Thank you for this excellent intro!

I would like to build my own local broker? Is there anything in the Paho API to do it?

Regards

Ayoub

Now that we have the data, we compare it with prior data of every character, and look at what’s the closest in similarity. We do this by calculating the standard deviation, which gives a measure of how much deviation, or dissimilarity is there between 2 terms. We calculate it for each of the 8 frequency bins for the 8 directions, comparing each with the corresponding bin of the pre-stored character.

For example, someone who draws a “1” may have data like this-> 0, 0, 0, 30, 23, 0, 0, 0. Since bin 3 and 4 would be filled.

We compare it with pre-existing data on what a “1” would look like, for example-> 0, 0, 0, 36, 20, 0, 0, 0.

We compare 0 with 0, 30 with 36, 23 with 20 and so on. Since these are quite similar, we’ll have a low standard deviation.