Normalize y axis in Matlab histogram

Hi,

suppose I have this histogram:

histogram([resultCastMatch.jaccard]);

Is this the correct way to normalize y axis (so to show y axix as percentage)?

histogram([resultCastMatch.jaccard]);
ylabels = get(gca, 'YTickLabel');
ylabels = linspace(0,100,length(ylabels));
set(gca,'YTickLabel',ylabels);

Answer 1

[Geoff Hayes]

Mons - no, I don't think that is the way to go about converting the y-axis to percentages. Remember, the ticks along the y-axis will just tell you the number of elements that fall into each of your bins. So if you want to convert these ticks (or heights) to percentages, you would have to divide each tick by the total number of elements across all bins. Or, depending upon your version of MATLAB, you may be able to specify the Normalization property to have a value of probability (which would produce the desired result without you having to manipulate the ticks). See http://www.mathworks.com/help/matlab/ref/histogram.html#namevaluepairarguments and look for the section on Normalization.

[Mons Nasery]

thanks for your useful link :)

[Steven Lord]

I second Geoff's suggestion of specifying the 'probability' value for the 'Normalization' argument.

Answer 2

[Star Strider]

Not quite. You have to specify the YTick values as well as the YTickLabel labels:

figure(1)
h1 = histogram(randi(9, 1, 25))
yt = get(gca, 'YTick');                     % Get Y-Tick Values
set(gca, 'YTick',yt, 'YTickLabel',yt*4)     % Since Data Vector Is (1x25), Multiply ‘yt’ By 4 To Get Percent
ylabel('Percent')

[Mons Nasery]

thanks for your kind answer. So, in my case which the y axis shows the number of users and I have 164 users, is it correct?

figure;
h1 = histogram([resultGenreMatch.jaccard]);
yt = get(gca, 'YTick');                    
set(gca, 'YTick',yt, 'YTickLabel',yt*(100/164));

[Star Strider]

My pleasure.

That looks correct to me. It should work.