Different kernels and arguments can be passed to pfocal. A set of functions is available to provide all possible values for these arguments.
pfocal_info_transform()
pfocal_info_reduce()
pfocal_info_nan_policy()
pfocal_info_mean_divisor()
pfocal_info_variance()An object of type matrix: first column shows the possible argument value,
second column is used internally and third column describe the argument.
# Retrieve info on different arguments
pfocal_info_transform()
#> [,1] [,2]
#> [1,] "MULTIPLY" "0"
#> [2,] "ADD" "1"
#> [3,] "R_EXP" "2"
#> [4,] "L_EXP" "3"
#> [,3]
#> [1,] "For data value 'd' and kernal value 'k', intermediate_value = (d * k)"
#> [2,] "For data value 'd' and kernal value 'k', intermediate_value = (d + k)"
#> [3,] "For data value 'd' and kernal value 'k', intermediate_value = (d ^ k)"
#> [4,] "For data value 'd' and kernal value 'k', intermediate_value = (k ^ d)"
pfocal_info_reduce()
#> [,1] [,2]
#> [1,] "SUM" "0"
#> [2,] "ABS_SUM" "1"
#> [3,] "PRODUCT" "2"
#> [4,] "ABS_PRODUCT" "3"
#> [5,] "MIN" "4"
#> [6,] "MAX" "5"
#> [,3]
#> [1,] "Accumulator starts at 0. For each intermediate value, in no particular order, acc = ( acc + iv )"
#> [2,] "Accumulator starts at 0. For each intermediate value, in no particular order, acc = ( acc + abs(iv) )"
#> [3,] "Accumulator starts at 1. For each intermediate value, in no particular order, acc = ( acc * iv )"
#> [4,] "Accumulator starts at 1. For each intermediate value, in no particular order, acc = ( acc * abs(iv) )"
#> [5,] "Accumulator starts at the highest possible value. For each intermediate value, in no particular order, acc = min( acc , iv )"
#> [6,] "Accumulator starts at the lowest possible value. For each intermediate value, in no particular order, acc = max( acc , iv )"
pfocal_info_nan_policy()
#> [,1] [,2]
#> [1,] "NA" "0"
#> [2,] "FALSE" "1"
#> [3,] "TRUE" "2"
#> [,3]
#> [1,] "Fastest. Assume that there will be no NAN values. Will not crash if it is given a NAN, but makes no guarantee other than that."
#> [2,] "Will propagate NANs aggressively."
#> [3,] "Will discard NAN values early."
pfocal_info_mean_divisor()
#> [,1] [,2]
#> [1,] "ONE" "0"
#> [2,] "KERNEL_SIZE" "1"
#> [3,] "KERNEL_COUNT" "2"
#> [4,] "KERNEL_SUM" "3"
#> [5,] "KERNEL_ABS_SUM" "4"
#> [6,] "KERNEL_PROD" "5"
#> [7,] "KERNEL_ABS_PROD" "6"
#> [8,] "DYNAMIC_COUNT" "7"
#> [9,] "DYNAMIC_SUM" "8"
#> [10,] "DYNAMIC_ABS_SUM" "9"
#> [11,] "DYNAMIC_PROD" "10"
#> [12,] "DYNAMIC_ABS_PROD" "11"
#> [13,] "DYNAMIC_DATA_SUM" "12"
#> [14,] "DYNAMIC_DATA_ABS_SUM" "13"
#> [15,] "DYNAMIC_DATA_PROD" "14"
#> [16,] "DYNAMIC_DATA_ABS_PROD" "15"
#> [,3]
#> [1,] "Does not divide the final value by anything"
#> [2,] "Divide the final value at each point by nrow(k)*ncol(k)"
#> [3,] "Divide the final value at each point by sum(+!is.na(k))"
#> [4,] "Divide the final value at each point by sum(k[!is.na(k)])"
#> [5,] "Divide the final value at each point by sum(abs(k[!is.na(k)]))"
#> [6,] "Divide the final value at each point by prod(k[!is.na(k)])"
#> [7,] "Divide the final value at each point by prod(abs(k[!is.na(k)]))"
#> [8,] "Divide the final value at each point by sum(!is.na( intermediate_data )), recalculated at every point"
#> [9,] "Divide the final value at each point by sum(intermediate_data[!is.na( intermediate_data )]), recalculated at every point"
#> [10,] "Divide the final value at each point by sum(abs(intermediate_data[!is.na( intermediate_data )])), recalculated at every point"
#> [11,] "Divide the final value at each point by prod(intermediate_data[!is.na( intermediate_data )]), recalculated at every point"
#> [12,] "Divide the final value at each point by prod(abs(intermediate_data[!is.na( intermediate_data )])), recalculated at every point"
#> [13,] "Divide the final value at each point by sum(local_data[!is.na( intermediate_data )]), recalculated at every point"
#> [14,] "Divide the final value at each point by sum(abs(local_data[!is.na( intermediate_data )])), recalculated at every point"
#> [15,] "Divide the final value at each point by prod(local_data[!is.na( intermediate_data )]), recalculated at every point"
#> [16,] "Divide the final value at each point by prod(abs(local_data[!is.na( intermediate_data )])), recalculated at every point"
pfocal_info_variance()
#> [,1] [,2]
#> [1,] "FALSE" "0"
#> [2,] "TRUE" "1"
#> [,3]
#> [1,] "Returns the value at each point"
#> [2,] "Returns something like the variance of the intermediate values at each point. First calculate the value as normal. Then find the 'mean' by dividing the value by the mean_divisor. Then for each intermidiate value, calculate the square of the difference and 'reduce' them using the reduce_function. (ie: sum them if SUM, multiply them if PROD, take their max if MAX etc.)"