Description Usage Arguments Value Note References
Replacing missing values with randomly sampled values from normal distribution, with width SD x width and down-shifted Median-Sd x shift compared to observed sample distribution. This is building upon the assumption that missing values have arisen due to low expression that can't be quantified. Therfore, shifting the median to lower expression levels will provide a proxy of this. In contrast to, locf imputation, this ensures that the variance is not reduced which would consequently impact the moderated t.test.
1 | impute_gaussian(df, width = 0.3, shift = -1.8, verbose = F)
|
df |
a data.frame with numeric columns |
width |
numeric. change the factor of the standard deviation. |
shift |
numeric. Negative values will shift the median distribution downwards. |
data.frame with missing values imputed.
No down-shifting and stdwith of 0.5 do not simualte low abudant missing values. down-shifting of 0.8 and stdwidth of 0.5 simulates low abundant missing values. down-shifting of 3.6 and stdwith of 0.5 results in (usually undesired) bi-modal distribution.
(Perseus, Tyanova et al. 2016)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.