Skip to main content

Table 1 Comparisons of WeNNI, KNNimpute, and LSimpute adaptive using two different measures. MSD is the mean squared deviation calculated over all spots, whereas MSD_imputed is calculated over spots with SNR smaller than β, i.e., the spots imputed in non-weighted methods. β was chosen to yield the lowest MSD for LSimpute adaptive. WeNNI is more accurate than LSimpute and KNNimpute, even though β was tuned to optimise the performance of LSimpute.

From: Improving missing value imputation of microarray data by using spot quality weights

Data set Measure β WeNNI KNNimpute LSimpute adaptive
Breast cancer MSD 0.2 0.345 0.369 0.368
  MSD_imputed   1.59 1.81 1.75
Melanoma MSD 0.6 0.995 1.08 1.05
  MSD_imputed   3.41 3.77 3.64
Mycorrhiza MSD 0.2 0.216 0.241 0.244
  MSD_imputed   0.840 0.902 0.954