How to estimate weekly Variance

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

How to estimate weekly Variance

Felipe Carrillo
Dear All:
Using the Excel formula below I am able to estimate
daily variance(Daily VarPD)for PD,but I can't figure
out how to estimate the Weekly variance for PD. Any
hints?

Daily VarPD = (PD * (1 - TD))/TD + VarTD *
((PD*(1-TD)+ PD^2*TD))/TD^3


WK TD          VARTD      PD  Daily VarPD
28 0.014735 0.0000449 0 0
28 0.012421 0.0000449 0 0
28 0.014396 0.0000449 71 7005
28 0.014898 0.0000449 132 14016
28 0.013512 0.0000449 223 32506
28 0.013886 0.0000449 224 31287
28 0.013778 0.0000449 218 30528
29 0.013274 0.0000449 228 34505
29 0.013072 0.0000449 224 34531
29 0.012518 0.0000449 488 117720
29 0.013264 0.0000449 525 119323
29 0.013233 0.0000449 80 9134
29 0.010194 0.0000449 417 133067
29 0.012567 0.0000449 82 10184
30 0.014695 0.0000449 413 68893
30 0.010634 0.0000449 373 103688
30 0.013934 0.0000449 914 272719
30 0.014107 0.0000449 213 28473
30 0.014339 0.0000449 651 147016
30 0.014375 0.0000449 521 102429
30 0.016373 0.0000449 979 229104


Felipe D. Carrillo
  Fishery Biologist
  US Fish & Wildlife Service
  California, USA



      ____________________________________________________________________________________
Be a better friend, newshound, and

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: How to estimate weekly Variance

David Winsemius
Felipe Carrillo <[hidden email]> wrote in
news:[hidden email]:

> Dear All:
> Using the Excel formula below I am able to estimate
> daily variance(Daily VarPD)for PD,but I can't figure
> out how to estimate the Weekly variance for PD. Any
> hints?
>
> Daily VarPD = (PD * (1 - TD))/TD + VarTD *
> ((PD*(1-TD)+ PD^2*TD))/TD^3
>
>
> WK     TD               VARTD      PD       Daily VarPD
> 28     0.014735     0.0000449     0     0
> 28     0.012421     0.0000449     0     0
> 28     0.014396     0.0000449     71     7005
> 28     0.014898     0.0000449     132     14016
> 28     0.013512     0.0000449     223     32506
> 28     0.013886     0.0000449     224     31287
> 28     0.013778     0.0000449     218     30528
> 29     0.013274     0.0000449     228     34505
> 29     0.013072     0.0000449     224     34531
> 29     0.012518     0.0000449     488     117720
> 29     0.013264     0.0000449     525     119323
> 29     0.013233     0.0000449     80     9134
> 29     0.010194     0.0000449     417     133067
> 29     0.012567     0.0000449     82     10184
> 30     0.014695     0.0000449     413     68893
> 30     0.010634     0.0000449     373     103688
> 30     0.013934     0.0000449     914     272719
> 30     0.014107     0.0000449     213     28473
> 30     0.014339     0.0000449     651     147016
> 30     0.014375     0.0000449     521     102429
> 30     0.016373     0.0000449     979     229104

I am going to assume on the basis of your earlier posts that you have
gotten this into a dataframe. You should know, since Prof Ripley told
you so six months ago, that variable names cannot have spaces. The  
formula you offer makes no sense to me, and I could not see what
bearing it has on calculating a variance of a week's worth of PD
values. In the future, you should offer code to create a minimal
dataset.

Assuming your dataframe is fishwk, here is a method to get the
variances of PD by week:

by(data=fishwk, INDICES=fishwk$WK, function(x) var(x$PD))


--
David Winsemius

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: How to estimate weekly Variance

jholtman
In reply to this post by Felipe Carrillo
Here is one way of doing it:

