Averages/Mode: Difference between revisions
Adding MUMPS example |
m Fixed markup broken by the MUMPS entry |
||
Line 871: | Line 871: | ||
2^3 |
2^3 |
||
USER>W $$MODE^ROSETTA("") |
USER>W $$MODE^ROSETTA("") |
||
Empty Set |
Empty Set</pre> |
||
=={{header|Objective-C}}== |
=={{header|Objective-C}}== |
Revision as of 06:58, 25 October 2011
You are encouraged to solve this task according to the task description, using any language you may know.
Write a program to find the mode value of a collection. The case where the collection is empty may be ignored. Care must be taken to handle the case where the mode is non-unique.
If it is not appropriate or possible to support a general collection, use a vector (array), if possible. If it is not appropriate or possible to support an unspecified value type, use integers.
ActionScript
This implementation does not work with arbitrary collections. However, it works with arrays containing mixed data, including strings and other arrays. <lang ActionScript>function Mode(arr:Array):Array { //Create an associative array to count how many times each element occurs, //an array to contain the modes, and a variable to store how many times each mode appears. var count:Array = new Array(); var modeList:Array; var maxCount:uint=0; for (var i:String in arr) { //Record how many times an element has occurred. Note that each element in the cuont array //has to be initialized explicitly, since it is an associative array. if (count[arr[i]]==undefined) { count[arr[i]]=1; } else { count[arr[i]]++; } //If this is now the most common element, clear the list of modes, and add this element. if(count[arr[i]] > maxCount) { maxCount=count[arr[i]]; modeList = new Array(); modeList.push(arr[i]); } //If this is a mode, add it to the list. else if(count[arr[i]] == maxCount){ modeList.push(arr[i]); } } return modeList; }</lang>
Ada
mode.ads: <lang Ada>generic
type Element_Type is private; type Element_Array is array (Positive range <>) of Element_Type;
package Mode is
function Get_Mode (Set : Element_Array) return Element_Array;
end Mode;</lang>
mode.adb: <lang Ada>with Ada.Containers.Indefinite_Vectors;
package body Mode is
-- map Count to Elements package Count_Vectors is new Ada.Containers.Indefinite_Vectors (Element_Type => Element_Array, Index_Type => Positive);
procedure Add (To : in out Count_Vectors.Vector; Item : Element_Type) is use type Count_Vectors.Cursor; Position : Count_Vectors.Cursor := To.First; Found : Boolean := False; begin while not Found and then Position /= Count_Vectors.No_Element loop declare Elements : Element_Array := Count_Vectors.Element (Position); begin for I in Elements'Range loop if Elements (I) = Item then Found := True; end if; end loop; end; if not Found then Position := Count_Vectors.Next (Position); end if; end loop; if Position /= Count_Vectors.No_Element then -- element found, remove it and insert to next count declare New_Position : Count_Vectors.Cursor := Count_Vectors.Next (Position); begin -- remove from old position declare Old_Elements : Element_Array := Count_Vectors.Element (Position); New_Elements : Element_Array (1 .. Old_Elements'Length - 1); New_Index : Positive := New_Elements'First; begin for I in Old_Elements'Range loop if Old_Elements (I) /= Item then New_Elements (New_Index) := Old_Elements (I); New_Index := New_Index + 1; end if; end loop; To.Replace_Element (Position, New_Elements); end; -- new position not already there? if New_Position = Count_Vectors.No_Element then declare New_Array : Element_Array (1 .. 1) := (1 => Item); begin To.Append (New_Array); end; else -- add to new position declare Old_Elements : Element_Array := Count_Vectors.Element (New_Position); New_Elements : Element_Array (1 .. Old_Elements'Length + 1); begin New_Elements (1 .. Old_Elements'Length) := Old_Elements; New_Elements (New_Elements'Last) := Item; To.Replace_Element (New_Position, New_Elements); end; end if; end; else -- element not found, add to count 1 Position := To.First; if Position = Count_Vectors.No_Element then declare New_Array : Element_Array (1 .. 1) := (1 => Item); begin To.Append (New_Array); end; else declare Old_Elements : Element_Array := Count_Vectors.Element (Position); New_Elements : Element_Array (1 .. Old_Elements'Length + 1); begin New_Elements (1 .. Old_Elements'Length) := Old_Elements; New_Elements (New_Elements'Last) := Item; To.Replace_Element (Position, New_Elements); end; end if; end if; end Add;
function Get_Mode (Set : Element_Array) return Element_Array is Counts : Count_Vectors.Vector; begin for I in Set'Range loop Add (Counts, Set (I)); end loop; return Counts.Last_Element; end Get_Mode;
end Mode;</lang>
example use: <lang Ada>with Ada.Text_IO; with Mode; procedure Main is
type Int_Array is array (Positive range <>) of Integer; package Int_Mode is new Mode (Integer, Int_Array);
Test_1 : Int_Array := (1, 2, 3, 1, 2, 4, 2, 5, 2, 3, 3, 1, 3, 6); Result : Int_Array := Int_Mode.Get_Mode (Test_1);
begin
Ada.Text_IO.Put ("Input: "); for I in Test_1'Range loop Ada.Text_IO.Put (Integer'Image (Test_1 (I))); if I /= Test_1'Last then Ada.Text_IO.Put (","); end if; end loop; Ada.Text_IO.New_Line; Ada.Text_IO.Put ("Result:"); for I in Result'Range loop Ada.Text_IO.Put (Integer'Image (Result (I))); if I /= Result'Last then Ada.Text_IO.Put (","); end if; end loop; Ada.Text_IO.New_Line;
end Main;</lang>
output:
Input: 1, 2, 3, 1, 2, 4, 2, 5, 2, 3, 3, 1, 3, 6 Result: 2, 3
AutoHotkey
Search autohotkey.com: [1]
Source: AutoHotkey forum by Laszlo <lang autohotkey>MsgBox % Mode("1 2 3") MsgBox % Mode("1 2 0 3 0.0") MsgBox % Mode("0.1 2.2 -0.1 0.22e1 2.20 0.1")
Mode(a, d=" ") { ; the number that occurs most frequently in a list delimited by d (space)
Sort a, ND%d% Loop Parse, a, %d% If (V != A_LoopField) { If (Ct > MxCt) MxV := V, MxCt := Ct V := A_LoopField, Ct := 1 } Else Ct++ Return Ct>MxCt ? V : MxV
}</lang>
C
Using an array of doubles. If another data type is desired, the cmp_dbl
and vcount
definitions should be changed accordingly.
