I have an array which is increasing. At one point it starts to decrease. This trend starts repeating many times in array. How to extract only the decreasing portion from that array?

I have an array which is increasing. At one point it starts to decrease. This trend starts repeating many times in array. How to extract only the decreasing portions from that array and plot another array values which corresponds to that decreasing portion?

Answer 1

[Walter Roberson]

d = diff( YourArray(:).' );    %I assume it is a vector of unknown orientation
goes_down = strfind([0 d>0], [1 0]) + 1;
stops_going_down = strfind([0 d<0], [1 0]);
n_seg = length(goes_down);
segs = cell(1, n_seg);
for K = 1 : n_seg;
  segs{K} = YourArray(goes_down(K) : stops_going_down(K));
end

[Image Analyst]

TESNA, please attach your data, or at least a screenshot of it. If it's noisy, you may have to apply denoising methods first to get results that are what you're hoping for.

[Tez]

Third column is the increasing and decreasing array from which decreasing values are to be extracted and I have to plot corresponding values of the second column..

[Walter Roberson]

It is not clear how you want the values to be plotted against each other.

fid = fopen('2variables.txt', 'rt');
YourTable = textscan(fid, '%s%f%f', 'Headerlines', 1);
fclose(fid);

Col2 = YourTable{2};
Col3 = YourTable{3};

d = diff( Col3 .' );    %I assume it is a vector of unknown orientation
mask = d < 0;
goes_down = strfind([0 mask], [0 1]) + 1;
stops_going_down = strfind([mask 0], [1 0]) + 1;
n_seg = length(goes_down);
idx2s = cell(1, n_seg);
seg2s = cell(1, n_seg);
seg3s = cell(1, n_seg);
for K = 1 : n_seg
  idx2s{K} = goes_down(K) : stops_going_down(K);
  seg2s{K} = Col3(idx2s{K});
  seg3s{K} = Col2(idx2s{K});
end

segs23 = [seg2s; seg3s];
plot(segs23{:});

[Tez]

Thank you for your help. It works.

[Walter Roberson]

Please mark the Answer as Accepted if this is what you were looking for.

... I still wonder whether perhaps you want to plot

segs32 = [seg3s; seg2s];
plot(segs32{:});

Hard to say... neither one is all that clear.

[Tez]

How can I apply this code in multiple files?

d = diff( Col3 .' );    %I assume it is a vector of unknown orientation
mask = d < 0;
goes_down = strfind([0 mask], [0 1]) + 1;
stops_going_down = strfind([mask 0], [1 0]) + 1;
n_seg = length(goes_down);
idx2s = cell(1, n_seg);
seg2s = cell(1, n_seg);
seg3s = cell(1, n_seg);
for K = 1 : n_seg
idx2s{K} = goes_down(K) : stops_going_down(K);
seg2s{K} = Col3(idx2s{K});
seg3s{K} = Col2(idx2s{K});
end
segs23 = [seg2s; seg3s];

[Walter Roberson]

When you process the multiple files, at each step do you want to forget about the previous file (except perhaps for the plot that has been drawn), or at each step do you want to add on to the seg2s / seg3s arrays so that you end up with one long cell array of those that includes all of the files, or do you want to use a cell array wrapping around each of those so that you can index into the cell array to get the seg2s / seg3s relevant to that file, or ....?

http://matlab.wikia.com/wiki/FAQ#How_can_I_process_a_sequence_of_files.3F