The sign function  or signum function  (from Latin  signum    'sign'  ) is a function  in mathematics  that assigns its  sign  to a  real  or  complex number  .
 
Sign function on the real numbers definition The real sign function maps the set of real numbers  into the set   and is usually defined as follows:
  
    
      
        { 
        - 
        1 
        , 
        0 
        , 
        1 
        } 
       
     
    {\ displaystyle \ {- 1,0,1 \}} 
   
 
  
    
      
        so-called 
         
        ( 
        x 
        ) 
        : = 
        
          
            { 
            
              
                
                  + 
                  1 
                 
                
                  
                    if 
                   
                  x 
                  > 
                  0 
                 
               
              
                
                  0 
                 
                
                  
                    if 
                   
                  x 
                  = 
                  0 
                 
               
              
                
                  - 
                  1 
                 
                
                  
                    if 
                   
                  x 
                  < 
                  0 
                 
               
             
             
         
       
     
    {\ displaystyle \ operatorname {sgn} (x): = {\ begin {cases} +1 & \; {\ text {falls}} \ quad x> 0 \\\; \; \, 0 & \; {\ text { if}} \ quad x = 0 \\ - 1 & \; {\ text {falls}} \ quad x <0 \\\ end {cases}}} 
   
 She assigns the positive  numbers the value +1, the negative numbers the value −1 and the 0 the value 0.
In computing technology applications, the special position of 0 is sometimes dispensed with; this is then assigned to the positive, negative or both number ranges. This allows the sign of a number to be encoded in a single bit  . Since the zero is a zero set  below the Lebesgue measure  , this is not important for all practical applications.
In the event that is set, there is the following relationship to the Heaviside function  :
  
    
      
        so-called 
         
        ( 
        0 
        ) 
        = 
        1 
       
     
    {\ displaystyle \ operatorname {sgn} (0) = 1} 
   
 
  
    
      
        Θ 
        ( 
        x 
        ) 
       
     
    {\ displaystyle \ Theta (x)} 
   
  
  
    
      
        so-called 
         
        ( 
        x 
        ) 
        = 
        2 
        Θ 
        ( 
        x 
        ) 
        - 
        1 
       
     
    {\ displaystyle \ operatorname {sgn} (x) = 2 \ Theta (x) -1} 
   
 Calculation rules By differentiating between cases it is easy to prove:
For everyone with amount  applies .
  
    
      
        x 
        ∈ 
        
          R. 
         
       
     
    {\ displaystyle x \ in \ mathbb {R}} 
   
 
  
    
      
        
          | 
         
        x 
        
          | 
         
       
     
    {\ displaystyle | x |} 
   
 
  
    
      
        x 
        = 
        
          | 
         
        x 
        
          | 
         
        ⋅ 
        so-called 
         
        ( 
        x 
        ) 
        
          such as 
         
        x 
        ⋅ 
        so-called 
         
        ( 
        x 
        ) 
        = 
        
          | 
         
        x 
        
          | 
         
       
     
    {\ displaystyle x = | x | \ cdot \ operatorname {sgn} (x) \; {\ text {and}} \; x \ cdot \ operatorname {sgn} (x) = | x |} 
   
   
  
    
      
        so-called 
         
        ( 
        - 
        x 
        ) 
        = 
        - 
        so-called 
         
        ( 
        x 
        ) 
       
     
    {\ displaystyle \ operatorname {sgn} (- x) = - \ operatorname {sgn} (x)} 
   
 
  
    
      
        x 
        ∈ 
        
          R. 
         
       
     
    {\ displaystyle x \ in \ mathbb {R}} 
   
  If is a constant and an odd function  , then
  
    
      
        k 
        ∈ 
        
          R. 
         
       
     
    {\ displaystyle k \ in \ mathbb {R}} 
   
 
  
    
      
        f 
       
     
    {\ displaystyle f} 
   
   
  
    
      
        f 
        ( 
        k 
        ⋅ 
        x 
        ) 
        = 
        f 
        ( 
        so-called 
         
        ( 
        k 
        ) 
        ⋅ 
        
          | 
         
        k 
        
          | 
         
        ⋅ 
        x 
        ) 
        = 
        so-called 
         
        ( 
        k 
        ) 
        ⋅ 
        f 
        ( 
        
          | 
         
        k 
        
          | 
         
        ⋅ 
        x 
        ) 
       
