Help: Index that Falls above the lower variation limit if more than 50% of the data points in the following 20 point window are also above?

Hello,

I would like ot replicate an anlysis in the paper where they defined the offset as

"the first datum point that falls above the lower variation limit if more than 50% of the data points in the following 20 point window are also above the lower variation limit"

I've created an array that indicates the all indices that are above the lower variation threshold

Does anyone have a good solution to find the which datum fits th above criteria?

Right now, I've simply used diff(IndexAboveThresh) just to indicate consecutive datapoints, but I'm having trouble finding a way to complete the code.

Answer 1

[Mohammad Sami]

You should create a logical array which is true for a point which is above the limit. Then we use a moving sum with a kf = 20 and kb = 0. We want a sum of 12 since sum is inclusive of current index plus next 20 points. So it would mean 11 points after this index satisfy the criteria which more then 50%. if you want more then or equal to 50% then we look for sum of 11.

% generate test data;
a = logical(randi([0 1],1000,1));
window_sum = movsum(a,[0 20]);
thresh = 12;
% a = true & window_sum >= thresh
firstdatumpoint = find(a & (window_sum >= thresh),1,'first');

firstdatumpoint = 7

[Dc215905]

Thank you! This works great!