finding sub sequences of a data set

clc
clear
ts = xlsread('SwedishLeaf_TRAIN.xlsx');
l=length(ts);
 for i = 1:500
     p=ts(i,:)
          for ii = 1:129
    subSequence(ii,:) = p((ii-1)*30+ 1:ii*30);
end

I used the above code for finding the subsequences,but it not working.How can I find the subsequence of this.

[Star Strider]

How do you define ‘subsequence’?

[Silpa K]

Iam trying to divide the sequence into subsequence,for that Iam trying to split 30 points each.

[Star Strider]

You need to describe in detail what your data are, and what you want to do.

[Silpa K]

I attached data set.I need to split each row.Each 30 points I need to separate from each row.

Answer 1

[Ajay Pattassery]

I assume you want to extract a sub-vector of length 30 from each row. But the total length of the column is 129 which is not divisible by 30. Hence the 5th sub-vector will take 21 elements from the second row and it goes on.

subsequence = (reshape(ts',30,[]))'

If you have a constraint to keep the column length 129 and wish to have sub-parts elements from the same row together, you could split the matrix into two parts. One exactly divisible by 30.

ts1 = ts(:,1:120);
Subsequence1 = (reshape(ts1',30,[]))';
Subsequence2 = ts(:,121:end);

Here Subsequence1 will be 2000x30 and Subsequence2 will have the remaining elements of each row.

[Silpa K]

It is not working in my code section.I need sequences ,every sequence contain 30 array points the remaining one is considering as next sequence.

clc

clear

b=zeros(175,1);

ts = xlsread('SwedishLeaf_TRAIN'.xlsx');

l=length(ts);

for i = 1:550

p=ts(i,:);

subSequence{1} = p(1:30);

subSequence{2} = p(31:60);

subSequence{3} = p(61:90);

subSequence{4} = p(91:120);

subSequence{5} = p(121:150);

subSequence{6} = p(151:180);

subSequence{7} = p(181:220);

subSequence{8} = p(220:250);

subSequence{9} = p(251:300);

subSequence{10} = p(301:end);

end

I tried this method for finding subsequence,but it is manually.How can I find subsequences.

[Ajay Pattassery]

please post the error message you are getting when using

subsequence = (reshape(ts',30,[]))'

Note, the transpose operator " ' " used in the above code.

Also post the output of

size(ts)

[Silpa K]

If it is not divisible by 30,then how can I do this.

[Silpa K]

Using the above Iam getting errors like not divisible by 30 in some datasets.

[Image Analyst]

So what would you LIKE to do when you have only a partial set (not the full 30)?

[Silpa K]

I need all subsequences,If a dataset row is not divisible by 30 ,then I need to take the remaining set also(every sequences)I need to apply this to each row.

[Ajay Pattassery]

1.So say your row comprises of 40 numbers ([1 2 . . . 40]), you want two subsequence.

Seq1 = [1 2 . . . 30] and Seq2 = [31 32 . . . 40], am I right?

2. You can not store the subsequence in a matrix as you were doing above since the subsequece length is not fixed. If the order of the subsequence does not matters you could try my second suggestion in the above answer. (Split the matrix into two, one with number of coulumns equals to mulitple of 30 and second matrix with remaining elements)

[Walter Roberson]

buffer() from the signal processing toolbox is useful. But otherwise mat2cell()

L = length(Ts1);
BL = 30;
if mod(L, BL)
    Sequences = cell2mat(Ts1, 1, [BL * ones(1, floor(L/BL)), mod(L,BL)]);
else
    Sequences = cell2mat(Ts1, 1, BL * ones(1, floor(L/BL)));
end

[Silpa K]

I used this in below code,Iam getting errors.Please help me

ts = xlsread('ArrowHead_TRAIN.xlsx');

l=length(ts);

for i = 1:24

p=ts(i,:)

fa = movstd(p,20,1);

secarray=movstd(fa,20,1) ;

k=maxk(secarray,10);

mpt=find(p);

mp=p(mpt(round(numel(mpt)/2)));

G=min(abs(mp-k));

[~,ii] = min(abs(p(:) - k(:)'));

out = p(unique(ii));

L = length(p);

BL = 30;

if mod(L, BL)

Sequences = cell2mat(p, 1, [BL * ones(1, floor(L/BL)), mod(L,BL)]);

else

Sequences = cell2mat(p, 1, BL * ones(1, floor(L/BL)));

end

A = [];

A = [];

for ii = 1:length(SubSequences)

if any(ismember(Subsequences,out))

if (k-mp<=G+l/2)

A{end+1} = Subsequences;

end

end

end

disp(aa);

Z=Subsequences;

idx = p(1:1);

q=[idx Z];

data = q;

cellReference = sprintf('A%d', i);

xlswrite('swetrain.xlsx', data, 1, cellReference);

end