Statistical Analysis of Sample-Size Effects in ICA
نویسندگان
چکیده
Independent component analysis (ICA) solves the blind source separation problem by evaluating higher-order statistics, e.g. by estimating fourthorder moments. While estimation errors of the kurtosis can be shown to asymptotically decay with sample size according to a square-root law, they are subject to two further effects for finite samples. Firstly, errors in the estimation of kurtosis increase with the deviation from Gaussianity. Secondly, errors in kurtosis-based ICA algorithms increase when approaching the Gaussian case. These considerations allow us to derive a strict lower bound for the sample size to achieve a given separation quality, which we study analytically for a specific family of distributions and a particular algorithm (fastICA). We further provide results from simulations that support the relevance of the analytical results.
منابع مشابه
Improving effect size estimation and statistical power with multi-echo fMRI and its impact on understanding the neural systems supporting mentalizing
Functional magnetic resonance imaging (fMRI) research is routinely criticized for being statistically underpowered due to characteristically small sample sizes and much larger sample sizes are being increasingly recommended. Additionally, various sources of artifact inherent in fMRI data can have detrimental impact on effect size estimates and statistical power. Here we show how specific remova...
متن کاملImproving Mimo Channel Equalization with Independent Component Analysis
The blind equalization of multi-input multi-output (MIMO) digital communication channels is addressed in this paper. A variety of error sources (e.g., stemming from finite sample size) can cause considerable discrepancies between practical performance and theoretical limits of multichannel linear equalizers. In this contribution, we show that a post-detection strategy based on the statistical t...
متن کاملFast Kernel Density Independent Component Analysis
We develop a super-fast kernel density estimation algorithm (FastKDE) and based on this a fast kernel independent component analysis algorithm (KDICA). FastKDE calculates the kernel density estimator exactly and its computation only requires sorting n numbers plus roughly 2n evaluations of the exponential function, where n is the sample size. KDICA converges as quickly as parametric ICA algorit...
متن کاملImproving the Performance of ICA Algorithm for fMRI Simulated Data Analysis Using Temporal and Spatial Filters in the Preprocessing Phase
Introduction: The accuracy of analyzing Functional MRI (fMRI) data is usually decreases in the presence of noise and artifact sources. A common solution in for analyzing fMRI data having high noise is to use suitable preprocessing methods with the aim of data denoising. Some effects of preprocessing methods on the parametric methods such as general linear model (GLM) have previously been evalua...
متن کاملBayesian Determination of Sample Size in Longitudinal Studies with Binary Responses Using Random Effects Models
Sample size determination is important in all statistical studies including longitudinal studies. This is usually done by considering a target power to reduce the costs of sampling. Choosing the right sample size using efficient methods, ensures that the researcher achieve goal of the study, by spending the least amount of energy, time and money. In this article, using a method based on simulat...
متن کامل