How can i process the elements of a vector in this way

Hello everyone

I have this vector

v=[-0.5 -0.3 0 1.2 1.3 1.7 -0.4 -0.6 0 0.9 0.2 0.5 -1.2 -2.3 3.2 0 -1.0 0 -5.5 -5.2 -6.4]

i want to divide the elements of this vector both between positive and negative (I want to neglect the zeros of the vector) and according to how many times they appear consecutively (that is according to the length of the sequences they appear). I mean, if i consider the negative values, i want to obtain

value_negative(1)=[-1.0]
value_negative(2)=[-0.5 -0.3; -0.4 -0.6; -1.2 -2.3]
value_negative(3)=[-5.5 -5.2 -6.4]

With "value_negative(n)" i indicate the negative elements of the vector that appear in a sequence of length n. If a sequence of negative values appear more that once, i want to obtain a matrix like in the case of value_negative(2).

Could you help me?

Thank you

Answer 1

[Rik]

A good first start would be to split your vector into a cell array. I suspect you will need diff, find, sign, and mat2cell.

Then you can use cellfun to find the number of elements in each section. You can decide to remove the zeros at this stage, or to remove them before you convert to a cell array. I would suggest removing them afterwards to keep your code more understandable.

Once you have the length, you can merge them with a loop and cat.

Let me know if you have any issues in implementing this.

[salvatore vergine]

Thank you for your answer. I tried but i don't get the result :(

[Rik]

What did you try? Did you read the documentation for the functions I suggested? What errors are you getting?

[salvatore vergine]

I read all the documentation of the functions but I am not so good in matlab so I don't understand the logical process to follow. I have never used most of these functions.

[Rik]

Let's try the splitting first. The general idea is to find the run length of each type (pos/neg/0). That way we can use that in mat2cell to split the vector into a cell array.

The first step is to convert your array to the same values, so all negative numbers are the same, all positive numbers are the same, and all zeros are the same. We can use the sign function to do this.

Next, we need to find all positions where type (pos/neg/0) changes. We can use the diff function. (Note that this will reduce the length of the vector by 1, so we will need to be wary of that.) For every position where the type changes the result will be non-zero.

That leads us to the next step: find will return the indices of the positions with non-zero elements.

Those indices are very helpful, but don't yet tell us the length of each section. Which function do you think we need again?

Once we have the length of each section, we can use that and the original vector as inputs to the mat2cell function.

Can you try implementing these steps?