     
    {\ displaystyle f (k \ cdot x) \ quad = f (\ operatorname {sgn} (k) \ cdot | k | \ cdot x) \ quad = \ operatorname {sgn} (k) \ cdot f (| k | \ cdot x)} 
   
 For the transition to the reciprocal number is compatible with the signum function and does not change its value:
  
    
      
        x 
        ≠ 
        0 
       
     
    {\ displaystyle x \ neq 0} 
   
  
  
    
      
        so-called 
         
        ( 
        
          x 
          
            - 
            1 
           
         
        ) 
        = 
        ( 
        so-called 
         
        ( 
        x 
        ) 
        
          ) 
          
            - 
            1 
           
         
        = 
        so-called 
         
        ( 
        x 
        ) 
       
     
    {\ displaystyle \ operatorname {sgn} (x ^ {- 1}) = (\ operatorname {sgn} (x)) ^ {- 1} = \ operatorname {sgn} (x)} 
   
 
  
    
      
        0 
        ≠ 
        x 
        ∈ 
        
          R. 
         
       
     
    {\ displaystyle 0 \ neq x \ in \ mathbb {R}} 
   
  The signum function is compatible with multiplication: 
  
    
      
        so-called 
         
        ( 
        x 
        ) 
        ⋅ 
        so-called 
         
        ( 
        y 
        ) 
        = 
        so-called 
         
        ( 
        x 
        ⋅ 
        y 
        ) 
       
     
    {\ displaystyle \ operatorname {sgn} (x) \ cdot \ operatorname {sgn} (y) = \ operatorname {sgn} (x \ cdot y)} 
   
 
  
    
      
        x 
        , 
        y 
        ∈ 
        
          R. 
         
       
     
    {\ displaystyle x, y \ in \ mathbb {R}} 
   
  
  
    
      
        so-called 
         
        ( 
        so-called 
         
        ( 
        x 
        ) 
        ) 
        = 
        so-called 
         
        ( 
        x 
        ) 
       
     
    {\ displaystyle \ operatorname {sgn} (\ operatorname {sgn} (x)) = \ operatorname {sgn} (x)} 
   
 
  
    
      
        x 
        ∈ 
        
          R. 
         
       
     
    {\ displaystyle x \ in \ mathbb {R}} 
   
  From the last two calculation rules mentioned, it follows, for example, that the signum function in an argument composed of any number of factors can be replaced by a factor without changing the function value:
  
    
      
        
          x 
          
            j 
           
         
       
     
    {\ displaystyle x_ {j}} 
   
 
  
    
      
        so-called 
         
        ( 
        
          x 
          
            j 
           
         
        ) 
       
     
    {\ displaystyle \ operatorname {sgn} (x_ {j})} 
   
 
  
    
      
        so-called 
         
        
          
            ( 
           
         
        
          x 
          
            j 
           
         
        ⋅ 
        
          ∏ 
          
            i 
           
         
        
          x 
          
            i 
           
         
        
          
            ) 
           
         
        = 
        so-called 
         
        
          
            ( 
           
         
        so-called 
         
        ( 
        
          x 
          
            j 
           
         
        ) 
        ⋅ 
        
          ∏ 
          
            i 
           
         
        
          x 
          
            i 
           
         
        
          
            ) 
           
         
       
     
    {\ displaystyle \ operatorname {sgn} {\ bigg (} x_ {j} \ cdot \ prod _ {i} x_ {i} {\ bigg)} = \ operatorname {sgn} {\ bigg (} \ operatorname {sgn} (x_ {j}) \ cdot \ prod _ {i} x_ {i} {\ bigg)}} 
   
 
  
    
      
        
          x 
          
            i 
           
         
        , 
        
          x 
          
            j 
           
         
        ∈ 
        
          R. 
         
       
     
    {\ displaystyle x_ {i}, x_ {j} \ in \ mathbb {R}} 
   
  Derivative and integral 
The sign function is not continuous at the 0 position.
 
 The sign function is not continuous at this point and therefore not classically differentiable there. For all other places the sign function can be differentiated with . The sign function also has no weak derivative  . However, it is differentiable in the sense of  distributions  , and its derivative is , where denotes the  delta distribution  .
  
