Algorithm 1: The learning process details of the RankBoot algorithm for approaching the final ranking H.

Require: Initial values of (D $c_0$, $c_1$) over each code example pair in the training data. 1: Initialize: $D_1(c_0,c_1)=D(c_0,c_1)$ 2: for t = 1, …, T (T = 12) do 3: Build ranking function f t (c) based on the ranking feature. 4: Choose ${\alpha }_t$ using Equation (8). 5: Update: $D_{t+1}(c_0,c_1)=D_t(c_0,c_1)exp\left(t(f_t\left(c_0\right)-f_t\left(c_1\right))\right)$ 6: end for 7: Output the final ranking $H\left(c\right)={\sum }_{t=1}^T{\alpha }_t{f}_t\left(c\right)$