List of available predicates

bug/0, list/0, suggestion/0, mean/2, median/2, mode/2, percentile/3, quartile/3, iqr/2, rms/2, sum_of_squares/2, variance/2, pop_variance/2, std_dev/2, pop_std_dev/2, range/2, midrange/2, mean_absolute_deviation/2, covariance/3, correlation/3, pearson_correlation/3, spearman_correlation/3, weighted_mean/3, harmonic_mean/2, trimmed_mean/4, trimmed_variance/4, skew/2, kurtosis/2, moment/3, sum/2, prod/2, rescale/2, rescale/4, mean_normalize/2, standardize/2, entropy/2, entropy/3, min_val/2, max_val/2, min_max_val/3, rank/2, rank/3, nth_row/3, nth_column/3, swap_rows_columns/2, split_n_parts/3, occurrences/2, occurrences/3, normalize_prob/2, delete_nth/3, sample/3, sample/4, sample/5, empirical_distribution/3, seq/4, factorial/2, choose/3, search_position_sorted/3, search_position_sorted/4.

Predicates details

Mean

mean(+List:number,-Mean:float)

Mean is the mean of the list List. List can also be multidimensional (list of lists).

mean([1,2,3],M).
% Expected: M = 2.

mean([[1,3,4],[7,67]],L).
% Expected: [2.666,37].

Median

median(+List:number,-Median:number)

Median is the median of the list List. List can also be multidimensional (list of lists)

median([1,2,3],M).
% Expected: M = 2.

median([[1,5,64],[27,67]],M).
% Expected: M = [5, 47].

Mode

mode(+List:number,-Mode:list)

Mode is the mode of the list List. List can also be multidimensional (list of lists)

mode([1,2,3,1],M).
% Expected: M = 1.

mode([[1,5,64],[27,67]],M).
% Expected: M = [[1,5,64], [27,67]].

Percentile

percentile(+List:number,+K:number,-Percentile:number)

Percentile is the K-th percentile of the list List. Both List and K can be multidimensional (lists of lists).

Algorithm: arrange the data in ascending order, compute r = (p/100)* (n-1) + 1 where p is the percentile. If r is integer, then the r-th element is the desired percentile. Otherwise, it is `x_ceil(r) + (r - ceil(r)) (x_(ceil(r)+ 1) - x_ceil(r))`. The last formula is valid in both cases.

percentile([1,2,3,4,6,5,9],40,P).
% Expected: P = 3.4.

percentile([1,2,3,4,6,5],40,P).
% Expected: P = 3.0.

percentile([1,2,3,4,6,5],[10,40],P)
% Expected: P = [1.5,3.0].

percentile([[1,2,3,4,6,5],[15,25]],[10,40],P).
% Expected: P = [[1.5,3.0],[16.0,19.0]].

Quartile

quartile(+List:number,+Q:number,-Quartile:number)

Quartile is the Q quartile of the list List. Both List and Q can be multidimensional (lists of lists). Wrapper for percentile.

  Example: quartile([15,25],2,Q).
  Expected: Q = 20.

Inter quartile range

iqr(+List:number,-IQR:number)

IQR is the inter quartile range of list List, computed as the difference between 3rd - 1st quartile. List can also be multidimensional (list of lists).

iqr([1,2,3,4,6,5],I).
% Expected: I = 2.5.

RMS

rms(+List:number,-RMS:number)

RMS is the root mean square of the list List. List can also be multidimensional (list of lists). Square root of the sum of the squared data values divided by the number of values.

rms([1,5,8,3],S).
% Expected: S = 4.97493.

Sum of squares

sum_of_squares(+List:number,-SumOfSquares:number)

SumOfSquares is the sum of squares of the list List. List can also be multidimensional (list of lists). Formula: \sum_n (x - \mu)^2

sum_of_squares([1,2,3],S)
% Expected: S = 2.

Variance

variance(+List:number,-Variance:number)

Variance is the sample variance of the list List. List can also be multidimensional (list of lists). Formula: (1/(N - 1)) \sum_n (x_i - \mu)^2

variance([1,2,4,6,7,8,9],V).
% Expected: V = 9.2380952.

Population variance

pop_variance(+List:number,-Variance:number)

Variance is the population variance of the list List. List can also be multidimensional (list of lists). Formula: (1/N) \sum_n (x_i - \mu)^2

pop_variance([1,4,6,72,1],V).
% Expected: V = 765.3600.

