- 2
2: . .
:
, . 30 . , ..
:
.
.
, , .
, .
.
Statistica, Eviews, R.
: , . .
(): .
- 2. :
(: : 2015 , )
, .
, , . . . , , .
, , .
:
- , -10 , 4 .
- . .
|
|
- , .
- , STATISTICA, . Statistics Advanced Linear/ Nonelinear Models ( / ), Nonelinear Estimation ( ) (. 22).
-
-
- 22 -
- :
- User-specified regression least squares ( );
- User-specified regression, Custom loss function ( );
- Quick Logit regression ( );
- Quick Probit regression ( );
- Exponential growth regression ( );
- Piecewise linear regression (- );
- Nonlinear Estimation (.23) : OK, Cancel, Options, Open Data. , . , User-specified regression, custom loss function ( , ).
-
- 23 -
-
- (. 24), User-specified regression, custom loss function, : ; Cancel (); Options (); Open Data ( ); Analysis/Graph Selection Conditions ( /); Analysis/Graph Case Weights ( ); MD deletion ( ) 2 : (Casewise) (Mean substitution).
- , Function to be estimated& loss function (, & ), (. 25).
-
- 24 , .
-
- 25 .
- Estimated function ( ), , . Loss function , . (OBS-PRED)2. -, : Y = a 0* K ^ a 1* L ^ a 2. (), K , L , Y .
- OK, Model Estimation ( ) (. 26), Number of parameters to be estimated ( ), Loss function ( ), Dependent variable ( ), Independent variables ( ), number of valid cases ( ).
|
|
- 3 - Quick ( ), Advanced ( ), Review (); Quick () Estimation method ( )- Quasi-Newton,(-), Simplex (), Simplex and quasi-Newton ( -), Hooke-Jeeves (-) pattern moves, Hooke-Jeeves and quasi-Newton (- -), Rosenbrock pattern search. (). .
- -.
-
- 26 .
- Advanced (), (. 27), Maximum numbers of iterations ( ) convergence criterion ( ). Start values Initial step sizes .
-
- 27 .
- Result s (. 28). Quick ( ) : Summary Parameter estimates ( ), Observed, predicted, residual values (, , ), Fitted 3D function & observed values ( 3D & () ). (. 28) : Dependent variable ( ), Loss function ( ), Final value ( , ), Proportion of variance accounted for.. ( ..) ( ), Independent variables ( ), R- , .
-
- 28 .
- Summary Parameter estimates ( ), ( 0, 1, 2) 29.
-
- 29 .
- , : Y=0,0075*K^7,347*L^(-5,752).
- .
- , . Residuals () (. 30): Observed, predicted, residual values (, , ), Normal probability plot of residuals,( ), save predicted and residual values ( ), histogram of residuals ( ), half-normal probability plot ( - ), Predicted observed values ( ), Predicted residual values ( ).
-
- 30 .
- Observed, predicted, residual values (, , ), (. 31).
-
- 31 , , .
- . Frequency distribution residuals ( ), histogram of residuals ( ) , . 32:
|
|
-
- 32 .
- . Normal probability plot of residuals, (. 33):
-
- 33 .
- 32 33 , .
- . Predicted observed values ( 34):
-
- 34 .
- Predicted residual values ( 35).
-
- 35 .
- 34 35 , , .
- (. 36) Advanced ( ) : Flitted 3D function & observed values ( 3D & ). 3D .
-
- 36 3D .
- ( 37). , , , (R 2 = 0,2):
-
-
- 37 .
- , Review () ( 38), : Means & standard deviations ( ), Matrix plot for all variables ( ), Box & whisker plot for all variables ( ). .
-
- 38 .
, , . .
1. :
, .
2. - () :
, , .
:
,
, L , y , , . , , (: ).
- .
. . . , , . :
|
|
.
.
.
: , . L ≠0 :
,
: ; , :
, .
, , Z .
:
Y/K ();
K/Y ;
L/Y .
.
, , :
1. ;
2. , ;
3. .
, .
1. :
, (9)
, , n , , .
0 1. , . , R =1, .
2. R2 . , . :
3. , . :
.
5-7%, .
0 . 1, . , :
,
R2 , n , m .
- α F, . > , 0 .
Excel.
1
Y | Y | (1)-(2) | (3)/(2) | ABS((3)/(2)) |
0,18138176 | 0,18138176 | |||
-194496 | -0,19387117 | 0,193871166 | ||
0,1402792 | 0,1402792 | |||
0,013516522 | 0,013516522 | |||
-32021 | -0,04839289 | 0,048392894 | ||
0,323367026 | 0,323367026 | |||
0,255346885 | 0,255346885 | |||
-98027 | -0,14881192 | 0,148811922 | ||
0,079350034 | 0,079350034 | |||
0,042059677 | 0,042059677 | |||
-6098 | -0,01527327 | 0,015273273 | ||
0,000908675 | 0,000908675 | |||
-141369 | -0,28753796 | 0,287537959 | ||
-46342 | -0,11690711 | 0,116907106 | ||
-53725 | -0,14310212 | 0,143102124 | ||
0,054822241 | 0,054822241 | |||
-11652 | -0,03559494 | 0,035594942 | ||
-6332 | -0,02017774 | 0,020177743 | ||
0,004668069 | 0,004668069 | |||
-48840 | -0,16042627 | 0,160426272 | ||
2,265795489 |
=11,329>7%, .
|
|
0 . 1 . :
. Probability Calculator ( ) Statistica Distributions () (. 51).
51. F.
df1 df2.
, . .
> , 0 , p=0.95.
31,45173>8,6829, , 0 0,95.