scholarly journals On-Line Maximum Likelihood Prediction with Respect to General Loss Functions

1997 ◽  
Vol 55 (1) ◽  
pp. 105-118 ◽  
Author(s):  
Kenji Yamanishi
Keyword(s):  
Maximum Likelihood ◽  
Loss Functions ◽  
On Line ◽  
General Loss Functions ◽  
Maximum Likelihood Prediction

Statistically optimal event detection using small array data

1990 ◽  
Vol 80 (6B) ◽  
pp. 1934-1950 ◽  
Author(s):  
A. F. Kushnir ◽  
V. M. Lapshin ◽  
V. I. Pinsky ◽  
J. Fyen
Keyword(s):  
Maximum Likelihood ◽  
Power Spectrum ◽  
Event Detection ◽  
Underground Explosion ◽  
Seismic Signals ◽  
Array Data ◽  
On Line ◽  
Small Array ◽  
Maximum Likelihood Technique ◽  
Channel Filter

Abstract A generalization of Capon's maximum-likelihood technique for detection and estimation of seismic signals is introduced. By using a multi-dimensional autoregressive approximation of seismic array noise, we have developed a technique to use Capon's multi-channel filter for on-line processing. Such autoregressive adaptation to the curent noise matrix power spectrum is shown to yield good suppression of mutually correlated array noise processes. As an example, this technique is applied to detection of a small Semipalatinsk underground explosion recorded at the ARCESS array.

EVALUATION OF BEEF SIRES ACROSS BREEDS FOR CALVING EASE

1977 ◽  
Vol 57 (4) ◽  
pp. 635-645 ◽  
Author(s):  
L. R. SCHAEFFER ◽  
J. W. WILTON
Keyword(s):  
Maximum Likelihood ◽  
Scoring System ◽  
Variance Components ◽  
Error Variance ◽  
Best Linear Unbiased Prediction ◽  
Linear Unbiased Prediction ◽  
Calving Ease ◽  
Best Linear Unbiased ◽  
Maximum Likelihood Prediction ◽  
Unbiased Prediction

Agriculture Canada and Alberta Record of Performance calving ease records on 54,139 calves from 3,338 sires of 18 breeds were used to evaluate sires by comparisons across breeds of sire. An objective scoring system was applied to the calving ease codes to derive appropriate weights for each category rather than using percentage of unassisted births or assuming equal intervals between categories. Common sire and error variance components were assumed for all breeds of sire. Heritability of calving ease under the model used was estimated to be.10 by maximum likelihood. Prediction of sire values for calving ease scores of future calves were calculated by best linear unbiased prediction procedures. Shorthorn, Hereford, and Angus sires caused relatively few calving difficulties, while Maine-Anjou sires caused more difficulties. Age of dam and sex of calf differences were also important. The range of sire evaluations for calving ease was narrow, but the bulls in either extreme could be identified.

scholarly journals On the discrepancy principle and generalised maximum likelihood for regularisation

1995 ◽  
Vol 52 (3) ◽  
pp. 399-424 ◽  
Author(s):  
Mark A. Lukas
Keyword(s):  
Maximum Likelihood ◽  
Asymptotic Properties ◽  
Noisy Data ◽  
Loss Functions ◽  
Discrepancy Principle ◽  
Random Errors ◽  
Asymptotically Optimal ◽  
Large N ◽  
Class Of Loss Functions ◽  
Input Error

Let fnλ be the regularised solution of a general, linear operator equation, K f0 = g, from discrete, noisy data yi = g(xi ) + εi, i = 1, …, n, where εi are uncorrelated random errors with variance σ2. In this paper, we consider the two well–known methods – the discrepancy principle and generalised maximum likelihood (GML), for choosing the crucial regularisation parameter λ. We investigate the asymptotic properties as n → ∞ of the “expected” estimates λD and λM corresponding to these two methods respectively. It is shown that if f0 is sufficiently smooth, then λD is weakly asymptotically optimal (ao) with respect to the risk and an L2 norm on the output error. However, λD oversmooths for all sufficiently large n and also for all sufficiently small σ2. If f0 is not too smooth relative to the regularisation space W, then λD can also be weakly ao with respect to a whole class of loss functions involving stronger norms on the input error. For the GML method, we show that if f0 is smooth relative to W (for example f0 ∈ Wθ, 2, θ > m, if W = Wm, 2), then λM is asymptotically sub-optimal and undersmoothing with respect to all of the loss functions above.

Maximum likelihood prediction

1985 ◽  
Vol 37 (3) ◽  
pp. 507-517 ◽  
Author(s):  
Kenneth S. Kaminsky ◽  
Lennart S. Rhodin
Keyword(s):  
Maximum Likelihood ◽  
Maximum Likelihood Prediction

Maximum likelihood prediction of lake acidity based on sedimented diatoms

1990 ◽  
Vol 1 (1) ◽  
pp. 49-56 ◽  
Author(s):  
Jari Oksanen ◽  
Esa Läärä ◽  
Pertti Huttunen ◽  
Jouko Meriläinen
Keyword(s):  
Maximum Likelihood ◽  
Lake Acidity ◽  
Maximum Likelihood Prediction

A influência da qualidade do e-service bancário na satisfação com o serviço

2014 ◽  
Vol 7 (1) ◽  
pp. 74
Author(s):  
Antonio Erivando Xavier Júnior ◽  
Luciana Batista Sales ◽  
Heitor Takashi Kato ◽  
Eliane Cristine Francisco Maffezzolli
Keyword(s):  
Maximum Likelihood ◽  
On Line

O setor de serviços tem demonstrado crescente importância na economia brasileira. Especialmente no setor bancário, o investimento em serviços on-line oferece maior comodidade ao consumidor e menor custo de operação à empresa, e tem crescido muito nos últimos anos. Nesse contexto, esta pesquisa tem visa analisar se a qualidade do e-service bancário resulta em níveis mais elevados de satisfação com o serviço. Para isso, utilizou-se a escala E-SERVQUAL de Herington e Weaven (2009) e de satisfação com o serviço de Ribbink et al. (2004). Essas escalas foram compostas por 20 itens, distribuídos entre as dimensões: eficiência do site, necessidades pessoais, facilidade deutilização, organização do site e satisfação. Foram utilizados 204 questionários válidos, com a coleta realizada via online, com usuários do serviço em questão. Os resultados revelam que o modelo de mensuração da qualidade dos e-services parece adequado, com ênfase para a dimensão de eficiência do site. A hipótese da pesquisa foi testada com a utilização da Modelagem de Equações Estruturais, tendo como método de extração o Maximum Likelihood – ML. Foi observado o impacto positivo e significativo da qualidade do e-service bancário sobre a satisfação com o serviço. Esse resultado reforça a importância do investimento em serviços online que proporcionem eficiência, comodidade e agilidade dos serviços bancários para o consumidor final.

On-line Maximum Likelihood Identification of Open and Closed-Loop Processes

1979 ◽  
Vol 12 (8) ◽  
pp. 619-625
Author(s):  
A. van der Pol ◽  
H.W. Klesser ◽  
H.B. Verbruggen
Keyword(s):  
Maximum Likelihood ◽  
Closed Loop ◽  
On Line
Sign in / Sign up

Export Citation Format

Share Document