    
      
        x 
        = 
        0 
       
     
    {\ displaystyle x = 0} 
   
 
  
    
      
        x 
        ≠ 
        0 
       
     
    {\ displaystyle x \ neq 0} 
   
 
  
    
      
        
          so-called 
          
            ′ 
           
         
         
        ( 
        x 
        ) 
        = 
        0 
       
     
    {\ displaystyle \ operatorname {sgn} ^ {\ prime} (x) = 0} 
   
 
  
    
      
        2 
        δ 
       
     
    {\ displaystyle 2 \ delta} 
   
 
  
    
      
        δ 
       
     
    {\ displaystyle \ delta} 
   
  
Furthermore applies to everyone 
  
    
      
        x 
        ∈ 
        
          R. 
         
       
     
    {\ displaystyle x \ in \ mathbb {R}} 
   
 
  
    
      
        
          | 
         
        x 
        
          | 
         
        = 
        
          ∫ 
          
            0 
           
          
            x 
           
         
        so-called 
         
        ( 
        t 
        ) 
        d 
        t 
        . 
       
     
    {\ displaystyle | x | = \ int _ {0} ^ {x} \ operatorname {sgn} (t) \, dt \ ,.} 
   
 The sign function is also the weak derivative of  the absolute value function  .
Sign function on the complex numbers definition 
Signum of four complex numbers
 
 Compared to the sign function of real numbers, the following extension to complex numbers is rarely considered:
  
    
      
        so-called 
         
        ( 
        z 
        ) 
        : = 
        
          
            { 
            
              
                
                  
                    
                      z 
                      
                        
                          | 
                         
                        z 
                        
                          | 
                         
                       
                     
                   
                 
                
                  
                    if 
                   
                  z 
                  ≠ 
                  0 
                 
               
              
                
                  0 
                 
                
                  
                    if 
                   
                  z 
                  = 
                  0 
                 
               
             
             
         
       
     
    {\ displaystyle \ operatorname {sgn} (z): = {\ begin {cases} {\ frac {z} {| z |}} & \; {\ text {if}} \ quad z \ neq 0 \\ 0 & \; {\ text {falls}} \ quad z = 0 \\\ end {cases}}} 
   
 The result of this function is on the unit circle  and has the same argument as the output value, in particular applies
  
    
      
        z 
        ≠ 
        0 
       
     
    {\ displaystyle z \ neq 0} 
   
  
  
    
      
        so-called 
         
        
          ( 
          
            r 
            
              
                e 
               
              
                
                  i 
                 
                φ 
               
             
           
          ) 
         
        = 
        
          
            e 
           
          
            
              i 
             
            φ 
           
         
        , 
        
          f 
          a 
          l 
          l 
          s 
         
          
        r 
        > 
        0. 
       
     
    {\ displaystyle \ operatorname {sgn} \ left (r \ mathrm {e} ^ {\ mathrm {i} \ varphi} \ right) = \ mathrm {e} ^ {\ mathrm {i} \ varphi}, \ qquad \ mathrm {if} \ r> 0.} 
   
 Example:   (red in the picture)
  
    
      
        
          z 
          
            1 
           
         
        = 
        2 
        + 
        2 
        
          i 
         
       
     
    {\ displaystyle z_ {1} = 2 + 2 \ mathrm {i}} 
   
 
  
    
      
        so-called 
         
        ( 
        
          z 
          
            1 
           
         
        ) 
        = 
        so-called 
         
        ( 
        2 
        + 
        2 
        
          i 
         
        ) 
        = 
        
          
            
              2 
              + 
              2 
              
                i 
               
             
            
              | 
              
                2 
                + 
                2 
                
                  i 
                 
               
              | 
             
           
         
        = 
        
          
            
              2 
              + 
              2 
              
                i 
               
             
            
              2 
              
                
                  2 
                 
               
             
           
         
        = 
        
          
            
              1 
              + 
              
                i 
               
             
            
              2 
             
           
         
        = 
        
          
            
              2 
             
            2 
           
         
        + 
        
          
            
              2 
             
            2 
           
         
        
          
            i 
           
         
        . 
       
