How does boosting actually work




I recently read the below piece while I was learning about ensembles/boosting:

From what I understand:
If there are 10 variables in the X’s,at first all the variables have equal weight but in consequent stages the same dataset has variables whose weights are assigned according to how they have performed in the previous stage.
So if the variable income(say) has given a wrong classification for the response Y it is given more weight .
But I am not sure that this is the correct interpretation,so can someone please help me with this??


That is right so with ten variables you start to allocate a weight of 1/10, and then based on the result you change it. And then on each round you change the weight based on the result (badly classified) output of the base learners.
If you are really interested to get a deep understanding of Boosting, try to get the book Boosting form R Shapire. Robert invented adaboost so the reference.