<lang C>#include <stdio.h>
- include <stdlib.h>
typedef struct { double v; int c; } vcount;
int cmp_dbl(const void *a, const void *b) { double x = *(double*)a - *(double*)b; return x < 0 ? -1 : x > 0; }
int vc_cmp(const void *a, const void *b) { return ((vcount*)b)->c - ((vcount*)a)->c; }
int get_mode(double* x, int len, vcount **list) { int i, j; vcount *vc;
/* sort values */ qsort(x, len, sizeof(double), cmp_dbl);
/* count occurence of each value */ for (i = 0, j = 1; i < len - 1; i++, j += (x[i] != x[i + 1]));
*list = vc = malloc(sizeof(vcount) * j); vc[0].v = x[0]; vc[0].c = 1;
/* generate list value-count pairs */ for (i = j = 0; i < len - 1; i++, vc[j].c++) if (x[i] != x[i + 1]) vc[++j].v = x[i + 1];
/* sort that by count in descending order */ qsort(vc, j + 1, sizeof(vcount), vc_cmp);
/* the number of entries with same count as the highest */ for (i = 0; i <= j && vc[i].c == vc[0].c; i++);
return i; }
int main() { double values[] = { 1, 3, 6, 6, 6, 6, 7, 7, 12, 12, 12, 12, 17 };
- define len sizeof(values)/sizeof(double)
vcount *vc;
int i, n_modes = get_mode(values, len, &vc);
printf("got %d modes:\n", n_modes); for (i = 0; i < n_modes; i++) printf("\tvalue = %g, count = %d\n", vc[i].v, vc[i].c);
free(vc); return 0; }</lang>output<lang>got 2 modes:
value = 6, count = 4 value = 12, count = 4</lang>
C++
<lang cpp>
- include <iterator>
- include <utility>
- include <algorithm>
- include <list>
- include <iostream>
// helper struct template<typename T> struct referring {
referring(T const& t): value(t) {} template<typename Iter> bool operator()(std::pair<Iter, int> const& p) { return *p.first == value; } T const& value;
};
// requires: // FwdIterator is a ForwardIterator // The value_type of FwdIterator is EqualityComparable // OutIterator is an output iterator // the value_type of FwdIterator is convertible to the value_type of OutIterator // [first, last) is a valid range // provides: // the mode is written to result template<typename FwdIterator, typename OutIterator>
void mode(FwdIterator first, FwdIterator last, OutIterator result)
{
typedef typename std::iterator_traits<FwdIterator>::value_type value_type; typedef std::list<std::pair<FwdIterator, int> > count_type; typedef typename count_type::iterator count_iterator;
// count elements count_type counts;
while (first != last) { count_iterator element = std::find_if(counts.begin(), counts.end(), referring<value_type>(*first)); if (element == counts.end()) counts.push_back(std::make_pair(first, 1)); else ++element->second; ++first; }
// find maximum int max = 0; for (count_iterator i = counts.begin(); i != counts.end(); ++i) if (i->second > max) max = i->second;
// copy corresponding elements to output sequence for (count_iterator i = counts.begin(); i != counts.end(); ++i) if (i->second == max) *result++ = *i->first;
}
// example usage int main() {
int values[] = { 1, 2, 3, 1, 2, 4, 2, 5, 2, 3, 3, 1, 3, 6 }; median(values, values + sizeof(values)/sizeof(int), std::ostream_iterator<int>(std::cout, " ")); std::cout << std::endl;
} </lang> Output:
2 3
C#
<lang csharp> using System; using System.Collections; using System.Collections.Generic; using System.Linq;
namespace Test {
class Program { static void Main(string[] args) { /* * We Use Linq To Determine The Mode */ List<int> myList = new List<int>() { 1, 1, 2, 4, 4 };
var query = from numbers in myList //select the numbers group numbers by numbers //group them together so we can get the count into groupedNumbers select new { Number = groupedNumbers.Key, Count = groupedNumbers.Count() }; //so we got a query //find the max of the occurence of the mode int max = query.Max(g => g.Count); IEnumerable<int> modes = query.Where(x => x.Count == max).Select(x => x.Number);//match the frequence and select the number foreach (var item in modes) { Console.WriteLine(item); } Console.ReadLine(); }
}
} </lang>
Clojure
<lang clojure>(defn modes [coll]
(let [distrib (frequencies coll) [value freq] [first second] ; name the key/value pairs in the distrib (map) entries sorted (sort-by (comp - freq) distrib) maxfq (freq (first sorted))] (map value (take-while #(= maxfq (freq %)) sorted))))
</lang>
Common Lisp
The following returns a list of the modes of a sequence as the primary value, and the frequency as the secondary value. E.g., (mode '(a b c d a b c a b))
produces (A B)
and 3
. hash-table-options can be used to customize the hash table, e.g., to specify the test by which elements are compared.
