Mean squared error

|Coefficient |Std. Error |t-Statistic |Prob. | |C |0.006797 |0.001141 |5.955348 |0.0000 | |Zit? |0.796464 |0.035906 |22.18189 |0.0000 | |Weighted Statistics | | | | | |R-squared |0.634029 | Mean dependent var |0.046236 | |Adjusted R-squared |0.633398

19:29 Included observations: 22 after adjusting end points Variable C LADPGR LREDUCGR LRSANTEGR R-squared Coefficient 3.219729 0.002606 -0.142340 0.193574 Std. Error 1.114356 0.043986 0.044247 0.024011 Ex pe rt PD F 0.798524 0.764945 0.034240 0.021103 45.22674 Mean dependent var Adjusted R-squared S.E. of regression S.D.dependentvar Akaike info criterion Schwarz criterion F-statistic Sum squared resid Log likelihood Durbin-Watson stat 1.241408 ~ Prob(F-statistic) Source: d'après résultats

credit spread prediction errors generated by prior-period estimates of the relation between the level and slope of the Treasury term structure and the credit spreads. These prediction errors shed light on the optimal trade-off between reducing sample error by using a long as possible time series and causing bias by using outdated data in estimates of the current relations between the Treasury term structure and the credit spread. The results provide a comprehensive analysis of the relation between credit

in which these algorithms are seen as learning eigenfunctions of a kernel. LLE and Isomap pose special challenges as the kernel is training-data dependent. Numerical experiments on real data show that the generalizations performed have a level of error comparable to the variability of the embedding algorithms to the choice of training data. 1 Introduction In the last few years, many unsupervised learning algorithms have been proposed which share the use of an eigendecomposition for obtaining

Variable: LOGY | Method: Least Squares | Date: 02/26/11 Time: 15:44 | Sample(adjusted): 2000:02 2009:12 | Included observations: 119 after adjusting endpoints | Convergence achieved after 3 iterations | Variable | Coefficient | Std. Error | t-Statistic | Prob. | T | 0.003161 | 0.000500 | 6.324724 | 0.0000 | D1 | 5.420490 | 0.045991 | 117.8610 | 0.0000 | D2 | 5.526848 | 0.044311 | 124.7272 | 0.0000 | D3 | 5.792309 | 0.043679 | 132.6119 | 0.0000 | D4 | 6.012761 | 0.043523

qualitatives. e e Avec une variable qualitative la valeur retourn´e est un vecteur. Avec deux e une data.frame... Pour sp´cifier les variables qualitatives on passe par une liste : e > aggregate(iris$Sepal.Length,list(Species=iris$Species,Random=gl(2,75)), mean ) 2 Species Random x 1 setosa 1 5.006 2 versicolor 1 6.012 3 versicolor 2 5.860 4 virginica 2 6.588 2.2 apply La fonction permet de r´aliser tr`s rapidement des calculs sur des data.frames. e e La fonction prend comme argument : – le nom

Méthode d’Extraction: Analyse en Composantes Principales Totale variance expliquée Component 1 Valeurs propres initiales % de la Total %Cumulés variance 2,679 66,975 66,975 2 ,581 14,525 81,500 3 ,406 10,158 91,658 Extraction Sums of Squared Loadings % de la Total %Cumulées Variance 2,679 66,975 66,975 4 ,334 8,342 100,000 Méthode Extraction: Analyse en Composantes Principales Matrice de composantes (a) Composante 1 Vous vous intéressez aux informations concernant la Vous essayez de

cours du pétrole a connu son plus bas prix qui est de 10 à 11$, le 25 Novembre 1998. Figure 1 : Evolution des cours bruts du pétrole sur la période 1986-2013 Data 1: Crude Oil RWTC Cushing, OK WTI Spot Price FOB (Dollars p 160 140 120 100 Stat 80 Mean 39,3 Std. Dev. 28,85 Skew. 1,21 Kurtosis 3,293 60 40 20 0 86 88 90 92 94 96 98 00 02 04 06 08 10 12 Source : WTI, nos calculs.  De 1999 à 2007 Pendant cette période on constate une croissance lente du cours du pétrole. Cette

18:45 Sample: 2000 2009 Included observations: 10 Variable Coefficient Std. Error t-Statistic Prob. C 9.244108 0.425019 21.74989 0.0000 Y 0.052569 0.005661 9.286755 0.0000 R-squared 0.915114 Mean dependent var 13.14000 Adjusted R-squared 0.904503 S.D. dependent var 0.697933 S.E. of regression 0.215679 Akaike info criterion -0.053190 Sum squared resid 0.372141 Schwarz criterion 0.007327 Log likelihood 2.265951 Hannan-Quinn criter

S- Fès En plus d’un bon ajustement, nous avons des résidus ayant une forme normale, mais la probabilité de Jarque Bera nous dit le contraire ; une stabilité de la série comme nous renseigne les figures ci-contre. 8 7 6 5 4 3 2 1 0 -0.1 0.0 0.1 0.2 Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis Jarque-Bera Probability 1.53e-16 -0.009417 0.224617 -0.077427 0.068469 1.843683 7.134155 24.29459 0.000005 Series: Residuals Sample 1982 2001 Observations 19 Normalité des résidus 1.6 1.2 0.8