Forecasting with AR(1) model and fixed parameters

2 次查看（过去 30 天）
Panty
Panty 2014-9-12
编辑： Panty 2014-9-12
Hi guys. Allow me to share my issue with you. I have a variable with 150 observations and I want to generate forecasts with an AR(1) model. Also what I want is to estimate the AR(1) model first, using the initial 57 observations and then keeping my estimated parameters fixed, I want a loop generating 4-periods ahead forecasts by adding an extra observation every time. So estimate the model just one time, keep the estimates, and by expanding my sample by 1 obs each time I want to generate my forecasts.
Here I give you the script I wrote for estimating the AR(1) model. What I miss is the forecasting part.
------------------------------------------------
% load data
data=MatlabDataRealtimeandRevisedS1;
% construct complete Variables ,x1 is the lagged(-1).
y1=data(2:end,1);
x1=data(1:(end-1),1);
[Traw,col]=size(y1);
T_full=T_raw;
% data for initial estimation
T_estimation=57;
x2=x1(1:T_estimation,1);
Y=y1(1:T_estimation,1);
% Build regressors matrix and ones for the constant
X=[ones(length(x2),1),x2];
% Doing OLS
[Coef,CoefConfidInter,e]=regress(Y,X);
------------------------------------------------
Any help please??
Please let me know if you got question. Much appreciated

请先登录，再回答此问题。

回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Model Predictive Control Toolbox 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by