BURA Collection:
http://bura.brunel.ac.uk/handle/2438/13038
2019-02-05T01:08:12ZA study of logistic classifier: uniform consistency in finite-dimensional linear spaces
http://bura.brunel.ac.uk/handle/2438/17347
Title: A study of logistic classifier: uniform consistency in finite-dimensional linear spaces
Authors: Kazakeviciute, A; Olivo, M
Abstract: Let X be a random variable taking values in a finite
dimensional linear space and Y ∈ {0, 1} its associated label. We
study the case, where conditional distribution p(x) = P(Y = 1 |
X = x) depends on x through some linear form θx. We show
that in this case, under a mild assumption on the distribution µ
of X, a maximum-likelihood estimator pˆ, as well as the induced
class of logistic classifiers, are uniformly (w.r.t. p) consistent.2016-01-01T00:00:00ZA Modified Sequential Monte Carlo Procedure for the Efficient Recursive Estimation of Extreme Quantiles
http://bura.brunel.ac.uk/handle/2438/17330
Title: A Modified Sequential Monte Carlo Procedure for the Efficient Recursive Estimation of Extreme Quantiles
Authors: Date, P; Neslihanoglu, S
Abstract: Many applications in science involve finding estimates of unobserved variables
from observed data, by combining model predictions with observations. The
Sequential Monte Carlo (SMC) is a well-established technique for estimating
the distribution of unobserved variables which are conditional on current
observations. While the SMC is very successful at estimating the first central
moments, estimating the extreme quantiles of a distribution via the current
SMC methods is computationally very expensive. The purpose of this paper
is to develop a new framework using probability distortion. We use an SMC
with distorted weights in order to make computationally efficient inferences
about tail probabilities of future interest rates using the Cox-Ingersoll-Ross
(CIR) model, as well as with an observed yield curve. We show that the
proposed method yields acceptable estimates about tail quantiles at a fraction of the computational cost of the full Monte Carlo.2019-01-01T00:00:00ZComposite quantile regression for massive datasets
http://bura.brunel.ac.uk/handle/2438/17163
Title: Composite quantile regression for massive datasets
Authors: Jiang, R; Hu, X; Yu, K; Qian, W2018-09-03T00:00:00ZMaking use of external corrosion defect assessment (ECDA) data to predict DCVG %IR drop and coating defect area 10.1002/maco.201810085
http://bura.brunel.ac.uk/handle/2438/17126
Title: Making use of external corrosion defect assessment (ECDA) data to predict DCVG %IR drop and coating defect area 10.1002/maco.201810085
Authors: Yu, K
Abstract: Buried pipelines are vulnerable to the threat of corrosion. Hence these pipelines are
coated with a protective layer (coating) to isolate the metal substrate from the
surrounding environment. With time, the coating will deteriorate which could lead to
corrosion. The condition of the coating can be investigated by the external corrosion
direct assessment (ECDA) procedure to investigate and monitor corrosion activity on
unpiggable pipelines and provides a guideline in maintaining its structural integrity.
This paper highlights the results obtained from the ECDA process which was
conducted on 250 km of buried pipelines. The results from the indirect and direct
assessment part of the ECDA were modeled using the classical quantile regression
(QR) and the Bayesian quantile regression (BQR) method to investigate the effect of
factors toward the IR drop (%IR) and the coating defect size (TCDA). It was found
that the classical method and the Bayesian approach produces similar predictions on
the regression coefficients. However, the Bayesian method has the added advantage
of the posterior distribution which considers parameter uncertainties and can be
incorporated in future ECDAs.2018-01-01T00:00:00Z