192.168.1.1 at 19:42, 4 October 2016

2016-10-04T19:42:19Z

← Older revision		Revision as of 15:42, 4 October 2016
Line 61:		Line 61:
	end		end

	[[Category: ~~Matlab~~ functions]]		[[Category: MATLAB functions]]

192.168.1.1 at 19:41, 4 October 2016

2016-10-04T19:41:45Z

← Older revision		Revision as of 15:41, 4 October 2016
Line 61:		Line 61:
	end		end

	[[Category: Matlab ~~Functions~~]]		[[Category: Matlab functions]]

192.168.1.1 at 19:41, 4 October 2016

2016-10-04T19:41:31Z

← Older revision		Revision as of 15:41, 4 October 2016
Line 60:		Line 60:
	end		end
	end		end

			[[Category: Matlab Functions]]

192.168.1.1: Created page with "A MATLAB function was created that subtracts the nonlinear trend, obtained from the global mean, from each of the time series vectors. In the course of doing so, the function..."

2016-10-04T19:40:15Z

Created page with "A MATLAB function was created that subtracts the nonlinear trend, obtained from the global mean, from each of the time series vectors. In the course of doing so, the function..."

New page

A MATLAB function was created that subtracts the nonlinear trend, obtained from the global mean, from each of the time series vectors. In the course of doing so, the function plots the original data, the global mean, and the transformed data (with the trend removed).

== NLDetrendFSTS.m ==
function dtM = NLDetrendFSTS(M, droprows)
%Removes non-linear trends from a cell array of matrices, M.
%Parameters:
% M: a cell array of matrices.
% drop: a vector of indices indicating the rows to be dropped before
% detrending (default: empty; all rows are kept). Positive values are
% relative to the start of each matrix and values less than 1 are
% relative to the end of each matrix (e.g., [-1 0 1 2 3 4] drops the
% first four rows, 1:4, and the last two rows, -1:0)
if isempty(droprows) || ~exist('droprows', 'var')
droprows=[];
end

nmatrices=length(M);
dtM=cell(nmatrices,1);

figure(1);

for i=1:nmatrices
A=M{i};
%drop rows as requred
A=dropRows(A,droprows);
%get global mean waveform
gmean=mean(A')';
%find polynomial line of best fit to mean waveform
polyord=6;
t=(1:length(gmean))';
[p,s,mu]=polyfit(t,gmean,polyord);
f_y=polyval(p,t,[],mu);
%subtract the line of best fit from the time series
f_y_mat=repmat(f_y,1,size(A,2));
dtA=A-f_y_mat;
subplot(nmatrices,3, ((i-1)*3)+1);
plot(A);
subplot(nmatrices,3, ((i-1)*3)+2);
plot(gmean);
subplot(nmatrices,3, ((i-1)*3)+3);
plot(dtA);
dtM{i}=dtA;
end

%% Nested utility function dropRows(M,droprows)
function M=dropRows(M,droprows)
%receives matrix M and drops rows indexed by positive values in
%droprows, and indexed from the end (e.g., M(end-1,:) for values
%less than 1
sz=size(M);
keeprows=1:size(M,1); %rows to keep
%convert negative droprows indices into end-relative values
droprows=[droprows(droprows>0) droprows(droprows<1)+size(M,1)];
keeprows(droprows)=[];%removes from rows to keep the set of rows to drop
M=M(keeprows, :); %keep only the rows not flagged to be dropped
%if M was 3D+, it will need to be reshaped
sz(1)=size(M,1); %new number of rows in sz matrix, all other dimensions maintained as they were
M=reshape(M, sz);
end
end

NLDetrendFSTS - Revision history

192.168.1.1 at 19:42, 4 October 2016

192.168.1.1 at 19:41, 4 October 2016

192.168.1.1 at 19:41, 4 October 2016

192.168.1.1: Created page with "A MATLAB function was created that subtracts the nonlinear trend, obtained from the global mean, from each of the time series vectors. In the course of doing so, the function..."