Skip to content

predict returns with the fSeries package

2 messages · Ricardo Zambrano Aguilera, Diethelm Wuertz

#
Dear List
 how i can predict returns with new data??
 > ajuste1

Title:
 GARCH Modelling 

Call:
 garchFit(formula.mean = ~arma(2, 0), formula.var = ~garch(1, 
    1), series = r_peso, cond.dist = "dnorm") 

Mean and Variance Equation:
 ~arma(2, 0) + ~garch(1, 1) 

Conditional Distribution:
 dnorm 

Coefficient(s):
          mu           ar1           ar2         omega        alpha1         beta1  
-1.25313e-04   6.10406e-02  -7.06526e-02   2.11754e-06   9.47503e-02   8.45540e-01  

Error Analysis:
         Estimate  Std. Error  t value Pr(>|t|)    
mu     -1.253e-04   1.466e-04   -0.855   0.3925    
ar1     6.104e-02   2.765e-02    2.208   0.0272 *  
ar2    -7.065e-02   2.785e-02   -2.537   0.0112 *  
omega   2.118e-06   1.114e-06    1.902   0.0572 .  
alpha1  9.475e-02   2.411e-02    3.929 8.51e-05 ***
beta1   8.455e-01   5.195e-02   16.277  < 2e-16 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 

Log Likelihood:
 -5473.802    normalized:  -3.751749 

Description:
 Thu Sep 07 11:50:50 2006 

#######
################## then....############################
meanForecast   meanError standardDeviation
1   0.0001657992 0.005873028       0.004574582
2  -0.0002516247 0.005883959       0.004668490
3  -0.0001535909 0.005897073       0.004755100
4  -0.0001181148 0.005897279       0.004835123
5  -0.0001228757 0.005897331       0.004909178
6  -0.0001256727 0.005897333       0.004977807
7  -0.0001255071 0.005897334       0.005041485
8  -0.0001252994 0.005897334       0.005100636
9  -0.0001252984 0.005897334       0.005155636
10 -0.0001253130 0.005897334       0.005206823



######### now if a put newdata it?s the same??, how i can know the returns for the next week if i put the returns of the last week????############
meanForecast   meanError standardDeviation
1   0.0001657992 0.005873028       0.004574582
2  -0.0002516247 0.005883959       0.004668490
3  -0.0001535909 0.005897073       0.004755100
4  -0.0001181148 0.005897279       0.004835123
5  -0.0001228757 0.005897331       0.004909178
6  -0.0001256727 0.005897333       0.004977807
7  -0.0001255071 0.005897334       0.005041485
8  -0.0001252994 0.005897334       0.005100636
9  -0.0001252984 0.005897334       0.005155636
10 -0.0001253130 0.005897334       0.005206823

My best regards Ricardo Z.
1 day later
#
Ricardo Zambrano Aguilera wrote:
Ther is no new data! It's like in the Arima case of R!

You model and fit the time series up to the end, and then you start
with your forecast at position n+1. since there can't is be no newdata,
you must get always the same result, as you did.

I hope this helps
Diethelm Wuertz