MATLAB: How to perform row operations on a large MATRIX or CELL efficiently

celllarge cellsmatrixmemory management

I have a large CELL of the order 18000×4 (Or a MATRIX of the similar order) Each CELL element is a 'name' and therefore a string (In case of Matrix it is a 'number') The operation I need to do is explained in the following example (Please find relevant files attached). Example:
CELL INPUT
'1' '3' '4' []
'1' '3' '5' []
'1' '3' '6' []
'1' '4' '4' []
'1' '4' '5' []
'1' '4' '6' []
'2' '3' '4' []
'2' '3' '5' []
'2' '3' '6' []
'2' '4' '4' []
'2' '4' '5' []
'2' '4' '6' []
'5' '4' [] []
'5' '7' [] []
'6' '8' '4' []
'6' '8' '7' []
'9' '10' [] []
CELL OUTPUT
'1' '4' [] []
'1' '3' '5' []
'1' '3' '6' []
'2' '4' [] []
'2' '3' '5' []
'2' '3' '6' []
'4' '5' [] []
'5' '7' [] []
'4' '6' '8' []
'6' '7' '8' []
'10' '9' [] []
There are 2 main operations done to get the output:
1) Delete the duplicates in the rows. (1 4 4) —> (1 4) and (2 4 4) —> (2 4) in the above example.
2) After step (1), compare every row and delete the supersets of the row if any. (1 3 4) (1 4 5) (1 4 6) deleted because they are supersets of (1 4). Similarly, (2 3 4) (2 4 5) (2 4 6) are deleted because they are supersets of (2 4)
I have written a MATLAB function called 'minimize' which does this operation. But, when I need to do this for a large CELL of order 8656×4, MATLAB keeps running and there is no solution after 48hours. I need to get this done for even larger orders of the CELL and within a time frame of 1-2 hours.
Could you please provide any method/suggestion to efficiently solve this?

Best Answer

Hello All,
Thank you for your time and effort to help me solve this problem. I have figured it out to solve the 'minimize' more efficiently. This function is part of a bigger program I am working on, and was the bottle-neck. There is a pattern in which the input to 'minimize' is generated. I track the duplicates in while generating the input and only compare those rows and delete the super-sets among them. This reduces the computations from thousands to hundreds and sometimes even to tens.
Carl, Thanks for the suggestion of vectorization. I used it in many instances.
Sharath, Thank you for pointing out the flaws in my Algorithm. It helped me to think in different ways to reduce the number of computations.
With all the optimization done, I am able to get my output using input of the order 8656x4 in 20 seconds.
Best Regards, Ashish
Related Question