Standard devation

std_dev(+List:numbers,-StdDev:number)

StdDev is the standard deviation of the list List (square root of the sample variance). List can also be multidimensional (list of lists).

std_dev([1,2,4,6,7,8,9],S).
Expected: S = 3.039424.

Population standard deviation

pop_std_dev(+List:numbers,-StdDev:number)

StdDev is the population standard deviation of the list List (square root of the population variance). List can also be multidimensional (list of lists).

pop_std_dev([1,2,4,6,7,8,9],S).
% Expected: S = 3.039424.

Range

range(+List:numbers,-Range:number)

Range is the difference between the biggest and the smallest element of the list List. List can also be multidimensional (list of lists).

range([1,2,4,6,7,8,9],R).
% Expected: R = 8.

Midrange

midrange(+List:numbers,-Midrange:number)

Midrange is (Max - Min) / 2 of the list List (half of the range). List can also be multidimensional (list of lists).

midrange([1,2,4,6,7,8,9],M).
% Expected: M = 4.

Mean absolute deviation

mean_absolute_deviation(+List:numbers,-MAS:number)

MAD is the sum of the absolute value of the differences between data values and the mean, divided by the sample size. Formula: MAD = 1/N \sum_i |x - \mu|. List can also be multidimensional (list of lists).

mean_absolute_deviation([1,2,4,6,7,8,9],M).
% Expected: M = 2.5306122.

Covariance

covariance(+List1:numbers,+List2:numbers,-Covariance:number)

Covariance is the covariance of the lists List1 and List2.

covariance([5,12,18,23,45],[2,8,18,20,28],C).
% Expected: C = 146.1

Correlation

correlation(+List1:numbers,-List2:numbers,-Correlation:number) pearson_correlation(+List1:numbers,-List2:numbers,-Correlation:number) spearman_correlation(+List1:numbers,-List2:numbers,-Correlation:number)

Correlation is the Pearson correlation of List1 and List2. Formula: covariance(List1,List2) / (std_dev(List1) std_dev(List2)). Other correlations are self explanatory.

correlation([5,12,18,23,45],[2,8,18,20,28],C).
% Expected: C = 0.9366.

spearman_correlation([5,12,18,23,45],[2,8,18,20,28],C).
% Expected: C = 0.999999

Weighted mean

weighted_mean(+List:numbers,+Weights:numbers,-WM:number)

WM is the weighted mean of the list List: \sum x_iw_i / \sum w_i.

weighted_mean([3,8,10,17,24,27],[2,8,10,13,18,20],WM).
% Expected: WM = 19.1972.

Harmonic mean

harmonic_mean(+List:numbers,-HM:number)

HM is the harmonic mean of list List. Formula: n / (1/x1 + 1/x2 + ... + 1/xn). List can also be multidimensional (list of lists).

harmonic_mean([1,2,3,4,5,6,7],HM).
Expected: HM = 2.69972

Trimmed mean

trimmed_mean(+List:numbers,+Lower:number,+Upper:number,-TM:number)

TM is the trimmed mean of the list List, i.e., the mean computed by considering only numbers in the range [Lower,Upper].

trimmed_mean([1,2,3,4,5,6,7],3,5,T).
% Expected: T = 4

Trimmed variance

trimmed_variance(+List:numbers,+Lower:number,+Upper:number,-TV:number)

TV is the trimmed variance of the list List, i.e, the variance computed by considering only numbers in the range [Lower,Upper].

trimmed_variance([1,2,3,4,5,6,7],3,5,V).
% Expected: V = 1.0

Moment

moment(+List:numbers,+M:integer,-Moment:number)

Moment is the M-th moment about the mean for the list List. Formula: 1/n \sum (x_i - x_mean) ^ M.

moment([1,2,3,4,5],2,MO).
Expected: MO = 2

Skew

skew(+List:numbers,-Skew:number)

Skew is the sample skewness of list List. Formula: m_3 / (m_2)^(3/2). List can also be multidimensional (list of lists).

skew([2,8,0,4,1,9,9,0],S).
% Expected: S = 0.26505541

Kurtosis

kurtosis(+List:numbers,-Kurtosis:number)

Kurtosis is the fourth central moment divided by the square of the variance. List can also be multidimensional (list of lists).

kurtosis([3,5,7,2,7],K).
% Expected: K = 1.3731508875.

