how to treat the last observation? Forecasting??
3 ビュー (過去 30 日間)
古いコメントを表示
Dear all,
I have the following cell matrix
A={
'24/09/2000' [4.1583]
'22/10/2000' [3.9389]
'19/11/2000' [4.3944]
'17/12/2000' [3.9313]
'14/01/2001' [3.9313]
'11/02/2001' [3.9313]
'11/03/2001' [4.0339]}
I can easily convert the above data points to monthly averages; through the calculation of the weighted average between successive observations.
For instance for the first observation I have
24*4.1583+ 6*3.9389
where 24 is the 24 days from month Octomber and 6=(last date of the month-24=30-24) The only problem is how to treat the last observation:
11*4.0339+?
One way is to extrapolate but I do not consider this method reliable. The above values are prices. So could I apply some other way? Maybe forecasting taking into account seasonality?
Any help/code provided will greatly appreciated
thanks
4 件のコメント
Walter Roberson
2012 年 8 月 18 日
You do not have enough data to forecast with, not with any meaningful accuracy.
採用された回答
Oleg Komarov
2012 年 8 月 18 日
A ={'24/09/2000' 4.1583
'22/10/2000' 3.9389
'19/11/2000' 4.3944
'17/12/2000' 3.9313
'14/01/2001' 3.9313
'11/02/2001' 3.9313
'11/03/2001' 4.0339};
% Convert to tractable double matrix, i.e. use serial dates
A = [datenum(A(:,1),'dd/mm/yyyy') cat(1,A{:,2})];
A =
730753 4.1583
730781 3.9389
730809 4.3944
730837 3.9313
730865 3.9313
730893 3.9313
730921 4.0339
% Create a list of end-of-month dates (using two tricks):
- create first-day-of-next-month dates and subtract one
- datenum() automatically adjusts the date in case one of the units exceeds the max accepted value
xi = datenum(2000,10:16,1)-1;
% Interpolate
B = interp1(A(:,1),A(:,2), xi,'linear','extrap');
% Plot
plot(A(:,1),A(:,2),'-dr',xi,B,'-ob')
set(gca,'Xtick',sort([A(:,1); xi']),'xgrid','on')
datetick('x','dd/mm','keepticks')
0 件のコメント
その他の回答 (0 件)
参考
カテゴリ
Help Center および File Exchange で Calendar についてさらに検索
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!