Sequential KNN imputation method

This function estimates missing values sequentially from the gene that has least missing rate.

SeqKNN: Sequential KNN imputation method
This function estimates missing values sequentially from the gene that has
least missing rate in microarray data, using weighted mean of k nearest neighbors.
<Usage>
imputed_data=SeqKNN(data, k);

<Arguments>
data: matrix or dataframe, 1 row corresponds to 1 gene, 1 column to 1
sample,colnames and rownames can be used
k: number of nearest neighbors

<Details>
’SeqKNN’ separates the dataset into incomplete and complete set that has or has not missing values respectively. The genes in incomplete set are imputed by the order of missing rate. Missing value is filled by the weighted mean value of corresponding column of the nearest neighbor genes in complete set. Once all missing values in a gene are imputed, the imputed gene is moved into the complete set and used for the imputation of the rest of genes in incomplete set. In this process, all missing values in one gene can be imputed simultaneously from the selected neighbor genes in complete set. This reduces execution time from previously developed KNN method that selects nearest neighbors for each imputation.

<Algorithm>
Ki-Yeol Kim, Byoung-Jin Kim, Gwan-Su Yi (2004.Oct.26) "Reuse of imputed data in microarray analysis increases imputation efficiency", BMC Bioinformatics 5:160.

<Data>
Javed Khan, Jun S. Wei, Markus Ringner, Lao H. Saal, Marc Ladanyi, Frank Westermann, Frank Berthold, Manfred Schwab, Cristina R. Antonescu, Carsten Peterson, and Paul S. Meltzer (2001). Classification and diagnostic prediction of cancers using gene expression profiling and artificial neural networks. Nature Medicine, Volume 7, Number 6, June

<Display Photo>
By Mjanson srf ([1]) [CC BY-SA 3.0 (http://creativecommons.org/licenses/by-sa/3.0)], via Wikimedia Commons