How can I split an array into equal overlapping sub-arrays?

How can I split an array into equal overlapping sub-arrays?

Simple example where the array divides evenly without any need for overlapping

ArrayX = 1:160;
ArraySplitSize = 4

Result is

1:40, 40:80, 80:120, 120:160

Example where overlapping is needed. I want to keep full size of the array and not cut off anything at the end.

ArrayX = 1:100;
Subarray = 40;
ArraySplitSize = 3;
Result is 1:40, 30:70, 60:100

[Image Analyst]

Sounds like homework.

http://www.mathworks.com/matlabcentral/answers/8626-how-do-i-get-help-on-homework-questions-on-matlab-answers

[Kelly Kearney]

Also, what's the rule you used to decide which of the overlapping blocks get shorted when the array isn't evenly divisible? E.g. in your latter example, you indicate subarrays of length 40, 41, and 41. Also, in your first example, you listed values with overlaps; I assume that should be 1:40, 41:80, 81:120, and 121:160, right?

[mark palmer]

Right.

The subarrays don't have to be exactly equal-divided across the main array, as closly as possible is fine, but they all should be the same size (in this case 40).

It's not homework at all, I'm just not a pro programmer, am doing this for an art project.

Answer 1

[Walter Roberson]

If you have the Communications Toolbox, the easiest approach is to use buffer() and take the columns of the resulting matrix. buffer() already has logic for overlapping windows built in. Otherwise,

bufflen = length(buffer);
advance = window_size - overlap;
starts = 1 : advance : bufflen;
ends = min(starts + window_size - 1, bufflen);
result = arrayfun(@(S,E) buffer(S:E), starts, ends, 'uniform', 0);

This uses a strictly uniform size. The last few entries might be shorter.

[mark palmer]

Thanks, but how do I make the computer calculate the overlap? Plus, I want each subarray evenly spaced, and no truncation of the last subarray at the end.

[Walter Roberson]

bufflen = length(ArrayX);
starts = floor(linspace(1, bufflen-Subarray+1, ArraySplitSize));
ends = starts + Subarray - 1;
result = arrayfun(@(S,E) ArrayX(S:E), starts, ends, 'uniform', 0);

[mark palmer]

Works! Thank you so much!