Loop is running only for certain value in the whole range

For a data set of 6000 values,a loop is being run to calculate certain value but it is showing value for only 60 elements. The excerpt from code is given as :

t=0:0.005:29.995; %time duration
v=importdata('BAG_20051214_071014.txt'); %input data file
z=v.^2;
l=6001;
ns=1;
h=(1/ns)*(t>=0);
nl=60;
g=(1/nl)*(t>=0);
i = nl+1;
while i<=l
    for j = i-nl:i-1  % Loop to compute LTA
           x(j)=z(i-j);
            lta(j)=g(j).*x(j);
    end
    i=i+nl;
end

This part x(j)=z(i-j) is showing values only for 60 values but the total range of data is 6000. Kindly help.

[Bob Thompson]

Have you double checked that z is not empty except for those values?

Some side notes about your code:

i = nl+1;

This line is not necessary, as you immediately redefine i at the beginning of the for loop.

       i=i+1;

Manually advancing a loop index is considered bad practice, and generally doesn't work. I strongly advise against using this in you code.

       for j = i-nl:i-1  % Loop to compute LTA
           x(j)=z(i-j);
        lta(j) = g(j).*z(i-j); % Sum LTA window
       end

Although this loop adjusts the index back to cover a set of 60 values, 59 of those values will be overwritten for the next i value. What is the point of running this loop? As an example, the first value of i is 61, so this loop will cover j values from 1 to 60. The second i value is 62, which means that the loop will cover values from 2 to 61.

[gulu]

Yes i have checked the z values,they are correct. But this loop is running for 60 values,i want to calculate the same things for 6000 values keeping the j limits same. Is there any other possible way to calculate values for 6000 datas keeping j limits same?

[Image Analyst]

Looks like you forgot to attach 'BAG_20051214_071014.txt' with the paper clip icon, so I didn't bother delving into your code yet. It's so much faster to debug when you have the actual data than try try to guess without data.

[gulu]

The data file is attached here.

[KSSV]

What is t in your code?

Undefined function or variable 't'.

Error in Junk (line 8)
h= (1/ns)*(t>=0);

[gulu]

t is the time period.

[KSSV]

I would like to know what you are trying to do? If this is known code can be simplified.

[gulu]

I am trying to calculate this formula in MATLAB.

[Walter Roberson]

while i<=l
    for j = i-nl:i-1  % Loop to compute LTA
           x(j)=z(i-j);
    end
end

You do not modify i or l inside your while loop, so this is an infinite loop: if i<=l is true at the time the loop is encountered, the loop will run forever.

[gulu]

Sorry, I have updated that. Stil the same problem.

[Walter Roberson]

It looks to me as if you should be reshaping the z array to nl rows, and flipud() that, and assign that to x, possibly followed by a reshape to vector.

[per isakson]

z(j) = v(i)^2 in the formula is that a typo?

[Guillaume]

@gulu,

I'm not sure what you're hoping to achieve by editing your question away, other than making it less likely that you'll receive help in the future. It's something we really don't like. It's also a complete waste of time since it will be restored by Mathworks.

[Rik]

If you want to ask a new question, post it as a separate question, don't edit this one. Your last edit after Guillaume's comment seems like a valid question, but has nothing to do with the comments or the answer below.

[Rena Berman]

(Answers Dev) Restored edit

Answer 1

[per isakson]

I'm not a fan of backward engineering of faulty code. The formula of one of your comments is the formula of a moving average. Matlab has a moving average function, movmean. One could also use convolution or a filter.

Here is a code that somewhat adheres to your code. I might have missed the exact position of the window by one.

[x1,z] = cssm_();
whos  x1 z
figure; plot([x1',z'],'.')
x2 = movmean( z, [0,59], 'Endpoints','discard' );
figure; plot([x1(61:end)',x2(2:end)'],'.')
function    [ x ,z ] = cssm_()
    %#ok<*AGROW>
    v = importdata('BAG_20051214_071014.txt'); %input data file
    z = reshape( v.^2, 1,[] );
    len=length(v);
    x = nan( 1, len );
    nl = 60;
    g = ones( nl, 1 ) / nl;
    for jj = nl+1:len
        x(jj) = z( jj-nl+1 : jj ) * g; 
    end
end

In response to comment

I still don't understanding what you mean by "code for x2", "theoretical values" and "match". Anyhow, the following check shows that the values of x1 and x2 agree (except for one value at the edge).

>> d = x1(61:end)-x2(2:end);
>> figure; plot(d,'.')

[per isakson]

The summation is implemented by

z( jj-nl+1 : jj ) * g; 

Is this homework?

[per isakson]

What do you mean by "practical and theoretical values" ?

[per isakson]

Answer by editing in comments is not a good idea!

See the addendum to my answer.