A novel algorithm to preprocess cancerous gene expression dataset for efficient gene selection

Abhishek, Abhishek; Singh, Sanjiv

doi:10.1109/i2ct.2017.8226206

Cited by 1 publication

(2 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The ordering used for both of these was the one induced by the scales of the columns in the matrices of predictors. As mentioned in the introduction, this order is often used more directly in a low variance filtering step, particularly with gene expression data (Singh et al 2017).…”

Section: Riboflavin and Prostate Datamentioning

confidence: 99%

“…For example, in a setting where one may expect measurement error to be distributed evenly over the variables, it is reasonable to suspect that variables with larger observed variance will be less corrupted by the error and hence contain more underlying signal. Perhaps as a result of this, it is also common to remove columns with the smallest variance as part of pre-processing, a step known in the machine learning community as applying a 'low variance filter' (see for example Silipo et al (2014), Singh et al (2017), Abou Elhamayed (2018), Langkun et al (2020), Saputra et al (2018), Kalambe et al (2020)). This is however a somewhat crude way of using this potential information, and may eliminate important variables that happen to have a small variation, a risk which the practice of scaling variables aims to mitigate.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

High-dimensional regression with potential prior information on variable importance

Stokell

Shah

2022

Stat Comput

View full text Add to dashboard Cite

There are a variety of settings where vague prior information may be available on the importance of predictors in high-dimensional regression settings. Examples include the ordering on the variables offered by their empirical variances (which is typically discarded through standardisation), the lag of predictors when fitting autoregressive models in time series settings, or the level of missingness of the variables. Whilst such orderings may not match the true importance of variables, we argue that there is little to be lost, and potentially much to be gained, by using them. We propose a simple scheme involving fitting a sequence of models indicated by the ordering. We show that the computational cost for fitting all models when ridge regression is used is no more than for a single fit of ridge regression, and describe a strategy for Lasso regression that makes use of previous fits to greatly speed up fitting the entire sequence of models. We propose to select a final estimator by cross-validation and provide a general result on the quality of the best performing estimator on a test set selected from among a number M of competing estimators in a high-dimensional linear regression setting. Our result requires no sparsity assumptions and shows that only a $$\log M$$ log M price is incurred compared to the unknown best estimator. We demonstrate the effectiveness of our approach when applied to missing or corrupted data, and in time series settings. An R package is available on github.

show abstract

Section: Riboflavin and Prostate Datamentioning

confidence: 99%