6
$\begingroup$

Binary variable is often used in applied statistics. However, I had a hard time to figure out how to create it. Somebody might have better idea how to do it. I have two variables say x and y as follows

x = Range[100];
y = Flatten[RandomInteger[{1, 100}, {100, 1}]];

I want to create a binary variable b1 such that b1 = 1 if x>y and 0 otherwise. I have done so far

b1 = TrueQ[#1 > #2] & @@@ Transpose[{x, y}] /. {True -> 1, False -> 0}

Any better way please?

$\endgroup$
0

4 Answers 4

5
$\begingroup$
Boole@Thread[Greater[x, y]] ==  MapThread[Boole@Greater@## &, {x, y}]
$\endgroup$
2
  • $\begingroup$ Thank you belisarius for your answer. $\endgroup$
    – ramesh
    Commented Feb 10, 2015 at 16:02
  • 2
    $\begingroup$ Or just Boole@Thread[x > y]. $\endgroup$
    – Karsten7
    Commented Feb 10, 2015 at 16:02
2
$\begingroup$
Inner[Greater, x, y, Boole @ {##} &]
Inner[Boole@Greater@## &, x, y, List]
Inner[Composition[Boole, Greater], x, y, List]
Boole @ Inner[Greater, x, y, List]

Update: Timings

ClearAll[r1, r2, r3, r4, r5, r6, r7, r8, r9, r10, 
 t1, t2, t3, t4, t5, t6, t7,  t8, t9, t10, 
 f1, f2, f3, f4, f5, f6, f7,  f8, f9, f10, functions, results, timings]
f1 = Inner[Greater, ##, Boole@{##} &] &;
f2 = Inner[Boole@Greater@## &, ##, List] &;
f3 = Inner[Composition[Boole, Greater], ##, List] &;
f4 = Boole@Inner[Greater, ##, List] &;
f5 = Boole[#1 > #2 & @@@ Transpose[{##}]] &;
f6 = Transpose[{##}] /. {x_, y_} :> Boole[x > y] &;
f7 = Boole@Thread[Greater[##]] &;
f8 = MapThread[Boole@Greater@## &, {##}] &;
f9 = 1 - UnitStep[#2 - #] &;
f10[x_, y_] := Subtract[1, UnitStep@Subtract[y, x]];

functions = {"f1", "f2", "f3", "f4", "f5", "f6", "f7", "f8", "f9", "f10"};
results = {r1, r2, r3, r4, r5, r6, r7, r8,r9, r10};
timings = {t1, t2, t3, t4, t5, t6, t7, t8, t9, t10};


SeedRandom[1]
x = Range[1000000];
y = Flatten[RandomInteger[{1, 100}, {1000000, 1}]]; 
(# = First[AbsoluteTiming[(#2 = ToExpression[#3][x, y]);]]) & @@@ 
 Transpose[{timings, results, functions}];

Equal @@ results

True

Grid[Prepend[SortBy[Transpose[{functions, ToExpression /@ functions, timings}], 
 Last], {"name", "function", "timing"}], Dividers -> All]

enter image description here

$\endgroup$
3
  • $\begingroup$ kguler, thank you for your answer. $\endgroup$
    – ramesh
    Commented Feb 10, 2015 at 16:03
  • $\begingroup$ Would you kindly add my code to your timings, unless I made a mistake or misunderstood? (In which case I'll delete.) $\endgroup$
    – Mr.Wizard
    Commented Jul 29, 2017 at 17:42
  • $\begingroup$ @Mr.Wizard, done ... $\endgroup$
    – kglr
    Commented Jul 29, 2017 at 17:58
1
$\begingroup$
Boole[#1 > #2 & @@@ Transpose[{x, y}]]

Or

Transpose[{x, y}] /. {x_, y_} :> Boole[x > y]
$\endgroup$
0
1
$\begingroup$

Maybe I'm missing something but I think I would just use UnitStep:

1 - UnitStep[y - x]

This will be a little faster with explicit Subtract due to (40927):

Subtract[1, UnitStep @ Subtract[y, x]]
$\endgroup$
1

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.