Selection bias in gene extraction on the basis of microarray gene-expression data

By C. Ambroise, G. J. McLachlan. Proceedings of the National Academy of Sciences of the United States of America, 2002

1,449 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/9c2bec7994fe1cb65e8fa8c84ede0b3f5ab36149

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Genetics

In Molecular Biology

Citations in the corpus

A Review of Feature Selection and Feature Extraction Methods Applied on Microarray Data (2015, background): “It adds features in an initially empty set and calculates the leave-one-out cross-validation error [29].”
Cluster validation techniques for genome expression data (2003, background): “For related work on clustering based on these data sets the reader is referred to [ 16 , 17, 18].”
Robustness of Random Forest-based gene selection methods (2013, methodology): “perly validate the results. In particular, if only a single application of a machine learning algorithm is applied to an entire dataset, subtle errors with very serious consequences may be introduced [3,14]. To avoid this limitation, the work presented here employed bootstrap [9], method where each ”
Analysis of Gene Expression Data Using BRB-Array Tools (2007, positive): “This partial crossvalidation is widely used (Ntzani and Ioannidis, 2003) but has been shown to be highly biased and misleading (Ambroise and McLachlan, 2002; Simon et al. 2003b).”
Markov blanket-embedded genetic algorithm for gene selection (2007, methodology): “This so called internal cross validation [8,47] has been demonstrated to suffer from selection bias [8] or overfitting [47] problem, especially on datasets of small sample size.”
Applied Predictive Modeling (2013, methodology): “Ambroise and McLachlan (2002) also showed that when the bootstrap, 10-fold cross-validation or repeated test set resampling methods were used properly, the model results were correctly determined to be around the value of the no-information rate.”
Partially supervised feature selection with regularized linear models (2009, background): “Such a protocol includes an optimistic performance bias ( Ambroise & McLachlan, 2002 ).”
Bias in error estimation when using cross-validation for model selection (2006, background): “[2], Ambroise and McLachlan [3] and Reunanen [4], this gives a very biased estimate of the true error; not much better than the resubstitution estimate.”
Comparison of Bayesian predictive methods for model selection (2015, methodology): “1998; Barbieri and Berger 2004; Narisetty andHe 2014). In fact, often in the Bayesian variable selection literature, the selection is assumed to be performed using one of these approaches, and the studies focus on choosing priors for the model space and parameters of the individual models (see the r”
A simple and efficient algorithm for gene selection using sparse logistic regression (2003, neutral): “This estimate can be obtained using techniques of external cross-validation suggested in (Ambroise and McLachlan, 2002).”
Feature Selection with the Boruta Package (2010, background): “For example Ambroise and McLachlan (2002) have shown that, with the help of feature selection procedure, one can obtain a classiﬁer, which is using only non-informative attributes and is 100% accurate on the training set.”
Toward more realistic drug–target interaction predictions (2014, neutral): “Another source of potential bias is that simple cross-validation (CV) cannot evaluate the effect of adjusting the model parameters, and may therefore easily lead to selection bias and overoptimistic prediction results [20–22].”
Resolving confusion of tongues in statistics and machine learning: A primer for biologists and bioinformaticians (2012, methodology): “If cross-validation is used as an alternative to a test set, the performance of such a classiﬁer can be assessed by using cross-validation twice [5].”