CODEWITHSUNDEEP
javamatlabhashmap
bhavs
bhavs

Reputation: 2301

java to matlab conversion

I am a newbie at matlab . As a part of a larger problem, I need to find maximum number of occurrences of a string in an array of Strings.

Since I have some experience in java I have written the partial code in java ( only until the number of occurences of the string within the string array can be computed I can sort the hashmap depending on the values and extract this) 

int incr = 0;
    String[] c = { "c1", "c2", "c3", "c1", "c2", "c2", "c2","c1","c2" };
    Map<String, Integer> classRegistry = new HashMap<String, Integer>();
    for (int i = 0; i < c.length; i++) {
        String classes = c[i];
        if (!(classRegistry.containsKey(classes))) {
            for (int j = i + 1; j < c.length; j++) {
                if (classes.equals(c[j])) {
                    incr++;
                }
            }
            classRegistry.put(classes, incr+1);
            incr = 0;
        }
    }

Any idea how i can use something like a hashMap in MATLAB to calculate the number of occurrences of all the strings in an array

Thanks,

Bhavya

Upvotes: 1

Views: 2335

Answers (3)

Amro
Amro

Reputation: 124563

MATLAB has a function TABULATE available in the Statistics Toolbox:

c = {'c1' 'c2' 'c3' 'c1' 'c2' 'c2' 'c2' 'c1' 'c2'};
t = tabulate(c)
t = t(:,1:2)

The result:

t = 
    'c1'    [3]
    'c2'    [5]
    'c3'    [1]

Alternatively, you can do the same using the UNIQUE and ACCUMARRAY functions:

c = {'c1' 'c2' 'c3' 'c1' 'c2' 'c2' 'c2' 'c1' 'c2'};
[classes,~,subs] = unique(c);
counts = accumarray(subs(:),1);

Again the result as before:

>> t = [classes(:) num2cell(counts)]
t = 
    'c1'    [3]
    'c2'    [5]
    'c3'    [1]

Then to find the string that occurred the most, use:

>> [~,idx] = max(counts);
>> classes(idx)
ans = 
    'c2'

Upvotes: 2

Kevin K
Kevin K

Reputation: 9584

Since you have experience with Java, you could just write your code in Java and call it from MATLAB. This techdoc article should help you get started if you decide to go this route. But it would probably be more useful to learn how to do it in m-script (see jomb87's answer)

Incidentally, you could improve the performance of your Java algorithm if you take further advantage of the hash map:

    for (int i = 0; i < c.length; i++) {
        String classes = c[i];
        if (classRegistry.containsKey(classes)) {
            classRegistry.put(classes, classRegistry.get(classes) + 1);
        } else {
            classRegistry.put(classes, 1);
        }
    }

Upvotes: 0

jornb87
jornb87

Reputation: 1461

You didn't specify exaclty how you would want your input and output data types to be, but I wrote this quick script you might find usefull.

c = {'c1' 'c2' 'c3' 'c1' 'c2' 'c2' 'c2' 'c1' 'c2'};
count = struct();
for ic=1:length(c)
    field = c{ic};
    if isfield(count, field)
        count = setfield(count, field, getfield(count, field) + 1);
    else
        count = setfield(count, field, 1);
    end
end

The output for this specific c would be

count =

    c1: 3
    c2: 5
    c3: 1

Upvotes: 1

Related Questions