etude econometrique
D’après les graphes on remarque X2 X3 X4 X6 ne sont pas stationnaires
On va travailler avec le logarithme népérien sur les variables X2 X3 X4 X5 X6
Tableau de régression
Dependent Variable: LVENTE
Method: Least Squares
Date: 06/26/14 Time: 15:19
Sample: 1971 1984
Included observations: 14
Variable
Coefficient
Std. Error t-Statistic Prob.
LX2
0.108262
0.040065
2.702182
0.0270
LX3
-0.243122
0.074207
-3.276263
0.0113
LX4
0.130651
0.057756
2.262106
0.0535
LX5
-0.001937
0.005632
-0.343933
0.7398
LX6
-0.004261
0.091519
-0.046554
0.9640
C
0.636051
0.858231
0.741119
0.4798
R-squared
0.885952
Mean dependent var
0.796445
Adjusted R-squared
0.814672
S.D. dependent var
0.005776
S.E. of regression
0.002487
Akaike info criterion
-8.858152
Sum squared resid
4.95E-05
Schwarz criterion
-8.584270
Log likelihood
68.00706
Hannan-Quinn criter.
-8.883504
F-statistic
12.42917
Durbin-Watson stat
1.862562
Prob(F-statistic)
0.001323
D’après le tableau d’Eviews on remarque qu’il ya plusieurs variables non significatives (L’effet masque) donc il ya un problème de multi colinéarité
Matrice de corrélation
LVENTE
LX2
LX3
LX4
LX5
LX6
LVENTE
1
-0.1005564240959154
-0.14626775376035
-0.07172365684092617
-0.5119627841398956
0.01655601283032362
LX2
-0.1005564240959154
1
0.99581906719296
0.9930380259044055
0.585051462261492
0.9737095026595129
LX3
-0.14626775376035
0.99581906719296
1
0.9960254722074346
0.6187418394115306
0.9739734804488466
LX4
-0.07172365684092617
0.9930380259044055
0.9960254722074346
1
0.5849956142550936
0.986779566386667
LX5
-0.5119627841398956
0.585051462261492
0.6187418394115306
0.5849956142550936
1
0.5995276520357094
LX6
0.01655601283032362
0.9737095026595129
0.9739734804488466
0.986779566386667
0.5995276520357094
1
il ya forte corrélation entre lx3 et Lx4 et LX2 et LX6 et entre LX4 LX6 ce qui va entrainer un problème de multi colinéarité
Variance Inflation