Practical metrics for performance evaluation of estimation algorithms are discussed. A variety of metrics useful for evaluating various aspects of the performance of an estimation algorithm is introduced and justified...
详细信息
Practical metrics for performance evaluation of estimation algorithms are discussed. A variety of metrics useful for evaluating various aspects of the performance of an estimation algorithm is introduced and justified. They can be classified in two different ways: 1) absolute error measures (without a reference), relative error measures (with a reference), or frequency counts (of some events), and 2) optimistic (i.e., how good the performance is), pessimistic (i.e., how bad the performance is), or balanced (neither optimistic nor pessimistic). Pros and cons of these metrics and the widely-used RMS error are explained. The paper advocates replacing the RMS error in many cases by a measure called average Euclidean error.
Stochastic volatility models (SVMs) represent an important framework for the analysis of financial time series data, together with ARCH-type models;but unlike the latter, the former, at least from the statistical poin...
详细信息
Stochastic volatility models (SVMs) represent an important framework for the analysis of financial time series data, together with ARCH-type models;but unlike the latter, the former, at least from the statistical point of view, cannot rely on the possibility of obtaining exact inference, in particular with regard to maximum likelihood estimates for the parameters of interest. For SVMs, usually only approximate results can be obtained, unless particularly sophisticated estimation strategies like exact non-gaussian filtering methods or simulation techniques are employed. In this paper we review SVM and present a new characterization for them, called 'generalized bilinear stochastic volatility'.
A unified statistical performance analysis using subspace perturbation expansions is applied to subspace-based algorithms for direction-of-arrival (DOA) estimation in the presence of sensor errors. In particular, the ...
详细信息
A unified statistical performance analysis using subspace perturbation expansions is applied to subspace-based algorithms for direction-of-arrival (DOA) estimation in the presence of sensor errors. In particular, the multiple signal classification (MUSIC), min-norm, state-space realization (TAM and DDA) and estimation of signal parameters via rotational invariance techniques (ESPRIT) algorithms are analyzed. This analysis assumes that only a finite amount of data is available. An analytical expression for the mean-squared error of the DOA estimates is developed for theoretical comparison in a simple and self-contained fashion. The tractable formulas provide insight into the algorithms. Simulation results verify the analysis.
Algebras over estimation algorithms in the set of regular problems with nonoverlapping classes are considered. A correctness criterion for the arbitrary degree algebraic closure of the model of estimation algorithms i...
详细信息
Algebras over estimation algorithms in the set of regular problems with nonoverlapping classes are considered. A correctness criterion for the arbitrary degree algebraic closure of the model of estimation algorithms in the classification problems of this type is proposed;this criterion can be efficiently verified. An estimate of the minimal degree of the algebraic closure that is sufficient for constructing a correct classifier in an arbitrary regular problem with nonoverlapping classes is found.
Linearizations of nonlinear functions that are based on Jacobian matrices often cannot be applied in practical applications of nonlinear estimation techniques. An alternative linearization method is presented in this ...
详细信息
Linearizations of nonlinear functions that are based on Jacobian matrices often cannot be applied in practical applications of nonlinear estimation techniques. An alternative linearization method is presented in this paper. The method assumes that covariance matrices are determined on a square root factored form. A factorization of the output covariance from a nonlinear vector function is directly determined by "perturbing" the nonlinear function with the columns of the factored input covariance, without explicitly calculating the linearization and with no differentiations involved. The output covariance is more accurate than that obtained with the ordinary Jacobian linearization method. It also has an advantage that Jacobian matrices do not have to be derived symbolically. (C) 1997 Elsevier Science Ltd.
This paper describes the procedure to reduce an estimation bias of interpolated DFT and single-model sine-fit algorithms. estimation bias reduction of more than an order of magnitude is attainable. The procedure does ...
详细信息
ISBN:
(纸本)9781467391344
This paper describes the procedure to reduce an estimation bias of interpolated DFT and single-model sine-fit algorithms. estimation bias reduction of more than an order of magnitude is attainable. The procedure does not modify the estimation algorithm itself and additional knowledge of the measurement system or additional sampled points are not required.
The purpose of this paper is to compare two estimation methods when identifying the coefficients of the Simplified Volterra Series (SVS) model, in order to linearize a class AB GaN Power Amplifier (PA) driven by a 20-...
详细信息
ISBN:
(纸本)9783030028497;9783030028480
The purpose of this paper is to compare two estimation methods when identifying the coefficients of the Simplified Volterra Series (SVS) model, in order to linearize a class AB GaN Power Amplifier (PA) driven by a 20-MHz LTE-A signal. First, a Digital Predistorter (DPD) design using the cholesky decomposition based inversion method and the Least Square QR (LSQR) algorithm is carried out, and next the performances of each method are analyzed in terms of computational complexity and suppressing distortions capability. The co-simulation test results show that the LSQR performs better than Cholesky decomposition in terms of Adjacent Channel Power Ratio (ACPR) and Normalized Mean Square Error (NMSE) by a margin of 3 dB and 4 dB, receptively.
This paper describes a simulation model of a sonar seabed mapping system based on echo sounder. The purpose of this simulation is to study the performance of linear predictors of different orders in estimating the fut...
详细信息
ISBN:
(纸本)9789077381281
This paper describes a simulation model of a sonar seabed mapping system based on echo sounder. The purpose of this simulation is to study the performance of linear predictors of different orders in estimating the future values of seabed depths. The simulated system consists of random seabed generator, echo sounder simulated signal generator and a predictor. The beam width of the echo sounder and the order of the estimator can be selected by the user. Labview and Matlab were used to build the simulation model. Results showed that a second order predictor gives the best results.
In this contribution, Genetic Programming (GP) is used to develop inferential estimation models using experimental data. GP performs symbolic optimisation, automatically determining both the structure and the complexi...
详细信息
In this contribution, Genetic Programming (GP) is used to develop inferential estimation models using experimental data. GP performs symbolic optimisation, automatically determining both the structure and the complexity of an empirical model. After a tutorial example, the usefulness of the technique is demonstrated by the development of an inferential estimation model of a plasticating extruder. A statistical analysis procedure is used as a guide in the selection of the final model structure. For the industrial case study, the inferential models obtained using the GP algorithm are compared to those obtained using a linear, finite impulse response model and a feedforward artificial neural network (FANN). For this application, the GP technique produces models with a significantly lower Root Mean Square (RMS) error than the other techniques.
A common challenge for exoskeleton control is discerning operator intent to provide seamless actuation of the device with the operator. One way to accomplish this is with joint angle estimation algorithms and multiple...
详细信息
A common challenge for exoskeleton control is discerning operator intent to provide seamless actuation of the device with the operator. One way to accomplish this is with joint angle estimation algorithms and multiple sensors on the human-machine system. However, the question remains of what can be accomplished with just one sensor. The objective of this study was to deploy a modular testing approach to test the performance of two joint angle estimation models-a kinematic extrapolation algorithm and a Random Forest machine learning algorithm-when each was informed solely with kinematic gait data from a single potentiometer on an ankle exoskeleton mock-up. This study demonstrates (i) the feasibility of implementing a modular approach to exoskeleton mock-up evaluation to promote continuity between testing configurations and (ii) that a Random Forest algorithm yielded lower realized errors of estimated joint angles and a decreased actuation time than the kinematic model when deployed on the physical device.
暂无评论