Hexagonal nearest neighbor based imputation — HnnImpute • spots

Data imputation and smoothing using hexagonal nearest neighbor.

HnnImpute(data, dist.hnn, dist.k = NULL, mu = 0, sigma = 1)

Arguments

data: A data matrix with features as rows and observations as columns.
dist.hnn: A hexagonal nearest neighbor distance matrix.
dist.k: The maximum distance used to calculate the weight. Default is NULL and all neighbor weights are calculated.
mu: The mean of Gaussian filter, default is 0.
sigma: The standard deviation of Gaussian filter, default is 1.

Value

Imputed data.

See also

Examples

{
data.use <- quakes[1:100,]
dist.use <- as.matrix(dist(data.use[,1:2]))
# transpose the data to have features in rows and observations in columns
res <- HnnImpute(t(data.use[,3:4]), dist.use)
}