#weighted-error

Boosting on 10 points uses four weak classifiers, misclassifying B: {1, 5, 6, 9}, D: {1, 4, 7, 8}, F: {2, 7}, I: {2, 3, 8, 10}. Weights start at $1/10$; each round picks the smallest weighted error (ties alphabetically) and then rescales the misclassified points to total weight $1/2$ and the rest to $1/2$. Round 1 picks F and round 2 picks B. What is the weighted error of the classifier picked in round 3? Round to 3 decimals.