wNNSel: Imputatin using wNNSel method.
In wNNSel: Weighted Nearest Neighbor Imputation of Missing Values using Selected Variables

Description Usage Arguments Details Value References See Also Examples

'wNNSel' is used to impute the missing values particularly in high dimensional data. It uses a cross validation procedure for selecting the best values of the tuning parameters. It also works when the samples are smaller than the covariates.

wNNSel(x, x.initial = NULL, x.true = NULL, k, useAll = TRUE,
  x.dist = "euclidean", kernel = "gaussian", method = "2", impute.fn,
  convex = TRUE, m.values = seq(2, 8, by = 2), c.values = seq(0.1, 0.5, by
  = 0.1), lambda.values = seq(0, 0.6, by = 0.01)[-1], times.max = 5,
  testNA.prop = 0.05, withinFolds = FALSE, folds, verbose = TRUE)

`x`	a numeric data `matrix` containing missing values
`x.initial`	an optional. A complete data matrix e.g. using mean imputation of `x`. If provided, it will be used for the computation of correlations.
`x.true`	a matrix of true or complete data. If provided, `MSIE` will be returned in the results list.
`k`	an optional, the number of nearest neighbors to use for imputation.
`useAll`	`logical`. If `TRUE`, all available neighbors are used for the imputation.
`x.dist`	distance to compute. The default is `x.dist="euclidean"`, that uses the Euclidean distance. Set `x.dist` to `NULL` for Manhattan distance.
`kernel`	kernel function to be used in nearest neighbors imputation. Default kernel function is "gaussian".
`method`	convex function, performs selection of variables. If `method="1"`, linear function is used and the power function is used when `method="2"`.
`impute.fn`	the imputation function to run on the length k vector of values for a missing feature. Defaults to a weighted mean of the neighboring values, weighted by the specified `kernel`. If not specified then wNN imputation will be used by default.
`convex`	logical. If `TRUE`, selected variables are used for the computation of distance. The default is `TRUE`.
`m.values`	a `vector` of integer values, required when mehtod="2".
`c.values`	a `vector` between 0 and less than 1. It is required when mehtod="1".
`lambda.values`	a `vector`, for the tuning parameter λ
`times.max`	maximum number of repititions for the cross validation procedure.
`testNA.prop`	proportion of values to be deleted artificially for cross validation in the missing matrix `x`. Default method uses 5 percent.
`withinFolds`	`logical`. Use only if the neighbors/rows belong to particular folds/groups. Default is set to `FALSE`.
`folds`	a `list` of vectors specifying folds/groups for neighbors. lenght of list is equal to the number of folds/groups. Each element/vector of the list indicates row indices belonging to that particular group/fold.
`verbose`	logical. If `TRUE`, prints status updates

For each sample, identify missinng features. For each missing feature find the nearest neighbors which have that feature. Impute the missing value using the imputation function on the selected vector of values found from the neighbors. By default the wNNSel method automatically searches for optimal values for a given data matrix.

The default method uses x.dist="euclidean" including selected covariates. The specific distancs are computed using important covariates only. If mehtod="1", the linear function in absolute value of r is used, defined by

\frac{|r|}{1-c} - \frac{c}{1-c},

for |r|>c, and, 0 , otherwise. By default, the power function |r|^m is used when mehtod="2". For more detailed discussion, see references.

a list containing imputed data matrix, and cross validation results

`x.impute`	imputed data matrix
`MSIE`	True error. Note it is only available when x.true is provided.
`lambda.opt`	optimal parameter selected by cross validation
`m.opt`	optimal parameter selected by cross validation
`MSIE.cv`	cross validation error

Tutz, G. and Ramzan,S. (2015). Improved methods for the imputation of missing data by nearest neighbor methods. Computational Statistics and Data Analysis, Vol. 90, pp. 84-99.

Faisal, S. and Tutz, G. (2017). Missing value imputation for gene expression data by tailored nearest neighbors. Statistical Application in Genetics and Molecular Biology. Vol. 16(2), pp. 95-106.

cv.wNNSel, wNNSel.impute

 set.seed(3)
 x.true = matrix(rnorm(100),10,10)
 ## create 10% missing values in x
 x.miss = artifNA(x.true, 0.10)
 ## imputed matrix
 result <- wNNSel(x.miss)
 result$x.impute
 ## cross validation result can be accessed using
 result$cross.val

[1] "Cross validation in process..."
[1] "Cross validation complete"
             [,1]       [,2]        [,3]        [,4]       [,5]       [,6]
 [1,] -0.96193342 -0.7447816 -0.57848372  0.90062473  0.7865069  0.7268389
 [2,] -0.29252572 -1.1312186 -0.94230073  0.85177045 -0.3104631 -0.8094409
 [3,] -0.32474639 -0.7163585 -0.20372818  0.72771517  1.6988848  0.2670851
 [4,] -1.15213189  0.2526524 -1.66647484  0.02576859 -0.7945937 -1.7372637
 [5,]  0.19578283  0.1520457 -0.48445511 -0.35212962  0.3484377 -1.4114251
 [6,]  0.03012394 -0.3076564 -0.74107266  0.70551551 -2.2654011 -0.4535512
 [7,]  0.08541773 -0.9530173  1.16061578  0.61171985 -0.1622053 -1.0354913
 [8,]  1.11661021 -0.6482428  1.01206712  0.03825201 -0.2672170  1.3621429
 [9,] -1.21885742  1.2243136 -0.07207847 -0.97928377 -0.4555460  0.9174567
[10,]  1.26736872  0.1998116 -1.13678230  0.79376123 -0.8991663 -0.7851422
            [,7]       [,8]       [,9]       [,10]
 [1,]  0.5735182 -0.0313255  1.7355352 -0.85381845
 [2,]  0.9181962  0.4670973  0.7308506 -0.98999433
 [3,]  0.2562873  1.0241977  0.6886400 -0.65087774
 [4,]  0.4180793  0.1770846  1.2244061  1.05394666
 [5,]  1.1743374  0.2318261  0.7942963 -0.39087803
 [6,] -0.4808464  0.7475925  0.4119010 -0.07058639
 [7,] -0.4188297  1.2170685  0.2191506 -0.46205081
 [8,]  0.9551128  0.5300453 -0.8864638  0.54090827
 [9,]  0.5428348 -0.9880528  0.4397603  0.93163497
[10,]  0.1861974 -0.1568529 -0.8863898 -0.20927435
lambda.opt      m.opt    MSIE.cv 
 0.1800000  4.0000000  0.8685304