Rank

rank(+List:numbers,-RankList:number)

rank(+List:numbers,+Method:atom,-RankList:number)

RankList is the rank of the list List according to method Method. If Method is not provided, by default it performs average/fractional ranking. Method must be one of the following:

average or fractional: items that compare equal receive the same rank, which is the mean of ordinal ranking values (see below)
min or competition: items that compare equal receive the same rank (there will be a gap in the ranking list)
max or modified_competition: as min, but the gap is left before, rather than after
dense: as min, but no gaps are left

ordinal: all the elements receive a different rank. If the same element appears more than one time, all the occurrences will have a different (increasing) rank.

rank([0,2,3,2],R).
% Expected: R = [1.0,2.5,4.0,2.5].

rank([0,2,3,2],average,R).
% Expected: R = [1.0,2.5,4.0,2.5].

example: rank([0,2,3,2],min,R).
% Expected: R = [1,2,4,2].

example: rank([0,2,3,2],max,R).
% Expected: R = [1,3,4,3].

example: rank([0,2,3,2],dense,R).
% Expected: R = [1,2,3,2].

example: rank([0,2,3,2],ordinal,R).
% Expected: R = [1,2,4,3].

rank([[0,2,3,2],[1,4,5]],max,R).
% Expected: R = [[1,4,5,4],[2,6,7]].

Nth-row

nth_row(+ListOfList:numbers,+Nth:integer,-NthRow:List)

NthRow is the nth row of ListOfList, counting from 1.

nth_row([[1,2],[3,4]],2,N).
% Expected: N = [3,4].

Nth-column

nth_column(+ListOfList:numbers,+Nth:integer,-NthColumn:List)

NthColumn is the nth column of ListOfList, counting from 1.

nth_column([[1,2],[3,4]],2,N).
% Expected: N = [2,4].

Swap rows columns

swap_rows_columns(+LRows:numbers,+ListOfLists:LColumns)

LColumns is LRows transposed (rows and columns swapped).

swap_rows_columns([[1,2,4],[3,6,7]],R).
Expected: R = [[1,3],[2,6],[4,7]].

Split n parts

split_n_parts(+List:numbers,+Parts:number,-PartsList:numbers)

PartsList is a list of lists obtained by splitting list List in Parts parts.

split_n_parts([1,2,4,3,7,6],2,[[1,2],[4,3],[7,6]]).
% Expected: S = [[1,2],[4,3],[7,6]].

Occurrences

occurrences(+List:number,-Occ:list)

occurrences(+Number:number,+List:numbers,-Occ:list)

Occ is the occurrences of Number in list List. If Number is not provided, Occ is a list [Value,Occurrences] for each element in list List.

occurrences([1,2,4,6,7,8,9,1],O).
% Expected: O = [[1,2],[2,1],[4,1],[6,1],[7,1],[8,1],[9,1]].
  
occurrences([1,2,4,6,7,8,9,1],1,O).
% Expected: O = 2.

Min value

min_val(+List:numbers,-Min:number)

Min is the smallest value of the list List. List can also be multidimensional (list of lists).

min_val([1,2,4,6,7,8,9,1],M).
% Expected: M = 1.

Max value

max_val(+List:numbers,-Max:number)

Max is the biggest value of the list List. List can also be multidimensional (list of lists).

max_val([1,2,4,6,7,8,9,1],M).
%  Expected: M = 9.

Min max value

min_max_val(+List:numbers,-Min:number,-Max:number)

Min and Max are the minimum and the maximum value of the list List respectively.

min_max_val([1,2,4,6,7,8,9,1],Min,Max).
% Expected: Min = 1, Max = 9.

Sum

sum(+List:numbers,-Sum:number)

Sum is the sum of the elements in the list List. List can also be multidimensional (list of lists).

sum([1,24,2,3,-1],S).
% Expected: S = 29.

Prod

prod(+List:numbers,-Prod:number)

Prod is the product of the elements in list List. List can also be multidimensional (list of lists).

prod([1,24,2,3,-1],P).
% Expected: P = -144.

Normalize probability

normalize_prob(+List:numbers,-NormalizedList:number)

NormalizedList is the list List normalized. List must contain only elements between 0 and 1 (included). List can also be multidimensional (list of lists). Formula: i / list_sum foreach i in List.

normalize_prob([0.07,0.14,0.07],L).
% Expected: L = [0.25,0.5,0.25].

