Separating data to 3 different groups according to specific range of number in a variable

Hello, I have a 32x1875 sized variable named SNR_GPS, each rows represent the number of signals to noise ratio (SNR) of each satellite in the 1875 seconds. The first thing I would like is to calculate the mean SNR of each satellite which I successfully done do. But when I would like to arrange the satellites into 3 separate variables according to the requirements ( First group: SNR < 25, Second group: SNR 25-35, Third group: SNR>35 ), my old code failed to compile that. Is there any method to complete that? Thanks for your help.

Answer 1

[Dyuman Joshi]

Note - NaN values are not included in the grouping

load('SNR_GPS.mat')
data = SNR_GPS(~isnan(SNR_GPS))

data = 20625x1

0 22 23 39 0 31 40 41 0 32

k = discretize(data, [-Inf 25 36 Inf])

k = 20625x1

1 1 1 3 1 2 3 3 1 2

out = accumarray(k, data, [], @(x) {x})

out = 3x1 cell array
    {7037x1 double}
    {5184x1 double}
    {8404x1 double}

%checking the data
z = out{2}

z = 5184x1

31 32 32 32 32 30 34 30 34 30

min(z)

ans = 25

max(z)

ans = 35

[Yat Chi]

Thanks, but what I would like is to rearrange the whole row of data into new groups. For example in picture SNR_GPS_Mean the 4th row is 16.2891, so copy the whole 4throw in SNR_GPS to first group. And the 8th row of SNR_GPS_Mean is 27.5701, so so copy the whole 8th row in SNR_GPS to second group something like that.

The code for SNR_GPS_Mean is here:

SNR_GPS_Mean = mean(SNR_GPS,2);

[Dyuman Joshi]

Okay, the partition is to be done based on mean values -

load('SNR_GPS.mat')
data = mean(SNR_GPS, 2);
%removing NaN data
data = rmmissing(data)

data = 11×1

0.1419 16.2891 27.5701 33.5483 1.2693 34.6277 39.4309 38.3797 0.6048 37.8459

k = discretize(data, [-Inf 25 36 Inf]);
out = accumarray(k, data, [], @(x) {x.'})

out = 3×1 cell array
    {[   0.1419 16.2891 1.2693 0.6048]}
    {[        27.5701 33.5483 34.6277]}
    {[39.4309 38.3797 37.8459 38.7296]}

[Yat Chi]

Thanks for the solution, worked well for me

[Dyuman Joshi]

Glad to have helped!

[Yat Chi]

Sorry to bother you but one more question, I tried to use your method of using accumarray to rearrange data into 3 groups. So I generated a new variable SNR_GPS_SD_2 having the same size as the SNR_GPS_Mean_2. I checked that the accumulating elements of SNR_GPS_Mean_3 is same as the SNR_GPS_SD_2. But when I run the file, it shows the error of using accumarray, stating Second input must be a vector with one element for each row in the first input, or a scalar. What is wrong behind that and how can I solve that? Thanks again for your help.

This is the original code:

SNR_GPS_Mean_4 = accumarray(SNR_GPS_Mean_3, SNR_GPS_Mean_2, [], @(x) {x.'});

This is the new code:

SNR_GPS_SD_3 = accumarray(SNR_GPS_Mean_3, SNR_GPS_SD_2, [], @(x) {x.'});

[Dyuman Joshi]

Did you remove the NaN values (for the new data) as I did, using rmmissing ?

[Yat Chi]

Nevermind it worked again. Originally i used the rmmissing but still got the problem, maybe there is conflict in matlab code. Sorry for bothering you long and thanks very much for your help today!

[Dyuman Joshi]

No problem, I'm always happy to help :) You're welcome!