1) fix: `survived_score` getting wrong value. When the Tree is not dropped, the value got is `survived_score` and not th_val. Is not crictical, because generally the trees require many trees, so the score is got correctly.
2) Add quality checks in `utils.LoadDicts`.
3) Add input parameter `ignore_errors` in `utils.LoadDicts`.
4) Add iterable protocol in `utils.LoadDicts`.
5) Refactor some type hints.
6) Add missing values prediction feature.