<lang lisp>(defun mode (sequence &rest hash-table-options)
(let ((frequencies (apply #'make-hash-table hash-table-options))) (map nil (lambda (element) (incf (gethash element frequencies 0))) sequence) (let ((modes '()) (hifreq 0 )) (maphash (lambda (element frequency) (cond ((> frequency hifreq) (setf hifreq frequency modes (list element))) ((= frequency hifreq) (push element modes)))) frequencies) (values modes hifreq))))</lang>
D
The mode function returns an array of all the mode items: <lang d>import std.stdio, std.algorithm, std.array;
T[] mode(T)(T[] items) {
int[T] aa; foreach (item; items) aa[item]++; int m = reduce!max(aa.byValue()); return array(filter!((k){ return aa[k] == m; })(aa.keys));
}
void main() {
auto data = [1, 2, 3, 1, 2, 4, 2, 5, 3, 3, 1, 3, 6]; writeln("Mode: ", mode(data));
data ~= 2; writeln("Mode: ", mode(data));
}</lang> Output:
Mode: [3] Mode: [2, 3]
E
<lang e>pragma.enable("accumulator") def mode(values) {
def counts := [].asMap().diverge() var maxCount := 0 for v in values { maxCount max= (counts[v] := counts.fetch(v, fn{0}) + 1) } return accum [].asSet() for v => ==maxCount in counts { _.with(v) }
}</lang>
<lang e>? mode([1,1,2,2,3,3,4,4,4,5,5,6,6,7,8,8,9,9,0,0,0])
- value: [4, 0].asSet()</lang>
In the line "maxCount max= (counts[v] := counts.fetch(v, fn{0}) + 1)
", max=
is an update-assignment operation like +=
. (The parentheses are unnecessary.) A more verbose version would be:
<lang e> def newCount := counts.fetch(v, fn { 0 }) + 1
counts[v] := newCount maxCount := maxCount.max(newCount)</lang>
In for loops, each key and value from the collection are pattern matched against the specified key pattern => value pattern
. In "for v => ==maxCount in counts
", the ==
is a pattern-match operator which fails unless the value examined is equal to the specified value; so this selects only the input values (keys in counts
) whose counts are equal to the maximum count.
Euphoria
<lang euphoria>include misc.e
function mode(sequence s)
sequence uniques, counts, modes integer j,max uniques = {} counts = {} for i = 1 to length(s) do j = find(s[i], uniques) if j then counts[j] += 1 else uniques = append(uniques, s[i]) counts = append(counts, 1) end if end for max = counts[1] for i = 2 to length(counts) do if counts[i] > max then max = counts[i] end if end for j = 1 modes = {} while j <= length(s) do j = find_from(max, counts, j) if j = 0 then exit end if modes = append(modes, uniques[j]) j += 1 end while return modes
end function
constant s = { 1, "blue", 2, 7.5, 5, "green", "red", 5, 2, "blue", "white" } pretty_print(1,mode(s),{3})</lang>
Output:
{ "blue", 2, 5 }
F#
The Unchecked.defaultof became available in version 1.9.4 I think. <lang fsharp>let mode (l:'a seq) =
l |> Seq.countBy (fun item -> item) // Count individual items |> Seq.fold // Find max counts (fun (cp, lst) (item, c) -> // State is (count, list of items with that count) if c > cp then (c, [item]) // New max - keep count and a list of the single item elif c = cp then (c, item :: lst) // New element with max count - prepend it to the list else (cp,lst)) // else just keep old count/list (0, [Unchecked.defaultof<'a>]) // Start with a count of 0 and a dummy item |> snd // From (count, list) we just want the second item (the list)</lang>
Example usage: <lang fsharp>> mode ["a"; "b"; "c"; "c"];; val it : string list = ["c"] > mode ["a"; "b"; "c"; "c";"a"];; val it : string list = ["c"; "a"] > mode [1;2;1;3;2;0;0];; val it : int list = [0; 2; 1]</lang>
Fortran
For the Qsort_Module see Sorting_algorithms/Quicksort#Fortran <lang fortran>program mode_test
use Qsort_Module only Qsort => sort implicit none
integer, parameter :: S = 10 integer, dimension(S) :: a1 = (/ -1, 7, 7, 2, 2, 2, -1, 7, -3, -3 /) integer, dimension(S) :: a2 = (/ 1, 1, 1, 1, 1, 0, 2, 2, 2, 2 /) integer, dimension(S) :: a3 = (/ 0, 0, -1, -1, 9, 9, 3, 3, 7, 7 /)
integer, dimension(S) :: o integer :: l, trash
print *, stat_mode(a1)