> x <- read.table(textConnection("WK      TD                VARTD      PD   DailyVarPD
+ 28      0.014735        0.0000449       0       0
+ 28      0.012421        0.0000449       0       0
+ 28      0.014396        0.0000449       71      7005
+ 28      0.014898        0.0000449       132     14016
+ 28      0.013512        0.0000449       223     32506
+ 28      0.013886        0.0000449       224     31287
+ 28      0.013778        0.0000449       218     30528
+ 29      0.013274        0.0000449       228     34505
+ 29      0.013072        0.0000449       224     34531
+ 29      0.012518        0.0000449       488     117720
+ 29      0.013264        0.0000449       525     119323
+ 29      0.013233        0.0000449       80      9134
+ 29      0.010194        0.0000449       417     133067
+ 29      0.012567        0.0000449       82      10184
+ 30      0.014695        0.0000449       413     68893
+ 30      0.010634        0.0000449       373     103688
+ 30      0.013934        0.0000449       914     272719
+ 30      0.014107        0.0000449       213     28473
+ 30      0.014339        0.0000449       651     147016
+ 30      0.014375        0.0000449       521     102429
+ 30      0.016373        0.0000449       979     229104"), header=TRUE)
> closeAllConnections()
> # daily using your formula
> x$newDaily <- with(x, (PD * (1 - TD))/TD + VARTD * ((PD*(1-TD)+ PD^2*TD))/TD^3)
> x
   WK       TD    VARTD  PD DailyVarPD   newDaily
1  28 0.014735 4.49e-05   0          0      0.000
2  28 0.012421 4.49e-05   0          0      0.000
3  28 0.014396 4.49e-05  71       7005   7006.197
4  28 0.014898 4.49e-05 132      14016  14018.783
5  28 0.013512 4.49e-05 223      32506  32514.484
6  28 0.013886 4.49e-05 224      31287  31295.424
7  28 0.013778 4.49e-05 218      30528  30535.647
8  29 0.013274 4.49e-05 228      34505  34514.170
9  29 0.013072 4.49e-05 224      34531  34539.992
10 29 0.012518 4.49e-05 488     117720 117762.642
11 29 0.013264 4.49e-05 525     119323 119365.443
12 29 0.013233 4.49e-05  80       9134   9136.090
13 29 0.010194 4.49e-05 417     133067 133116.531
14 29 0.012567 4.49e-05  82      10184  10186.465
15 30 0.014695 4.49e-05 413      68893  68915.217
16 30 0.010634 4.49e-05 373     103688 103724.478
17 30 0.013934 4.49e-05 914     272719 272829.966
18 30 0.014107 4.49e-05 213      28473  28480.576
19 30 0.014339 4.49e-05 651     147016 147070.746
20 30 0.014375 4.49e-05 521     102429 102464.631
21 30 0.016373 4.49e-05 979     229104 229195.063
> # not sure what you want for weekly, but we can take the averages of the values
> x.wk <- aggregate(x, list(x$WK), mean)
> x.wk$var <- with(x.wk, (PD * (1 - TD))/TD + VARTD * ((PD*(1-TD)+ PD^2*TD))/TD^3)
> x.wk
  Group.1 WK         TD    VARTD       PD DailyVarPD  newDaily       var
1      28 28 0.01394657 4.49e-05 124.0000   16477.43  16481.51  14340.27
2      29 29 0.01258886 4.49e-05 292.0000   65494.86  65517.33  53548.78
3      30 30 0.01406529 4.49e-05 580.5714  136046.00 136097.24 126432.61
>
>


On Feb 16, 2008 11:17 AM, Felipe Carrillo <[hidden email]> wrote:

> Jim:
> I actually would like to know how can I estimate Daily
> and Weekly Variance for PD using R and not Excel.I
> have read about R's variance function but I don't know
> how to apply it to PD in my dataset. I hope my
> question is clearer. Thanks
>
>
> > > Dear All:
> > > Using the Excel formula below I am able to
> > estimate
> > > daily variance(Daily VarPD)for PD,but I can't
> > figure
> > > out how to estimate the Weekly variance for PD.
> > Any
> > > hints?
> > >
> > > Daily VarPD = (PD * (1 - TD))/TD + VarTD *
> > > ((PD*(1-TD)+ PD^2*TD))/TD^3
> > >
> > >
> > > WK      TD                VARTD      PD   Daily
> > VarPD
> > > 28      0.014735        0.0000449       0       0
> > > 28      0.012421        0.0000449       0       0
> > > 28      0.014396        0.0000449       71
> > 7005
> > > 28      0.014898        0.0000449       132
> > 14016
> > > 28      0.013512        0.0000449       223
> > 32506
> > > 28      0.013886        0.0000449       224
> > 31287
> > > 28      0.013778        0.0000449       218
> > 30528
> > > 29      0.013274        0.0000449       228
> > 34505
> > > 29      0.013072        0.0000449       224
> > 34531
> > > 29      0.012518        0.0000449       488
> > 117720
> > > 29      0.013264        0.0000449       525
> > 119323
> > > 29      0.013233        0.0000449       80
> > 9134
> > > 29      0.010194        0.0000449       417
> > 133067
> > > 29      0.012567        0.0000449       82
> > 10184
> > > 30      0.014695        0.0000449       413
> > 68893
> > > 30      0.010634        0.0000449       373
> > 103688
> > > 30      0.013934        0.0000449       914
> > 272719
> > > 30      0.014107        0.0000449       213
> > 28473
> > > 30      0.014339        0.0000449       651
> > 147016
> > > 30      0.014375        0.0000449       521
> > 102429
> > > 30      0.016373        0.0000449       979
> > 229104
> > >
> > >
> > > Felipe D. Carrillo
> > >  Fishery Biologist
> > >  US Fish & Wildlife Service
> > >  California, USA
> > >
> > >
> > >
> > >
> >
> ____________________________________________________________________________________
> > > Be a better friend, newshound, and
> > >
> > > ______________________________________________
> > > [hidden email] mailing list
> > > https://stat.ethz.ch/mailman/listinfo/r-help
> > > PLEASE do read the posting guide
> > http://www.R-project.org/posting-guide.html
> > > and provide commented, minimal, self-contained,
> > reproducible code.
> > >
> >
> >
> > --
> > Jim Holtman
> > Cincinnati, OH
> > +1 513 646 9390
> >
> > What is the problem you are trying to solve?
> >
>
>
> Felipe D. Carrillo
>  Fishery Biologist
>  US Fish & Wildlife Service
>  California, USA
>
>
>
>      ____________________________________________________________________________________
> Never miss a thing.  Make Yahoo your home page.
> http://www.yahoo.com/r/hs
>



--
Jim Holtman
Cincinnati, OH
+1 513 646 9390

What is the problem you are trying to solve?

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.