Talking about mathematical codes

x =[0. 1 1 0.49 0.50 0.23 0.27 0.02-0.02 0.26-0.25 0.08 ...

0.52 0. 13 -0.0 1 0.52 0.57 0.0 1 0.32 -0. 15 0.45 0.07 0.66 ...

0.58 0.5 1 0.53 0.49 0.38 0.33 0.29 0.32 0.03 0.39];

Long format; % Set the calculation accuracy

If length (x (:,1)) =1%determines the input matrix, if it is not a one-dimensional column matrix, transpose it.

x = x’；

end

N = length (x); % take the sample size of input data.

z = 0；

For i= 1:n%, the cumulative value is calculated and assigned to the matrix be.

z=z+x(i，)；

be(i，)= z；

end

The original sequence is shifted in parallel for i=2:n%.

y(i- 1，)=x(i，)；

end

Calculate the data in the first column of data matrix B for i= 1:n- 1%.

c(i，)=-0.5*(be(i，)+be(i+ 1，)；

end

Calculate the data in the second column of data matrix B for j= 1:n- 1%.

e(j，)= 1；

end

The data matrix b is constructed for i= 1:n- 1%.

B(i， 1)=c(i，)；

B(i，2)=e(i，)；

end

alpha = inv(B ' * B)* B ' * y； % calculation parameter matrix

For i= 1:n+ 1% Calculate the cumulative series of estimated values of data. If it is predictable, change n+ 1 to n+m+ 1.

ago(i，:=(x( 1，:)-alpha(2，:)/alpha( 1，:)*exp(-alpha( 1，:)(i- 1))+alpha(2，:)/alpha( 1，:)；

end

var( 1，)=ago( 1，)

For i= 1:n% If n is changed to n+m- 1, the last m- 1 values can be predicted.

var(i+ 1，:)=ago(i+ 1，:)-ago(i，:)； % of the estimated value, and calculate the next prediction.

end

For i= 1:n

error(i，)=var(i，)-x(i，)； % calculation residual

end

C=std (error) /std(x)% Call the standard deviation function of the statistical toolbox to calculate the ratio of the variance after the test.