trash = stat_mode(a1, o, l) print *, o(1:l) trash = stat_mode(a2, o, l) print *, o(1:l) trash = stat_mode(a3, o, l) print *, o(1:l)
contains
! stat_mode returns the lowest (if not unique) mode ! others can hold other modes, if the mode is not unique ! if others is provided, otherslen should be provided too, and ! it says how many other modes are there. ! ok can be used to know if the return value has a meaning ! or the mode can't be found (void arrays) integer function stat_mode(a, others, otherslen, ok) integer, dimension(:), intent(in) :: a logical, optional, intent(out) :: ok integer, dimension(size(a,1)), optional, intent(out) :: others integer, optional, intent(out) :: otherslen
! ta is a copy of a, we sort ta modifying it, freq ! holds the frequencies and idx the index (for ta) so that ! the value appearing freq(i)-time is ta(idx(i)) integer, dimension(size(a, 1)) :: ta, freq, idx integer :: rs, i, tm, ml, tf
if ( present(ok) ) ok = .false.
select case ( size(a, 1) ) case (0) ! no mode... ok is false return case (1) if ( present(ok) ) ok = .true. stat_mode = a(1) return case default if ( present(ok) ) ok = .true. ta = a ! copy the array call sort(ta) ! sort it in place (cfr. sort algos on RC) freq = 1 idx = 0 rs = 1 ! rs will be the number of different values do i = 2, size(ta, 1) if ( ta(i-1) == ta(i) ) then freq(rs) = freq(rs) + 1 else idx(rs) = i-1 rs = rs + 1 end if end do idx(rs) = i-1 ml = maxloc(freq(1:rs), 1) ! index of the max value of freq tf = freq(ml) ! the max frequency tm = ta(idx(ml)) ! the value with that freq
! if we want all the possible modes, we provide others if ( present(others) ) then i = 1 others(1) = tm do freq(ml) = 0 ml = maxloc(freq(1:rs), 1) if ( tf == freq(ml) ) then ! the same freq i = i + 1 ! as the max one others(i) = ta(idx(ml)) else exit end if end do if ( present(otherslen) ) then otherslen = i end if
end if stat_mode = tm end select
end function stat_mode
end program mode_test</lang>
Haskell
<lang haskell>import Prelude (foldr, maximum, (==), (+)) import Data.Map (insertWith', empty, filter, elems, keys)
mode :: (Ord a) => [a] -> [a] mode xs = keys (filter (== maximum (elems counts)) counts)
where counts = foldr (\x -> insertWith' (+) x 1) empty xs</lang>
counts is a map from each value found in xs to the number of occurrences (foldr traverses the list, insertWith' increments the count). This map is then filtered to only those entries whose count is the maximum count, and their keys (the values from the input list) are returned.
> mode [1,2,3,3,2,1,1] [1] > mode [1,2,3,3,2,1] [1,2,3]
Alternately: <lang haskell>import Data.List (group, sort)
mode :: (Ord a) => [a] -> [a] mode xs = map fst $ filter ((==best).snd) counts
where counts = map (\l -> (head l, length l)) . group . sort $ xs best = maximum (map snd counts)</lang>
Another version that does not require an orderable type: <lang haskell>import Data.List (partition)
mode :: (Eq a) => [a] -> [a] mode = snd . modesWithCount
where modesWithCount :: (Eq a) => [a] -> (Int, [a]) modesWithCount [] = (0,[]) modesWithCount l@(x:_) | length xs > best = (length xs, [x]) | length xs < best = (best, modes) | otherwise = (best, x:modes) where (xs, notxs) = partition (== x) l (best, modes) = modesWithCount notxs</lang>
GAP
<lang gap>mode := function(v)
local c, m; c := Collected(SortedList(v)); m := Maximum(List(c, x -> x[2])); return List(Filtered(c, x -> x[2] = m), y -> y[1]);
end;
mode([ 7, 5, 6, 1, 5, 5, 7, 12, 17, 6, 6, 5, 12, 3, 6 ]);
- [ 5, 6 ]</lang>
Go
<lang go>package main
import (
"fmt"
)
func main() {
fmt.Println(mode([]int{2, 7, 1, 8, 2})) fmt.Println(mode([]int{2, 7, 1, 8, 2, 8}))
}
func mode(a []int) []int {
m := make(map[int]int) for _, v := range a { m[v] = m[v] + 1 } var mode []int var n int for k, v := range m { switch { case v < n: case v > n: n = v mode = append(mode[:0], k) default: mode = append(mode, k) } } return mode
}</lang>
Icon and Unicon
The mode procedure generates all n mode values if the collection is n-modal.
