net.maizegenetics.analysis.numericaltransform

Class kNearestNeighbors

net.maizegenetics.analysis.numericaltransform.kNearestNeighbors

public class kNearestNeighbors

Imputation of the missing data by k-nearest neighbors.

Method Detail

impute

public static kotlin.Array[] impute(kotlin.Array[] data,
                                    int k,
                                    boolean isManhattan,
                                    boolean isCosine)

Impute all the missing values.

Parameters:

data - matrix

k - - number of nearest neighbors to be computed.

isManhattan - - if true Manhattan distance will be used.

isCosine - - if true Cosine similarity will be used.

Returns:

org.jetbrains.dokka.NodeRenderContent@3dc3d8c8: imputed data matrix

columnMean

public static double columnMean(kotlin.Array[] data,
                                int col)

Parameters:

data - a matrix

col - the column of the matrix for which the mean should be computed

Returns:

org.jetbrains.dokka.NodeRenderContent@68ba1b58: the mean of the column, ignoring missing values

cosine

public static double cosine(kotlin.Array[] data1,
                            kotlin.Array[] data2)

Computes the cosine similarity of two vectors.

Parameters:

data1 - array

data2 - array

Returns:

org.jetbrains.dokka.NodeRenderContent@536910c1: angle between the two vectors.

cosineSimRank

public static kotlin.Array[] cosineSimRank(kotlin.Array[] data,
                                           int row,
                                           int col,
                                           int k)

Rank all the rows according to their cosine similarity with the given row.

Parameters:

data - matrix

row -

col -

k -

Returns:

org.jetbrains.dokka.NodeRenderContent@69f82a6b: matrix whose rows are ordered according to their similarity with given row.

distance

public static double distance(kotlin.Array[] data1,
                              kotlin.Array[] data2,
                              boolean isManhattan)

Compute Manhattan or Eulcidean distance between two vectors.

Parameters:

data1 - array

data2 - array

isManhattan - - true for Manhattan distance choice.

Returns:

org.jetbrains.dokka.NodeRenderContent@4adeaa04: value of the distance between two vectors.

KNearestNeighbor

public static kotlin.Array[] KNearestNeighbor(kotlin.Array[] data,
                                              int row,
                                              int k,
                                              boolean isManhattan)

Rank the rows based on their distance from the given row.

Parameters:

data - matrix

row -

k -

isManhattan -

Returns:

org.jetbrains.dokka.NodeRenderContent@728608dc: matrix