Reducing Bias without Prejudicing Sign
نویسندگان
چکیده
Jackknife and bootstrap bias corrections are based on a diierencing argument which does not necessarily respect the sign of the true parameter value. Depending on sampling variability they can over-correct, producing a nal estimator that is negative when one knows on physical grounds that it should be positive. To overcome this problem we suggest a simple, alternative bootstrap approach, based on biased-bootstrap methods. Our technique has similar properties to the standard uniform-bootstrap method in cases where the latter does not endanger sign, but it respects sign in a canonical way when the standard method disregards it.
منابع مشابه
Anomalous bias dependence of spin torque in magnetic tunnel junctions.
We predict an anomalous bias dependence of the spin transfer torque parallel to the interface, Tparallel, in magnetic tunnel junctions, which can be selectively tuned by the exchange splitting. It may exhibit a sign reversal without a corresponding sign reversal of the bias or even a quadratic bias dependence. We demonstrate that the underlying mechanism is the interplay of spin currents for th...
متن کاملMultilayer Keyboard: transition toward a new optimized layout
Reorganization of a keyboard layout based on linguistic characteristics would be an efficient way to improve input text speed. However, a new character layout imposes a learning period that often discourages users. Bi [1] aimed at easing a new layout acceptance by sacrificing the long term performance. We propose a solution based on the multilayer interface concept to achieve the same goal with...
متن کاملClustering and Co-evolution to Construct Neural Network Ensembles: an experimental study
This paper introduces an approach called Clustering and Co-evolution to Construct Neural Network Ensembles (CONE). This approach creates neural network ensembles in an innovative way, by explicitly partitioning the input space through a clustering method. The clustering method allows a reduction in the number of nodes of the neural networks that compose the ensemble, thus reducing the execution...
متن کاملAn Adaptive Self-adjusting Bandwidth Bandpass Filter without IIR Bias
In this paper we introduce a simple, computationally inxepentsive, adaptive recursive structure for enhancing bandpass signals highly corrupted by broad-band noise. This adaptive algorithm, enhancing input signals, enables us to estimate the center frequency and the bandwidth of the input signal. In addition, an important feature of the proposed structure is that the conventional bias existing ...
متن کاملClustering and co-evolution to construct neural network ensembles: An experimental study
This paper introduces an approach called Clustering and Co-evolution to Construct Neural Network Ensembles (CONE). This approach creates neural network ensembles in an innovative way, by explicitly partitioning the input space through a clustering method. The clustering method allows a reduction in the number of nodes of the neural networks that compose the ensemble, thus reducing the execution...
متن کامل