Calculating the dispersion in C - c

Question

I'm trying to calculate the dispersion between two values, one I get from and certain parameter('age') from each entry on the array of structures.
I've got an auxiliary .txt file and, the whole point of this is to go through each of the lines of the .txt file individually (each line is a string). It's better if I state an example, so, here we go:
matrix[n][n]:
1 2 3 4
1 2 3
1 2
So, the structure would look a bit like this:
struct person {
int id; //where im comparing between the matrix and the structure;
int age; //what I need to calculate the dispersion
}
I gotta compare the each value of each line of the .txt, and if it matches any of the id's on the structure, I gotta get its age. Now comes the tricky part.
To calculate the dispersion I need to get the following working out for me:
Let's take as example the first row of the .txt file: the dispersion would be:
let's say that 'age' = age of the id (n);
//in this case nGroups = 6 (number of age(n)-age(m) operations)
dispersion(first_row)= [ [|age(1)-age(2)|] + [|age(1)-age(3)|] + [|age(1)-age(4)|] + [|age(2)-age(3)|] + [|age(2)-age(4)|] + [|age(3)-age(4)|] ]/nGroups
So I have to do this for each row of the matrix. I've tried it and managed the following code, but in the 'math' part my brain freezes a bit.
// going through each line of the matrix
for(i=0; i<nconjuntos; i++) {
// going through each column of the matrix
for(j=0; j<strlen(aux[i]); j++) {
// going through all the structures in the structure array
for(k=0; k<TOTAL; k++) {
// comparing each number with it's id equivalent in
// each structure.id parameter in the array
if(aux[i][j] - 48 == pessoas[k].id) {
}
}
}
}
Any help that would help me to advance in my code would be really appreciated!

Answer 1

You have two problems here requiring a solution:
Match the id in the matrix with the id in the structure and get a list of ages for a given row
Calculate the dispersion
It's probably easiest if you separate these two tasks instead of doing them both inside the same nested for loop. You can write a function to calculate the dispersion according to the formula you gave like this:
double calculateDispersion(double* values, int count)
{
if(count < 2)
return 0.0; // need at least 2 values
double total = 0.0;
// iterate through each combination of values and sum the absolute
// differences of each combination:
int i, j;
for(i = 0; i < (count - 1); i++)
{
for(j = i + 1; j < count; j++)
{
double difference = values[i] - values[j];
if(difference < 0.0) // find absolute value
difference = 0.0 - difference;
total += difference; // add to the total
}
}
int groups = (count * (count - 1)) / 2;
return total / (double)groups;
}
Inside your main loop create an array of doubles equal in size to the largest row in your matrix, to store the ages for the current row (or floats, but using ints will give you rounding errors). Fill it up by cross-referencing with the structs containing the id and age. Then call calculateDispersion() to calculate the dispersion for the row.
Something like roughly this:
double rowArray[MAX_SIZE]; // or initialize to maximum matrix row length dynamically
// going through each line of the matrix
for(i=0; i<nconjuntos; i++) {
int count = 0;
// going through each column of the matrix
for(j=0; j<strlen(aux[i]); j++) {
// going through all the structures in the structure array
for(k=0; k<TOTAL; k++) {
// comparing each number with it's id equivalent in
// each structure.id parameter in the array
if(aux[i][j] - 48 == pessoas[k].id) {
// set array value:
rowArray[count++] = (double)pessoas[k].age;
break; // break out of the loop
}
}
}
// compute the dispersion for the row:
doubleDispersion = calculateDispersion(rowArray, count);
}