论文修改建议 (HuangP 20211102 算法描述)

1. Safe Co-training for semi-supervised regression
   $\to$ 在前面第一次出现时说明了这个缩写，现在就正常写
   Safe co-training for semi-supervised regression

2. P, T 是整数，为标量
   $\to$
   用 $ 括起来，直接变成斜体

3. t <T
   $\to$
   t < T

4. j 统一为斜体

5. no availabel data
   $\to$
   no available data

6. while ($t<T$ || no availabel data) do
   $\to$ 根据自己的程序来反向修改伪代码
   while ($t<T$ && has available data) do

7. select a confident sample according to Eq. (2) by the (3−j)-th regressor and add it to the selected sample set;
   $\to$ 明明是两个步骤, 就写成两个句. 慎用 and.
   Select a confident sample x_i according to Eq. (2) by the (3−j)-th regressor;
   X’ = X’ \cup {x_i};

8. labeled data set {X_l ∪ X′(j)};
   $\to$
   集合取并之后，外面不要打花括号

9. Initialize \Theta = \theta(1), \theta(2)
   $\to$
   \Theta = \theta(1), \theta(2); // Initialize the cotrainers

10. selected sample set X′(j)
    $\to$
    X′(j) = \emptyset; // Selected sample set