04.03.2014 Views

Complete Volume - Institute of Business Management

Complete Volume - Institute of Business Management

Complete Volume - Institute of Business Management

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Research<br />

Process Capability Analysis for Non Normal Data<br />

ˆ<br />

⎡ ∂<br />

V ( α ) = ⎢<br />

⎣∂α<br />

V<br />

The three-parameter Weibull distribution probability<br />

density function is given by<br />

f<br />

( x)<br />

β ⎛ x −γ<br />

=<br />

⎞<br />

⎜ ⎟<br />

α ⎝ α ⎠<br />

⎡<br />

⎢<br />

⎛ x −γ<br />

exp −<br />

⎞<br />

⎜ ⎟<br />

⎢⎣<br />

⎝ α ⎠<br />

β −1<br />

β<br />

where x > γ , γ > 0 , α > 0, β > 0 . We may derive the twoparameter<br />

Weibull distribution by assuming the location<br />

parameter “equal to zero”. Many distributions are special cases<br />

<strong>of</strong> Weibull distribution, for example exponential distribution is a<br />

transformed form <strong>of</strong> Weibull distribution with shape parameter<br />

. β = 1<br />

We consider the two-parameter Weibull distribution with<br />

β<br />

probability function given as F ( x) = 1 − exp −( x α ) ,<br />

where x>0. A linear regression model was developed using the<br />

cumulative distribution function, as described below.<br />

F x = 1 − exp − x α β<br />

[ ]<br />

( ) ( )<br />

⇒ ln( x) − ln( α ) = [ ln{ − ln ( 1 − F( x)<br />

)}]<br />

β<br />

⇒ ln( x) = ln ( α ) + [ ln{ − ln ( 1 − F( x)<br />

)}] β<br />

The last expression (equation 2) is equivalent to<br />

simple linear regression model, y = β 0 + β1<br />

z + e , where<br />

( ) ( ) [ { ( ( ))}<br />

]<br />

y = ln x , β0 = ln α , β1<br />

= 1 β ,and z = ln −ln<br />

1−F<br />

x<br />

Least square estimators <strong>of</strong> both parameters were determined and<br />

standard errors were obtained using the Delta method (Stuart<br />

and Ord, 1987) as given below:<br />

2<br />

2<br />

⎤<br />

⎡<br />

2<br />

2<br />

exp( ) (<br />

0<br />

) [ exp(<br />

0<br />

)] (<br />

0<br />

) [ exp(<br />

0)<br />

] ∑ y ⎤<br />

i<br />

β0<br />

⎥ V β = β V β = β ⎢<br />

2 ⎥σ<br />

⎦<br />

⎢⎣<br />

n∑<br />

( y<br />

i<br />

− y ) ⎥⎦<br />

2<br />

2<br />

4<br />

( ˆ<br />

⎡ ∂ ⎛ 1 ⎞⎤<br />

⎡ 1 ⎤ ⎡⎛<br />

1 ⎞⎤<br />

) (<br />

1<br />

) ⎢⎜<br />

⎛<br />

⎟<br />

⎞<br />

β = ⎜ ⎟ V β = − ⎥ V ( β<br />

1<br />

) = ⎜ ⎟ V ( β<br />

1<br />

)<br />

⎢ ⎥<br />

⎣ ∂β<br />

⎝ β1<br />

⎠⎦<br />

⎜<br />

⎢⎣<br />

⎝<br />

β<br />

2<br />

1<br />

⎟<br />

⎠⎥⎦<br />

⎢ ⎥<br />

⎣⎝<br />

β1<br />

⎠⎦<br />

⎡⎛<br />

1 ⎞⎤<br />

= ⎢⎜<br />

⎟⎥<br />

⎣⎝<br />

β1<br />

⎠⎦<br />

4<br />

2<br />

⎡ 1 ⎤<br />

⎢<br />

2<br />

⎥σ<br />

⎢⎣<br />

∑ ( y i − y ) ⎥⎦<br />

2<br />

⎤<br />

⎥<br />

⎥⎦<br />

[ ]<br />

− − − − − − − − − − −<br />

− − −<br />

......2<br />

(3)<br />

− − ( 4)<br />

PAKISTAN BUSINESS REVIEW JULY 2010<br />

239

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!