11-07-2014 04:17 PM - edited 11-07-2014 04:19 PM
i have this array.
i want to delete random noises.
but i dont know where to start.
i am thinking that i can delete row3 thats equal = 0.
how can i go about doing that?
i want it to look like this
Solved! Go to Solution.
11-07-2014 04:37 PM
You will need to be more specific in your definitions of (1) how to detect the values elements containing your "random noises," and (2) which rows/columns to delete. You mention deleting row 3 but your results image shows columns 3-5 deleted.
Testing floating point values for equality with zero is risky because the actual value might be very small (~1E-15) but not exactly zero.
Lynn
11-07-2014 04:39 PM
i believe i need to look at columns 3-5 at row3 for 0. and if 0 exist. delete those columns?
11-07-2014 04:48 PM
Use Array Subset to get row 3, columns 3-5 in a subarray. Run that through In Range & Coerce with the limits +/- some small value that is larger than the maximum noise and smaller than the smallest useful signal. Check the In Range? output. True values will indicate the presence of "0." If you get a True, then use Array Subset to get all rows and columns 0-2.
Lynn
11-07-2014 04:58 PM
yeah. i know that deleting rows or data is bad but i have no idea what to do in this case.
i have a system with 6 TC channels
known variables are,
channel 2,3,4 gets HOTTER
channel 4,5,6 almost always stay the same?
i am only interested in analyzing channel 2,3,4
thats why i want to delete channel 4,5,6?
this is bad because its not always channel 2,3,4 that will get HOT..WHAT IF it was 2,5,6 thats hot?
can you sugguest a solution to his?
would i use RANGE & COERCE and look at the limits..if the limit changes then keep that channel? if not..delete it?
11-07-2014 05:27 PM
Now that is a different problem.
The data file you posted does not have any zeros in it. The maximum and minimum of the temperature data are 51.1 and 12.9.
I found the minimum and maximum for each column and then calculated the ratio. The "quiet" channels have ratios ~1.04 while the channels with the large temperature increases have ratios >1.7 There are also some channels with some noise (like the 60 channel in your image) with rations between those values. Perhaps you could use the max/min ratio to determine which channels you want.
Lynn
11-07-2014 05:42 PM - edited 11-07-2014 05:43 PM
lol..
thats a genius idea!!!
THANK YOU!!!