if else statement for tabular data classification

I have data in a table that I would like to classify using an 'if else' statement.

My data is the differences in frequency (total_changes) at intervals on a curve. The data is currently in a 20x10 table and I am now looking to classify these frequency changes into the following categories:

big_rise: >=3000

small_rise: >=500 & <3000

constant: <500 & >-500

small_fall: <=-500 & >-3000

big_fall: >=-3000

From looking online, I have found this suggestion. It has not been working for me and has not given any error messages.

if total_changes<3000
        fprintf('large_rise');
    elseif total_changes (total_changes>500 & total_changes<3000)
        fprintf('small_rise');
    elseif total_changes (total_changes<500 & total_changes>-500)
        fprintf('constant');
    elseif total_changes (total_changes<-500 & total_changes>-3000)
        fprintf('small_fall');
    else 
        fprintf('large_fall')
end

I am pretty new to MATLAB so apologies if this is rudimentary stuff, but I would be extremely grateful for some help with this!

If any more information/screenshots are needed, please let me know.

Thanks

Answer 1

[Steven Lord]

discretize your data. Let's start off with some sample data. I'm using rng default so you receive the same results I did when I ran this example, so I know all five of your categories are represented.

rng default
x = randi([-5000 5000], 10, 1);

Define the categories and the edges for each.

cats = {'big fall', 'small fall', 'constant', 'small rise', 'big rise'};
edges = [-Inf -3000 -500 500 3000 Inf];

The interval [-Inf, -3000) is the 'big fall' category.

The interval [-3000, -500) is the 'small fall' category.

Now let's discretize. Note that by default MATLAB includes the left edge but not the right in each bin (except for the last bin, which includes both edes.) If you wanted to change that you'd add the 'IncludedEdge' option. See the documentation page for details on that option.

d = discretize(x, edges, 'categorical', cats);

Now let's see into which category each element of the data was binned.

t = table(x, d, 'VariableNames', {'Value', 'category'})

[RufusS]

This worked great! Thank you!

Answer 2

[darova]

Try for loop

[RufusS]

Hiya,

So I am trying the for loop like you suggested and adding in the variable i. I end up with only a number of 'large_rise' as they all seem to be classified the same.

Surely the numerical parameters need to be set? E.g. big_rise: >=3000

This is the edited code I wrote following your suggestion:

for i = 1:length(total_changes)
    if total_changes(i)
        fprintf('large_rise');
    elseif total_changes (total_changes(i) & total_changes(i))
        fprintf('small_rise');
    elseif total_changes (total_changes(i) & total_changes(i))
        fprintf('constant');
    elseif total_changes (total_changes(i) & total_changes(i))
        fprintf('small_fall');
    else 
        fprintf('large_fall');
    end
end

Thanks so much for your help!

[darova]

Sorry to misunderstand you. I mean just adding (i), not replacing

for i = 1:length(total_changes)
    if total_changes(i)<3000
            fprintf('large_rise');
        elseif total_changes (total_changes(i)>500 & total_changes(i)<3000)
            fprintf('small_rise');
        elseif total_changes (total_changes(i)<500 & total_changes(i)>-500)
            fprintf('constant');
        elseif total_changes (total_changes(i)<-500 & total_changes(i)>-3000)
            fprintf('small_fall');
        else 
            fprintf('large_fall')
    end
end