<lang icon>procedure main(args)
every write(!mode(args))
end
procedure mode(A)
hist := table(0) every hist[!A] +:= 1 hist := sort(hist, 2) modeCnt := hist[*hist][2] every modeP := hist[*hist to 1 by -1] do { if modeCnt = modeP[2] then suspend modeP[1] else fail }
end</lang>
Sample outputs:
->am 3 1 4 1 5 9 7 6 1 ->am 3 1 4 1 5 9 7 6 3 3 1 ->
J
<lang j>mode=: ~. #~ ( = >./ )@( #/.~ )</lang>
Example:
<lang> mode 1 1 2 2 3 3 4 4 4 5 5 6 6 7 8 8 9 9 0 0 0 4 0</lang>
Java
<lang java>import java.util.*;
public class Mode {
public static <T> List<T> mode(List<? extends T> coll) { Map<T, Integer> seen = new HashMap<T, Integer>(); int max = 0; List<T> maxElems = new ArrayList<T>(); for (T value : coll) { if (seen.containsKey(value)) seen.put(value, seen.get(value) + 1); else seen.put(value, 1); if (seen.get(value) > max) { max = seen.get(value); maxElems.clear(); maxElems.add(value); } else if (seen.get(value) == max) { maxElems.add(value); } } return maxElems; }
public static void main(String[] args) { System.out.println(mode(Arrays.asList(1, 3, 6, 6, 6, 6, 7, 7, 12, 12, 17))); // prints [6] System.out.println(mode(Arrays.asList(1, 1, 2, 4, 4))); // prints [1, 4] }
}</lang>
JavaScript
<lang javascript>function mode(ary) {
var counter = {}; var mode = []; var max = 0; for (var i in ary) { if (!(ary[i] in counter)) counter[ary[i]] = 0; counter[ary[i]]++;
if (counter[ary[i]] == max) mode.push(ary[i]); else if (counter[ary[i]] > max) { max = counter[ary[i]]; mode = [ary[i]]; } } return mode;
}
mode([1, 3, 6, 6, 6, 6, 7, 7, 12, 12, 17]); // [6] mode([1, 2, 4, 4, 1]); // [1,4]</lang>
K
<lang k> mode: {(?x)@&n=|/n:#:'=x}
mode 1 1 1 1 2 2 2 3 3 3 3 4 4 3 2 4 4 4
3 4</lang>
Lua
<lang lua>function mode (numlist)
if type(numlist) ~= 'table' then return numlist end local sets = {} local mode local modeValue = 0 table.foreach(numlist,function(i,v) if sets[v] then sets[v] = sets[v] + 1 else sets[v] = 1 end end) for i,v in next,sets do if v > modeValue then modeValue = v mode = i else if v == modeValue then if type(mode) == 'table' then table.insert(mode,i) else mode = {mode,i} end end end end return mode
end
result = mode({1,3,6,6,6,6,7,7,12,12,17}) print(result) result = mode({1, 1, 2, 4, 4}) if type(result) == 'table' then
for i,v in next,result do io.write(v..' ') end print ()
end</lang>
Mathematica
Built-in function commonest returns a list of the most common element(s), even is there is only one 'commonest' number. Example for multiple 'commonest' numbers and a single 'commonest' number: <lang Mathematica>
Commonest[{b, a, c, 2, a, b, 1, 2, 3}] Commonest[{1, 3, 2, 3}]
</lang> gives back: <lang Mathematica>
{b,a,2} {3}
</lang>
MATLAB
<lang Matlab>function modeValue = findmode(setOfValues)
modeValue = mode(setOfValues);
end</lang>
MUMPS
<lang MUMPS>MODE(X)
;X is assumed to be a list of numbers separated by "^" ;I is a loop index ;L is the length of X ;Y is a new array ;ML is the list of modes ;LOC is a placeholder to shorten the statement Q:'$DATA(X) "No data" Q:X="" "Empty Set" NEW Y,I,L,LOC SET L=$LENGTH(X,"^"),ML="" FOR I=1:1:L SET LOC=+$P(X,"^",I),Y(LOC)=$S($DATA(Y(LOC)):Y(LOC)+1,1:1) SET I="",I=$O(Y(I)),ML=I ;Prime the pump, rather than test for no data FOR S I=$O(Y(I)) Q:I="" S ML=$SELECT(Y($P(ML,"^"))>Y(I):ML,Y($P(ML,"^"))<Y(I):I,Y($P(ML,"^"))=Y(I):ML_"^"_I) QUIT ML</lang>
USER>W $$MODE^ROSETTA("1^2^3^2") 2 USER>W $$MODE^ROSETTA("1^2^3^2^3") 2^3 USER>W $$MODE^ROSETTA("") Empty Set
Objective-C
<lang objc>#import <Foundation/Foundation.h>
@interface NSArray (Mode) - (NSArray *)mode; @end
@implementation NSArray (Mode) - (NSArray *)mode {
NSCountedSet *seen = [NSCountedSet setWithArray:self]; int max = 0; NSMutableArray *maxElems = [NSMutableArray array]; NSEnumerator *enm = [seen objectEnumerator]; id obj; while( (obj = [enm nextObject]) ) { int count = [seen countForObject:obj]; if (count > max) { max = count; [maxElems removeAllObjects]; [maxElems addObject:obj]; } else if (count == max) { [maxElems addObject:obj]; } } return maxElems;
} @end</lang>
OCaml
<lang ocaml>let mode lst =
let seen = Hashtbl.create 42 in List.iter (fun x -> let old = if Hashtbl.mem seen x then Hashtbl.find seen x else 0 in Hashtbl.replace seen x (old + 1)) lst; let best = Hashtbl.fold (fun _ -> max) seen 0 in Hashtbl.fold (fun k v acc -> if v = best then k :: acc else acc) seen []</lang>
# mode [1;3;6;6;6;6;7;7;12;12;17];; - : int list = [6] # mode [1;1;2;4;4];; - : int list = [4; 1]
Octave
Of course Octave has the mode function; but it returns only the "lowest" mode if multiple modes are available.
