how to predict future values?

1 回表示 (過去 30 日間)
navid zed
navid zed 2016 年 12 月 8 日
コメント済み: navid zed 2016 年 12 月 11 日
hi there,I have been working on my thesis that is related prediction the land surface temperature. I want predict it with neural network,thus I downloaded images of 10 years from 1985 to 2014. I set the LST(land surface temperature) as target and some indices as inputs. however I can train the network with neural network toolbox,I do not know how to predict the LST of future years,for instance 2017. I will be thankful,if you help me. thanks

採用された回答

Greg Heath
Greg Heath 2016 年 12 月 10 日
編集済み: Greg Heath 2016 年 12 月 10 日
If you need to predict 3 years ahead, then you should practice on data spaced 3 years apart.
The autocorrelation function of the target and the cross-correlation function of the target and input will yield the significant predictive spacings between the current output with past outputs and the current input.
There are many helpful posts and examples in the NEWSGROUP and ANSWERS. Try searching in both with
greg nncorr narxnet
Hope this helps.
Thank you for formally accepting my answer
Greg
  1 件のコメント
navid zed
navid zed 2016 年 12 月 11 日
hi Greg, thanks for answering me, I didn't understand what you meant. let me clarify my question. I have downloaded images of 10 years(1985,1990,1994,2000,2002,2005,2008,2010,2012,2014). I clipped images, retrieved LST and other indices and put it in input and output like this: inputs = [{[albedo85;NDVI85;IBI85;NDISI85]},{[albedo90;NDVI90;IBI90;NDISI90]},... {[albedo94;NDVI94;IBI94;NDISI94]},{[albedo2000;NDVI2000;IBI2000;NDISI2000]},... {[albedo2002;NDVI2002;IBI2002;NDISI2002]},{[albedo2005;NDVI2005;IBI2005;NDISI2005]},... {[albedo2008;NDVI2008;IBI2008;NDISI2008]},{[albedo2010;NDVI2010;IBI2010;NDISI2010]},... {[albedo2012;NDVI2012;IBI2012;NDISI2012]},{[albedo2014;NDVI2014;IBI2014;NDISI2014]}]; outputs = {values_NLST85,values_NLST90,values_NLST94,values_NLST2000,values_NLST2002,values_NLST2005... ,values_NLST2008,values_NLST2010,values_NLST2012,values_NLST2014}; I could train it,but I think I can't predict with this style. please help me, thanks

サインインしてコメントする。

その他の回答 (0 件)

カテゴリ

Help Center および File ExchangeSequence and Numeric Feature Data Workflows 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!

Translated by