How to filter out 1sigma data

Hi,

I have below data:

1.02 
1.06 
1.10 
1.06 
1.02 
1.05 
1.12 
1.50 
1.01

I want to filter-out the data which is >1sigma values (if 1sigma is standard in Matlab, using that function), if there is no such standard, then 1sigma is by 1sigma mean value.

my desired output:

1.02 
1.06 
1.10 
1.06 
1.02 
1.05 
1.01

Answer 1

[Star Strider]

I can’t produce your desired output, since that does not match the criterion of data being less than 1 standard deviation of the mean:

v = [1.02 
    1.06 
    1.10 
    1.06 
    1.02 
    1.05 
    1.12 
    1.50 
    1.01];

vm = mean(v);
vs = std(v);
v_result = v(v < vm+vs);

The mean+std value is 1.2573, so only 1.50 fails to meet the criterion.

[Mekala balaji]

if we just only based on 1sigma mean? does it make sense according to you?

[Mekala balaji]

I also want the row index of outliers

[Star Strider]

If you set 1 sigma (standard deviation) as the criterion, then none are acceptable, since the standard deviation is 0.1528. All the values are above that.

The row index of the outliers is easy:

outlier_idx = find(v >= vm+vs);