Maximum minimum distance ordering

Returns the indices of an exact maximum-minimum distance ordering. This is similar to the order_maxmin_exact function. The main difference is that it allows the user to specify a distance function.

Usage

max_min_ordering(locs, dist.func = NULL)

Arguments

locs

A matrix with one row per location and any number of columns (x, y, time, etc.).

dist.func

Any distance function with a signature of dist(query_location, locations_matrix). Defaults to Euclidean distance.

Value

A vector of indices giving the ordering. Element i of this vector is the index of the ith location.

Details

For Euclidean distance, this function will return the same results as order_maxmin_exact, but will be much slower for large data sets. order_maxmin_exact should be used instead of this function when the distance function is Euclidean.

References

• Katzfuss, M., and Guinness, J. "A general framework for Vecchia approximations of Gaussian processes", Statistical Science (2021) 36(1):124-141.

• Guiness, J. "Permutation methods for sharpening Gaussian process approximations", Technometrics (2018) 60(4):415-429.

See also

order_maxmin_exact

Examples

data(weather)
locs <- weather[,3:4]
max_min_ordering(locs)
#>   [1] 131  97   1  45  82  69 109 155  39   5  81  27 129  89  99  14  29  32
#>  [19]  88   4  52 148 125  58 111 124 113  13  48  44 130  92  84  51  79 100
#>  [37]  41  86  34 116 122  16  11  40 112  57  60 103  26  98 133 108  49   2
#>  [55] 134  95  20 128  65 120  75  54 138 142 153 105  21  56 137  10  61  93
#>  [73] 152  83  47 156   3  24 119  68 140   9  50 135  96 146  91  42 149  37
#>  [91] 110 106  19 141  62 123 151  66  80  38 127 145  53 118  23 104 136  63
#> [109] 114  94  59  55  25 126  87  30 157  18  70 101  64 139 115 107  33   7
#> [127]  71  90 144 117  17  31 121 143  78  73  22 154 102  43  36 132  67  85
#> [145]  77  35 147  74 150  46  72  15   8  76  12  28   6