     
    {\ displaystyle \ operatorname {sgn} (z_ {1}) = \ operatorname {sgn} (2 + 2 \ mathrm {i}) = {\ frac {2 + 2 \ mathrm {i}} {\ left | 2+ 2 \ mathrm {i} \ right |}} = {\ frac {2 + 2 \ mathrm {i}} {2 {\ sqrt {2}}}} = {\ frac {1+ \ mathrm {i}} { \ sqrt {2}}} = {\ frac {\ sqrt {2}} {2}} + {\ frac {\ sqrt {2}} {2}} {\ mathrm {i}}.} 
   
 Calculation rules The following calculation rules apply to the complex sign function:
For all complex numbers   and the following applies:
  
    
      
        z 
       
     
    {\ displaystyle z} 
   
 
  
    
      
        w 
       
     
    {\ displaystyle w} 
   
 
  
    
      
        z 
        = 
        
          | 
         
        z 
        
          | 
         
        ⋅ 
        so-called 
         
        z 
       
     
    {\ displaystyle z = | z | \ cdot \ operatorname {sgn} z} 
   
 amount  of ;
  
    
      
        z 
        , 
       
     
    {\ displaystyle z,} 
   
 
  
    
      
        
          | 
         
        z 
        
          | 
         
       
     
    {\ displaystyle | z |} 
   
 
  
    
      
        z 
       
     
    {\ displaystyle z} 
   
   
  
    
      
        so-called 
         
        ( 
        
          
            
              z 
              ¯ 
             
           
         
        ) 
        = 
        
          
            
              so-called 
               
              ( 
              z 
              ) 
             
            ¯ 
           
         
       
     
    {\ displaystyle \ operatorname {sgn} ({\ bar {z}}) = {\ overline {\ operatorname {sgn} (z)}}} 
   
 complex conjugation  ; 
  
    
      
        so-called 
         
        ( 
        z 
        ⋅ 
        w 
        ) 
        = 
        so-called 
         
        z 
        ⋅ 
        so-called 
         
        w 
       
     
    {\ displaystyle \ operatorname {sgn} (z \ cdot w) = \ operatorname {sgn} z \ cdot \ operatorname {sgn} w} 
   
 
  
    
      
        so-called 
         
        ( 
        λ 
        ⋅ 
        z 
        ) 
        = 
        so-called 
         
        z 
       
     
    {\ displaystyle \ operatorname {sgn} (\ lambda \ cdot z) = \ operatorname {sgn} z} 
   
 
  
    
      
        λ 
       
     
    {\ displaystyle \ lambda} 
   
  
  
    
      
        so-called 
         
        ( 
        λ 
        ⋅ 
        z 
        ) 
        = 
        - 
        so-called 
         
        z 
       
     
    {\ displaystyle \ operatorname {sgn} (\ lambda \ cdot z) = - \ operatorname {sgn} z} 
   
 
  
    
      
        λ 
       
     
    {\ displaystyle \ lambda} 
   
  
  
    
      
        so-called 
         
        ( 
        - 
        z 
        ) 
        = 
        - 
        so-called 
         
        ( 
        z 
        ) 
       
     
    {\ displaystyle \ operatorname {sgn} (-z) = - \ operatorname {sgn} (z)} 
   
  
 
 
  
    
      
        so-called 
         
        ( 
        
          | 
         
        z 
        
          | 
         
        ) 
        = 
        
          | 
         
        so-called 
         
        ( 
        z 
        ) 
        
          | 
         
       
     
    {\ displaystyle \ operatorname {sgn} (| z |) = | \ operatorname {sgn} (z) |} 
   
  
If is, also applies
  
    
      
        z 
        ≠ 
        0 
       
     
    {\ displaystyle z \ neq 0} 
   
  
 
  
    
      
        so-called 
         
        ( 
        
          z 
          
            - 
            1 
           
         
        ) 
        = 
        so-called 
         
        ( 
        z 
        
          ) 
          
            - 
            1 
           
         
        = 
        
          
            
              so-called 
               
              ( 
              z 
              ) 
             
            ¯ 
           
         
       
     
    {\ displaystyle \ operatorname {sgn} (z ^ {- 1}) = \ operatorname {sgn} (z) ^ {- 1} = {\ overline {\ operatorname {sgn} (z)}}} 
   
  literature 
Königsberger: Analysis 1  . 6th edition. Springer, Berlin 2003, ISBN 3-540-40371-X  , p.  101  .  
 
Hildebrandt: Analysis 1  . 2nd Edition. Springer, Berlin 2005, ISBN 3-540-25368-8  , pp.  133  .  
 
 
Web links  
<img src="https://de.wikipedia.org//de.wikipedia.org/wiki/Special:CentralAutoLogin/start?type=1x1" alt="" title="" width="1" height="1" style="border: none; position: absolute;">