Search results for "Mean Square Error"
showing 7 items of 7 documents
Probabilistic cross-validation estimators for Gaussian process regression
2018
Gaussian Processes (GPs) are state-of-the-art tools for regression. Inference of GP hyperparameters is typically done by maximizing the marginal log-likelihood (ML). If the data truly follows the GP model, using the ML approach is optimal and computationally efficient. Unfortunately very often this is not case and suboptimal results are obtained in terms of prediction error. Alternative procedures such as cross-validation (CV) schemes are often employed instead, but they usually incur in high computational costs. We propose a probabilistic version of CV (PCV) based on two different model pieces in order to reduce the dependence on a specific model choice. PCV presents the benefits from both…
Efficient linear fusion of partial estimators
2018
Abstract Many signal processing applications require performing statistical inference on large datasets, where computational and/or memory restrictions become an issue. In this big data setting, computing an exact global centralized estimator is often either unfeasible or impractical. Hence, several authors have considered distributed inference approaches, where the data are divided among multiple workers (cores, machines or a combination of both). The computations are then performed in parallel and the resulting partial estimators are finally combined to approximate the intractable global estimator. In this paper, we focus on the scenario where no communication exists among the workers, de…
Effective state estimation of stochastic systems
2003
In the present paper, for constructing the minimum risk estimators of state of stochastic systems, a new technique of invariant embedding of sample statistics in a loss function is proposed. This technique represents a simple and computationally attractive statistical method based on the constructive use of the invariance principle in mathematical statistics. Unlike the Bayesian approach, an invariant embedding technique is independent of the choice of priors. It allows one to eliminate unknown parameters from the problem and to find the best invariant estimator, which has smaller risk than any of the well‐known estimators. There exists a class of control systems where observations are not …
Non Linear Image Restoration in Spatial Domain
2011
International audience; In the present work, a novel image restoration method from noisy data samples is presented. The restoration was per-formed by using some heuristic approach utilizing data samples and smoothness criteria in spatial domain. Unlike most existing techniques, this approach does not require prior modelling of either the image or noise statistics. The proposed method works in an interactive mode to find the best compromise between the data (mean square error) and the smoothing criteria. The method has been compared with the shrinkage approach, Wiener filter and Non Local Means algorithm as well. Experimental results showed that the proposed method gives better signal to noi…
Joint LMMSE equalizer for HSDPA in full-rate space time transmit diversity schemes
2005
This contribution presents a joint linear minimum mean square error (LMMSE) equalizer for full-rate space-time transmit diversity multi-code schemes based on W-CDMA, featuring two and four transmit antennas. Specifically, high-rate, high load systems are targeted, in order to analyze the high speed downlink packet access (HSDPA) of 3GPP Release 6. We derive an equivalent transmission scheme to equalize the received sufficient statistic in a single stage. To reduce the effects of multipath and multiuser interference, and to provide spatial and frequency diversity at the receiver, an LMMSE equalizer is employed. Computer simulations confirm that the proposed schemes achieve robust performance…
Graph Topology Learning and Signal Recovery Via Bayesian Inference
2019
The estimation of a meaningful affinity graph has become a crucial task for representation of data, since the underlying structure is not readily available in many applications. In this paper, a topology inference framework, called Bayesian Topology Learning, is proposed to estimate the underlying graph topology from a given set of noisy measurements of signals. It is assumed that the graph signals are generated from Gaussian Markov Random Field processes. First, using a factor analysis model, the noisy measured data is represented in a latent space and its posterior probability density function is found. Thereafter, by utilizing the minimum mean square error estimator and the Expectation M…
A reexamination of voltage distortion for classical carrier-based vs B-Spline modulation of three-phase Voltage Sources Inverters
2015
Voltage waveform improvement has been the object of several investigations for many years and a manifold of different solutions have been proposed to reduce the harmonic content in Voltage Source Inverters (VSI) power application. In many cases this improvements have been obtained by modifying the reference voltage modulating signal. The recent introduction of cardinal B-spline functions, used as carrier signals, has given rise to a new modulation technique whose main characteristic is a lower value of the Total Harmonic Distortion (THD). After the discussion on the B-Spline modulation principle and on its computational effort, a performance comparisons is carried out by means of Total Harm…