diff not working on a vector of values

Is there a reason why I cant calculate the difference between each of these values (x1)

This is the code Im using

         [x1,y1]=SpotFind_NXC(app,ax2);   %my function to locate x,y coordinates of spots in an image    
         x1
         class(x1)
         xdiff=abs(diff(x1))

and the command window:

x1 =

         98.00
         97.68
         97.75
         98.32
        221.00
        220.99
        221.32
        221.32
        344.00
        343.99
        344.33
        344.25


ans =

    'double'

Unrecognized function or variable 'diff'.

2nd Question, once it works, how can I group these into "similar values". so ideally I would want the median (or mean) of the 1st group (98.00, 97.68. 97.75, 98.32), and then the median of the 2s group around 221 and the 3rd group around 344.

Thanks

Jason

Answer 1

[Cris LaPierre]

There is nothing about the code you have shared that would prohibit you from using diff. For some reason, it is not on your MATLAB path. You may need to reset your path using restoredefaultpath

x1 = [98.00 97.68 97.75 98.32 221.00 220.99 221.32 221.32 344.00 343.99 344.33 344.25]';
xdiff=abs(diff(x1))

xdiff = 11x1

0.3200 0.0700 0.5700 122.6800 0.0100 0.3300 0 122.6800 0.0100 0.3400

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

Not sure what typical values will be. You could use discretize or histcounts to bin your data. Here's one approach assuming your data is sorted.

% use xdiff to determine 
nbins = sum(xdiff>mean(xdiff))+1

nbins = 3

[N,edges,bin] = histcounts(x1,nbins)

N = 1x3

4 4 4

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

edges = 1x4

80 170 260 350

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

bin = 12x1

1 1 1 1 2 2 2 2 3 3

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

splitapply(@mean,x1,bin)

ans = 3x1

97.9375 221.1575 344.1425

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

[Jason]

Thankyou for the answer.

Answer 2

[Matt J]

x=[         98.00
         97.68
         97.75
         98.32
        221.00
        220.99
        221.32
        221.32
        344.00
        343.99
        344.33
        344.25];

G=findgroups(round(x,-1));

Medians=splitapply(@median,x,G)

Medians = 3x1

97.8750 221.1600 344.1250

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

[Jason]

Thankyou Matt, I love this findgroups approach. Is it possible to define a value that defines whether its ina group or not

Thanks