Error estimating VaR, riskmetrics function?

Question

Mira 2015-8-6

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/232691-error-estimating-var-riskmetrics-function

回答： Mira 2015-8-6

*******I'm trying to estimate VaR with riskmetrics function. Could anyone please help. My code:

load SP500;
SPX2r_star=SPX2r(SPX2r~=0);
SPX2rv=SPX2rv(SPX2r~=0);
DateID=DateID(SPX2r~=0);
R=1000; 
load GJR_estimates_rolling_SP500
RV=reshape(SPX2rv,1,1,rows(SPX2rv));
RM_t = squeeze(riskmetrics(RV,0.94));

******I get the following error:

Undefined function or variable "T".

Error in riskmetrics (line 59)

endPoint = max(min(floor(log(.01)/log(lambda)),T),k);

Error in VaR_estimation_TV (line 48)

RM_t = squeeze(riskmetrics(RV,0.94));

*** *****riskmetrics function:

function Ht = riskmetrics(data,lambda,backCast)
switch nargin
    case 2
        backCast = [];
    case 3
        % nothing
    otherwise
        error('2 or 3 inputs required.')
end
if ndims(data)==2
    [T,k] = size(data);
    temp = zeros(k,k,T);
    for t=1:T
        temp(:,:,t) = data(t,:)'*data(t,:);
    end
    data = temp;
end
if lambda<=0 || lambda>=1
    error('LAMBDA must be between 0 and 1.')
end
if isempty(backCast)
    endPoint = max(min(floor(log(.01)/log(lambda)),T),k);
    weights = (1-lambda).*lambda.^(0:endPoint-1);
    weights = weights/sum(weights);
    backCast = zeros(k);
    for i=1:endPoint
        backCast = backCast + weights(i)*data(:,:,i);
    end
end
backCast = (backCast+backCast)/2;
if min(eig(backCast))<0
    error('BACKCAST must be positive semidefinite if provided.')
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Ht = zeros(k,k,T);
Ht(:,:,1) = backCast;
for i=2:T
    Ht(:,:,i) = (1-lambda)*data(:,:,i-1) + lambda * Ht(:,:,i-1);
end