How to to write a code to my n*1 matrix into differnt submatrices?

Hi guys, I have N*1 Index matrix with N rows. Where N varies for different files. My matrix consists N number of channels. I need to copy my channels into one group when difference between my channel is greater than one.I need to make differnt groups whenever differnce between the channels is greater than one. I tried to use Magic function and find function to save into different groups. When I use FIND function, I can find channels which difference are greater than one and less than one. I am unable to copy them into differnt groups.

Can anyone help me with this problem

Cheers,

Sudheer

Answer 1

[Matt Fig]

Can you give a small example matrix and what you expect the output to be?

It is still not clear what you want. I asked you to provide input and output. You provided input, but no output. How would you group these, in a cell array?

Here is a guess:

A = [164
165
166
167
175
176
177
189
190
195
196];

T = [1;diff(A)]==1;
t1 = [1 findstr((T).',[0 1])];
t2 = [findstr((T).',[1 0]) length(T)];

H = cell(1,length(t1));
for ii = 1:length(t1)
   H{ii} =  A(t1(ii):t2(ii));
end

R = num2cell(A(~ismembc(A,cat(1,H{:}))));
H(end+1:end+length(R)) = R

.

.

EDIT

A one-liner (FWIW):

groups = mat2cell(A,diff([0;find(diff(A) ~= 1);length(A)]),1);

[sudheer kumar reddy]

For instance, I have 20 channels, which are my 20 rows.

Channels

164

165

166

167

175

176

177

189

190

195

196

Now if you analyze difference between my first 4 channels is 1 and difference between 4 and 5 is greater than one and so on the cycle continues. When the difference is greater than one I want all previous channels into one group and next comes into different group and so on.

If you need more information please do not hesitate to reply back

Cheers

[sudheer kumar reddy]

Hi MattFig,

Thanks for your code. I would explain what I am doing in my project. The following values which I have give represents data points. Now when I get different groups of my data points I can calculate mean of my data points and plot them on my wave form. The only reason I need to make into different groups is, if I plot my data points I am plotting many points at a sudden change. Where I need only one point which specifies my peak. Please do not hesitate to ask any more information.

cheers

[sudheer kumar reddy]

HI Matt Fig,

groups = mat2cell(A,diff([0;find(diff(A) ~= 1);length(A)]),1);

The following function is straight forward and thats what I need to make groups. Thanks for that.

In the following function A is the indes of my data points and time. Data points is in voltage. Now I need to get data points of my A in groups and calcuate mean of my data points in my groups. I tried to call data(A), t(A), but I get an error message which doesnt accept data(A), t(A). Can you please suggest me with some solution. I can provide any more furter information if you need.

Cheers

[Paulo Silva]

This should work cellfun(@mean,groups)

[sudheer kumar reddy]

Hi Paulo Silva,

cellfun(@mean,groups)

with the following function I am able to find mean of my groups. But, I need to find mean of my data points which are represented by A. If I can get datapoints and time into groups I can use the following function and get find mean and plot.

I ll give an example,

I have got groups

group1= 162

163

164

165

group2= 170

171

172

173

these groups are index points of time and voltage. Now, I need to call my time and voltage into these groups and find maean and plot it.

Can you please give me futher advice

Cheers

[Paulo Silva]

time and voltage? are they stored in a cell or something? please give us one small example

[sudheer kumar reddy]

close all

[NUM,TXT,RAW]=xlsread('b3br.csv');

t=NUM(:,1);

data=NUM(:,2);

figure;

clf;

hold on;

plot(t,data);

figure(5); hold on;

Z1=smooth(t,'rlowess');

Z2=smooth(data,12,'rloess');

figure(5);

plot(Z1,Z2)

K = (diff(Z2)); %to calculate the difference in voltage

K1 = abs(diff(Z2));

figure; plot(K)

ind = find(K1>0.0275); %check whther the diff is >0.02

i=1;

j=1;

C1=diff(ind);

while(~isempty(ind))

if (K1(i)<0) || ((K1(i)>0) && K1(i+1)>0)

i = i+1;

else

break;

end

X = t(ind);

Y = data(ind);

figure(5);hold on;

plot(X,Y,'r*')

A=ind;

groups = mat2cell(A,diff([0;find(diff(A) ~= 1);length(A)]),1);

Z=data(groups);

S=t(groups);

end

I am unale to get Z and S by using mat2cell function. If I can get Z and S I can use cellfun function and fing mean of my groups and plot.

Please give me some advice.

Cheers

[Walter Roberson]

Z = cellfun(@(G) mean(data(G)), groups);

S = cellfun(@(G) mean(t(G)), groups);

[sudheer kumar reddy]

Thank you Walter Robertson, your tip helped me..

Cheers

Answer 2

[Jan]

It sounds like you need HIST or HISTC, especially the 2nd output argument.

[sudheer kumar reddy]

I went through HIST. But, this only helps me to find count of my array. Where as, I need to group my data points in different groups.

Can you please give me any further advice.

please do not hesitate to ask any more information.

Cheers

Answer 3

[Paulo Silva]

a=[164 165 166 167 175 176 177 189 190 195 196]';
b=diff(a); %find differences
idx=find(b>1); %find their indexes
idx=[idx;numel(a)]; %add the last value as index
cel=cell(1,numel(idx)); %make a cell to hold the groups
sv=1; %start value
for f=1:numel(idx)
cel{f}=a(sv:idx(f)); %take each part of a into a group
sv=idx(f)+1; %make the next start value 
end
bar(cellfun(@mean,cel))

[Matt Fig]

This is a better solution, only calling FIND once and preserving the order.

You could get sv directly by:

idx = find(diff(A)~=1);

idx = [idx;length(A)];

sv = [1; idx(1:end-1)+1];

+1 vote

[Paulo Silva]

Thanks Matt, your tips are always welcome :)

[sudheer kumar reddy]

Thanks for the code. In the following code I am not aware of how many a values I get. So, I am not sore about number of groups I get. If I get this I need to calculate mean between values in my group and plot.

Can you please help me with this. Please do not hesitate to ask any more information.

cheers

[Paulo Silva]

cellfun(@mean,cel)