<lang octave>function m = mode2(v)
sv = sort(v); % build two vectors, vals and c, so that % c(i) holds how many times vals(i) appears i = 1; c = []; vals = []; while (i <= numel(v) ) tc = sum(sv==sv(i)); % it would be faster to count % them "by hand", since sv is sorted... c = [c, tc]; vals = [vals, sv(i)]; i += tc; endwhile % stack vals and c building a 2-rows matrix x x = cat(1,vals,c); % sort the second row (frequencies) into t (most frequent % first) and take the "original indices" i ... [t, i] = sort(x(2,:), "descend"); % ... so that we can use them to sort columns according % to frequencies nv = x(1,i); % at last, collect into m (the result) all the values % having the same bigger frequency r = t(1); i = 1; m = []; while ( t(i) == r ) m = [m, nv(i)]; i++; endwhile
endfunction</lang>
<lang octave>a = [1, 3, 6, 6, 6, 6, 7, 7, 12, 12, 17]; mode2(a) mode(a)
a = [1, 1, 2, 4, 4]; mode2(a) % returns 1 and 4 mode(a) % returns 1 only</lang>
Oz
<lang oz>declare
fun {Mode Xs} Freq = {Dictionary.new} for X in Xs do Freq.X := {CondSelect Freq X 0} + 1 end MaxCount = {FoldL {Dictionary.items Freq} Max 0} in for Value#Count in {Dictionary.entries Freq} collect:C do if Count == MaxCount then {C Value} end end end
in
{Show {Mode [1 2 3 3 2 1 1]}} {Show {Mode [1 2 3 3 2 1]}}</lang>
PARI/GP
<lang parigp>mode(v)={
my(count=1,r=1,b=v[1]); v=vecsort(v); for(i=2,#v, if(v[i]==v[i-1], count++ , if(count>r, r=count; b=v[i-1] ); count=1 ) ); if(count>r,v[#v],b)
};</lang>
Perl
<lang perl>use strict; use List::Util qw(max);
sub mode {
my %c; foreach my $e ( @_ ) {
$c{$e}++;
} my $best = max(values %c); return grep { $c{$_} == $best } keys %c;
}</lang>
<lang perl>print "$_ " foreach mode(1, 3, 6, 6, 6, 6, 7, 7, 12, 12, 17); print "\n"; print "$_ " foreach mode(1, 1, 2, 4, 4); print "\n";</lang>
Perl 6
<lang perl6>sub mode (@a) {
my %counts; ++%counts{$_} for @a; my $max = [max] values %counts; return map { .key }, grep { .value == $max }, %counts.pairs;
}</lang>
PHP
Note: this function only works with strings and integers, as those are the only things that can be used as keys of an (associative) array in PHP. <lang php><?php function mode($arr) {
$count = array_count_values($arr); $best = max($count); return array_keys($count, $best);
}
print_r(mode(array(1, 3, 6, 6, 6, 6, 7, 7, 12, 12, 17))); print_r(mode(array(1, 1, 2, 4, 4))); ?></lang>
PicoLisp
<lang PicoLisp>(de modes (Lst)
(let A NIL (for X Lst (accu 'A X 1) ) (let M (cdr (maxi cdr A)) (extract '((X) (and (= M (cdr X)) (car X))) A ) ) ) )</lang>
Output:
: (modes (1 3 6 6 6 6 7 7 12 12 17)) -> (6) : (modes (1 1 2 4 4)) -> (4 1) : (modes (chop "ABRAHAMASANTACLARA")) -> ("A")
PureBasic
<lang PureBasic>Procedure mean(Array InArray(1))
Structure MyMean Value.i Cnt.i EndStructure
Protected i, max, found Protected NewList MyDatas.MyMean() Repeat found=#False ForEach MyDatas() If InArray(i)=MyDatas()\Value MyDatas()\Cnt+1 found=#True Break EndIf Next If Not found AddElement(MyDatas()) MyDatas()\Value=InArray(i) MyDatas()\cnt+1 EndIf If MyDatas()\Cnt>max max=MyDatas()\Cnt EndIf i+1 Until i>ArraySize(InArray()) ForEach MyDatas() If MyDatas()\Cnt=max For i=1 To max Print(Str(MyDatas()\Value)+" ") Next EndIf Next
EndProcedure</lang>
Python
The following solutions require that the elements be hashable.