Rescale

rescale(+List:numbers,-Rescaled:numbers)

rescale(+List:numbers,+Lower:number,+Upper:number,-Rescaled:number)

Rescaled is list List rescaled in the range [Lower,Upper]. Also known as min-max normalization. List can also be multidimensional (list of lists). If Lower and Upper are not provided, they are set by default to 0 and 1. Every x is rescaled as Lower + ((x - min_list)(Upper - Lower)) / (max_list - min_list)

rescale([0.07,0.14,0.07],L).
% Expected: L = [0.0,1.0,0.0]

rescale([0.07,0.14,0.07],2,3,L).
% Expected: L = [2.0,3.0,2.0]

Mean normalize

mean_normalize(+List:numbers,-Normalized:numbers)

Normalized is list List mean normalized. List can also be multidimensional (list of lists). Formula: (x - mean_list) / (Upper - Lower) foreach x in List.

mean_normalize([1,2,4],L).
% Expected: L = [-0.444, -0.111, 0.555].

Standardize

standardize(+List:numbers,-Standardized:numbers)

Standardized is list List standardized. Formula: (x - mean_list) / std_dev_list foreach x in List. List can also be multidimensional (list of lists). Population std_dev is considered (divided by n).

standardize([1,2,4],L).
% Expected: L = [-1.0690449,-0.2672612,1.336306].

Entropy

entropy(+List:numbers,-Entropy:number)

entropy(+List:numbers,+Probabilities:number,-Entropy:number)

Entropy is the entropy of the list List. Formula: is probabilities are not provided, then E = -sum(pk log(pk) else E = sum(pk log(pk / qk) Logarithm in base e (natural logarithm) is computed.

entropy([9/10,1/10],E).
% Expected: E = 0.325082.

entropy([1/2,1/2],[9/10,1/10],E).
% Expected: E = 0.5108256.

Delete nth

delete_nth(+List:numbers,+Index:numbers,-LDeleted:number)

LDeleted is List with the element at pos Index removed, counting from 1. If Index is greater than the length of the list, fails.

delete_nth([1,2,7,4],3,L).
% Expected: L = [1,2,4].

Sample

sample(+List:elements,+Size:number,-Result:list)

sample(+List:elements,+Size:number,+Replace:bool,-Result:list)

sample(+List:elements,+Size:number,+Replace:bool,+Probabilities:list,-Result:list) Takes a sample of size Size from list List. Replace can be true or false, if not provided is false. Probabilities is a list of probabilities. If Replace is false and a list of probabilities is specified, the list is normalized, after the removal of the element

sample([1,2,3,4,5],5,L).
sample([1,2,3,4,5],5,true,L).
sample([1,2,3,4,5],5,false,L).
sample([1,2,3,4,5],5,true,[0.1,0.1,0.1,0.1,0.6],L).
sample([1,2,3,4,5],5,false,[0.1,0.1,0.1,0.1,0.6],L).

Empirical distribution

empirical_distribution(+List:numbers,+X:number,-Result:list)

Result is the empirical distribution of list List at point X. List and X can also be multidimensional (lists of lists).

empirical_distribution([0,1,2,2,4,6,6,7],0,E).
% Expected: E = 0.125.

empirical_distribution([0,1,2,2,4,6,6,7],2,E).
% Expected: E = 0.5.

empirical_distribution([0,1,2,2,4,6,6,7],7,E).
% Expected: E = 1.

empirical_distribution([[0,1,2,2,4,6,6,7],[1,2,4]],[6,7,8],E).
% Expected: E = [[0.875,1,1],[1,1,1]].

Seq

seq(A:number,B:number,Seq:List).

seq(A:number,B:number,Step:number,Seq:List).

List is a list with a sequence ranging from A to B with step Step. If Step is not provided, 1 is assumed.

seq(1,10,1,S).
% Expected: S = [1,2,3,4,5,6,7,8,9,10].

Factorial

factorial(+N:int,-Factorial:int)

Factorial is N! = N*(N-1)*...2*1

factorial(10,F).
% Expected: F = 3628800.

Choose

choose(+N:int,+K:int,-C:int)

C is the binomial coefficient N,K fact(N) / (fact(N-K) fact(K))

choose(10,3,C).
% Expected: C = 120.

Search position sorted

search_position_sorted(+List:numbers,+Element:number,-Pos:integer)

search_position_sorted(+List:numbers,+Element:number,+Direction:term,-Pos:integer)

Pos is the position that the element Element would have when inserted in list List to preserve its order. 0 means the first location. If the element should be inserted in the last position, Pos = N where N is the length of List. Counting from 1. List and Element can also be multidimensional (lists of lists). Direction can be left (default) or right.

search_position_sorted([1,2,3,4,5],3,P).
% Expected result: P = 2.

search_position_sorted([1,2,3,4,5],3,right,P).
% Expected result: P = 3.