um exemplo de como as estatísticas Bayesian pode ser um excelente complemento para as técnicas de estatística freqüentista clássicos.


Weed Science ◽  
1987 ◽  
Vol 35 (5) ◽  
pp. 720-725 ◽  
Author(s):  
Roger Cousens ◽  
Philip Brain ◽  
John T. O'Donovan ◽  
P. Ashley O'Sullivan

A model, based on a rectangular hyperbola, has been developed to describe the relationship between population density and relative time of seedling emergence of wild oat (Avena fatuaL. # AVEFA) and yield of barley (Hordeum vulgareL.) and wheat (Triticum aestivumL.). The equation iswhere yLis percent yield loss, D is weed density, T is relative time of emergence of weed and crop, and a, b, and c are nonlinear regression coefficients. Significant differences in fitted equations were found between years. From the values of regression coefficients it was concluded that barley is a better competitor than wheat and is less affected by late-emerging wild oat. The model was tested on previously published data. It provided only a slightly better description of the data than a multiple-regression model, but avoided a number of undesirable, implausible properties inherent in the more frequently used approach. In particular, the model does not predict a loss in yield when no weeds are present or a yield increase from late-emerging weeds.


Author(s):  
Osval Antonio Montesinos López ◽  
Abelardo Montesinos López ◽  
Jose Crossa

AbstractThis chapter gives details of the linear multiple regression model including assumptions and some pros and cons, the maximum likelihood. Gradient descendent methods are described for learning the parameters under this model. Penalized linear multiple regression is derived under Ridge and Lasso penalties, which also emphasizes the estimation of the regularization parameter of importance for its successful implementation. Examples are given for both penalties (Ridge and Lasso) and but not for penalized regression multiple regression framework for illustrating the circumstances when the penalized versions should be preferred. Finally, the fundamentals of penalized and non-penalized logistic regression are provided under a gradient descendent framework. We give examples of logistic regression. Each example comes with the corresponding R codes to facilitate their quick understanding and use.


2018 ◽  
Vol 7 (2) ◽  
pp. 56
Author(s):  
M. J. Hossain ◽  
A. K. Majumder

In constructing estimation and hypothesis testing procedures, it is important that all available information such as sign of parameter is used in order to maximize power of the test. Often prior information are known about the sign of regression coefficients (parameter) under test, the best example being that variances cannot be negative. Ignoring information about the signs of regression parameters can lead to loss of power in small samples. With this problem in mind, this paper concerned with developing restricted estimation and hypothesis testing approach in the context of multivariate multiple regression model. Developing the technique of estimating constraint regression coefficients and testing restricted parameters with the aid of information theoretic distance are the main contribution of this paper. The distribution of the existing two-sided test follows central chi-square distribution whereas the test statistic of our proposed distance-based one-sided test follows weighted mixture of chi-square distribution. Monte Carlo simulation indicates that our newly proposed test performs better than existing tests.


2017 ◽  
Vol 1 (3) ◽  
pp. 68
Author(s):  
Doreen Beyo Kubochi ◽  
Dr. Makori Moronge

Purpose: The main objective of the study was to examine the influence of performance contracting on procurement performance among county governments in Kenya.Methodology: The study employed a descriptive research design, targeting procurement staff at county government headquarters. The researcher preferred this method because it allowed an in-depth study of the subject. The study population was the 13 county governments with annual budgetary allocation of Ksh 8 Billion and above in Kenya; the respondents were the procurement officers of these counties. 181 procurement officers were selected using simple random sampling and were issued with questionnaires. Data was collected using self-administered questionnaires. The data collected was analysed by use of descriptive and inferential statistics. Multiple regression model was used to show the relationship between the dependent variable and the independent variables. The data generated was keyed in and analysed by use of Statistical Package of Social Sciences (SPSS) version 21 to generate information which was presented using charts, frequencies and percentages.Results: First, in regard to key performance indicators, the regression coefficients of the study show that it has a significant influence of 0.537 on performance of county governments. Second in regard to monitoring and evaluation, the regression coefficients of the study show that it has a significant influence of 0.097 on performance of county governments. With regard to the third objective, the regression coefficients of the study show that it has a significant influence of 0.067 on performance of county governments. Lastly, in regard to the fourth objective, the regression coefficients of the study show that it has a significant influence of 0.080 on performance of county governments.Conclusion: Based on the study findings, the study concludes that performance of county governments can be improved by key performance indicators, monitoring and evaluation, balanced scorecard and governance structures.Policy recommendation: Existing literature indicates that as a future avenue of research, there is need to undertake similar research in other institutions and public sector organizations in Kenya and other countries in order to establish whether the explored practices herein can be generalized to affect performance in public entities.


2018 ◽  
Vol 18 (1) ◽  
pp. 20-32 ◽  
Author(s):  
Jong-Min Kim ◽  
Jaiwook Baik

Marketing ZFP ◽  
2019 ◽  
Vol 41 (4) ◽  
pp. 33-42
Author(s):  
Thomas Otter

Empirical research in marketing often is, at least in parts, exploratory. The goal of exploratory research, by definition, extends beyond the empirical calibration of parameters in well established models and includes the empirical assessment of different model specifications. In this context researchers often rely on the statistical information about parameters in a given model to learn about likely model structures. An example is the search for the 'true' set of covariates in a regression model based on confidence intervals of regression coefficients. The purpose of this paper is to illustrate and compare different measures of statistical information about model parameters in the context of a generalized linear model: classical confidence intervals, bootstrapped confidence intervals, and Bayesian posterior credible intervals from a model that adapts its dimensionality as a function of the information in the data. I find that inference from the adaptive Bayesian model dominates that based on classical and bootstrapped intervals in a given model.


2016 ◽  
Vol 136 (3) ◽  
pp. 363-372
Author(s):  
Takaaki Nakamura ◽  
Makoto Imamura ◽  
Masashi Tatedoko ◽  
Norio Hirai

Sign in / Sign up

Export Citation Format

Share Document