<lang python>>>> from collections import defaultdict >>> def modes(values): count = defaultdict(int) for v in values: count[v] +=1 best = max(count.values()) return [k for k,v in count.items() if v == best]
>>> modes([1,3,6,6,6,6,7,7,12,12,17]) [6] >>> modes((1,1,2,4,4)) [1, 4]</lang>
<lang python>>>> from collections import Counter >>> def modes(values): count = Counter(values) best = max(count.values()) return [k for k,v in count.items() if v == best]
>>> modes([1,3,6,6,6,6,7,7,12,12,17]) [6] >>> modes((1,1,2,4,4)) [1, 4]</lang>
If you just want one mode (instead of all of them), here's a one-liner for that: <lang python>def onemode(values):
return max(set(values), key=values.count)</lang>
R
<lang R>statmode <- function(v) {
a <- sort(table(v), decreasing=TRUE) r <- c() for(i in 1:length(a)) { if ( a1 == ai ) { r <- c(r, as.integer(names(a)[i])) } else break; # since it's sorted, once we find # a different value, we can stop } r
}
print(statmode(c(1, 3, 6, 6, 6, 6, 7, 7, 12, 12, 17))) print(statmode(c(1, 1, 2, 4, 4)))</lang>
REXX
<lang rexx> /*REXX program to find the mode of a vector. */
/*--------vector---------- ---show vector--- ----show result------ */
v='1 8 6 0 1 9 4 6 1 9 9 9' ; say 'vector='v; say 'mode='mode(v); say v='1 2 3 4 5 6 7 8 9 10 11' ; say 'vector='v; say 'mode='mode(v); say v='8 8 8 2 2 2' ; say 'vector='v; say 'mode='mode(v); say v='cat kat Cat emu emu Kat' ; say 'vector='v; say 'mode='mode(v); say
exit
/*-------------------------------------MAKEARRAY subroutine-------------*/
makeArray: procedure expose @.; parse arg v; @.0=words(v) /*make array*/
do j=1 for @.0 @.j=word(v,j) end
return
/*-------------------------------------ESORT subroutine-----------------*/
esort: procedure expose @.; h=@.0 /*exchange sort.*/
do while h>1; h=h%2 do i=1 for @.0-h;j=i;k=h+i do while @.k<@.j;t=@.j;@.j=@.k;@.k=t;if h>=j then leave;j=j-h;k=k-h;end end /*i*/ end
return
/*-------------------------------------MODE subroutine------------------*/
mode: procedure expose @.; parse arg x /*find the MODE of a vector. */
call makeArray x /*make an array out of the vector*/
call esort @.0 /*sort the array elements. */
?=@.1 /*assume 1st element is the mode.*/
freq=1 /*the frequency of the occurance.*/
do j=1 for @.0 _=j-freq if @.j==@._ then do freq=freq+1 /*bump the frequency count. */ ?=@.j /*this is the one. */ end end
return ? </lang> Output:
vector=1 8 6 0 1 9 4 6 1 9 9 9 mode=9 vector=1 2 3 4 5 6 7 8 9 10 11 mode=1 vector=8 8 8 2 2 2 mode=2 vector=cat kat Cat emu emu Kat mode=emu
Ruby
Here's two methods, the first more Ruby-ish, the second perhaps a bit more efficient. <lang ruby>def mode(ary)
seen = Hash.new(0) ary.each {|value| seen[value] += 1} max = seen.values.max seen.find_all {|key,value| value == max}.map {|key,value| key}
end
def mode_one_pass(ary)
seen = Hash.new(0) max = 0 max_elems = [] ary.each do |value| seen[value] += 1 if seen[value] > max max = seen[value] max_elems = [value] elsif seen[value] == max max_elems << value end end max_elems
end
p mode([1, 3, 6, 6, 6, 6, 7, 7, 12, 12, 17]) # => [6] p mode([1, 1, 2, 4, 4]) # => [1, 4] p mode_one_pass([1, 3, 6, 6, 6, 6, 7, 7, 12, 12, 17]) # => [6] p mode_one_pass([1, 1, 2, 4, 4]) # => [1, 4]</lang>
If you just want one mode (instead of all of them), here's a one-liner for that: <lang ruby>def one_mode(ary)
ary.max_by { |x| ary.count(x) }
end</lang>
Scala
Receiving any collection is easy. Returning the result in the same collection takes some doing.
<lang scala>import scala.collection.breakOut import scala.collection.generic.CanBuildFrom def mode
[T, CC[X] <: Seq[X]](coll: CC[T]) (implicit o: T => Ordered[T], cbf: CanBuildFrom[Nothing, T, CC[T]]) : CC[T] = { val grouped = coll.groupBy(x => x).mapValues(_.size).toSeq val max = grouped.map(_._2).max grouped.filter(_._2 == max).map(_._1)(breakOut)
}</lang>
Slate
<lang Slate> s@(Sequence traits) mode [| sortedCounts |
sortedCounts: (s as: Bag) sortedCounts. (sortedCounts mapSelect: [| :count :elem | sortedCounts last count = count]) valueSet
]. </lang>
Smalltalk
This code is able to find the mode of any collection of any kind of object. <lang smalltalk>OrderedCollection extend [
mode [ |s| s := self asBag sortedByCount. ^ (s select: [ :k | ((s at: 1) key) = (k key) ]) collect: [:k| k value] ]
].
- ( 1 3 6 6 6 6 7 7 12 12 17 ) asOrderedCollection
mode displayNl.
- ( 1 1 2 4 4) asOrderedCollection
mode displayNl.</lang>
Tcl
<lang tcl># Can find the modal value of any vector of values proc mode {n args} {
foreach n [list $n {*}$args] { dict incr counter $n } set counts [lsort -stride 2 -index 1 -decreasing $counter] set best {} foreach {n count} $counts { if {[lindex $counts 1] == $count} { lappend best $n } else break } return $best
}
- Testing
puts [mode 1 3 6 6 6 6 7 7 12 12 17]; # --> 6 puts [mode 1 1 2 4 4]; # --> 1 4</lang> Note that this works for any kind of value.
UNIX Shell
<lang bash>#!/bin/bash
function mode {
declare -A map max=0 for x in "$@"; do
tmp=$((${map[$x]} + 1)) map[$x]=$tmp ((tmp > max)) && max=$tmp
done for x in "${!map[@]}"; do
[[ ${map[$x]} == $max ]] && echo -n "$x "
done echo
}</lang> <lang bash>mode 1 2 1 2 a b a b a 2 a 2</lang>
Ursala
The mode function defined below works on lists of any type and returns a list of the modes. There is no concept of a general collection in Ursala. The algorithm is to partition the list by equality, then partition the classes by their lengths, and then select a representative from each member of the set of classes with the maximum length.
<lang Ursala>#import std
mode = ~&hS+ leql$^&h+ eql|=@K2
- cast %nLW
examples = mode~~ (<1,3,6,6,6,7,7,12,12,17>,<1,1,2,4,4>)</lang> The function is tested on a pair of lists, one with a unique mode and one with multiple modes. Here is the output.
(<6>,<4,1>)
Vedit macro language
Current edit buffer stores the collection. Each line is an item in the collection. The items can be any type (ascii text, numeric values in ascii, binary values). However, binary file with fixed record length would require some modifications to the code.
The "mode" item and it's count are displayed on status line. If there are multiple items with the same count, the smallest one is displayed. <lang vedit> BOF // Copy all data to a new buffer Reg_Copy(10, ALL) Buf_Switch(Buf_Free) Reg_Ins(10)
Sort(0, File_Size) // Sort the data
BOF repeat(ALL) { // Count & delete duplicate lines
#1 = 1 while(Match("^{.*}\N\1$", REGEXP)==0) { Del_Line(1) #1++ } Num_Ins(#1, NOCR) // Insert item count at the beginning of line Ins_Char(9) // TAB Line(1, ERRBREAK) // Next (different) line
}
Sort(0, File_Size, REVERSE) // Sort according to the count
BOF // Display the results Reg_Copy_Block(10, CP, EOL_pos) Buf_Quit(OK) Statline_Message(@10) </lang>
XEmacs Lisp
This returns a list of the modes. Any type(s) of data can be passed in, and any "equal" predicate function can be specified.
<lang xelisp>(defun mode ( predicate &rest values)
"Finds the mode of all values passed in.
Uses `predicate' to compare items."
(let ((modes nil) ; Declare local variables
(mode-count 0) (value-list nil) current-value)
(loop for value in values do (if (setq current-value (assoc* value value-list :test predicate)) ; Construct a linked list of cons cells, (value . count)
(incf (cdr current-value)) (push (cons value 1) value-list)))
(loop for (value . count) in value-list do ; Find modes in linked list (if (> count mode-count)
(setq modes (list value) mode-count count) (when (eq count mode-count) (push value modes))))
modes))
</lang>
- Programming Tasks
- Probability and statistics
- ActionScript
- Ada
- AutoHotkey
- C
- C++
- C sharp
- Clojure
- Common Lisp
- D
- E
- Euphoria
- F Sharp
- Fortran
- Haskell
- GAP
- Go
- Icon
- Unicon
- J
- Java
- JavaScript
- K
- Lua
- Mathematica
- MATLAB
- MUMPS
- Objective-C
- OCaml
- Octave
- Oz
- PARI/GP
- Perl
- Perl 6
- PHP
- PicoLisp
- PureBasic
- Python
- R
- REXX
- Ruby
- Scala
- Slate
- Smalltalk
- Tcl
- UNIX Shell
- Ursala
- Vedit macro language
- XEmacs Lisp