# 討論區要求與模板 Discussion Forum Requirements and Template ###### tags: `course` --- You can edit the page simply by clicking on the top-right button <i class="fa fa-pencil"></i> and the top-left bottom <i class="fa fa-columns"></i>. (按右上方<i class="fa fa-pencil"></i> 之後再按左上方<i class="fa fa-columns"></i>,即可編輯本頁) --- ### 討論區要求 Requirments: * Please follow the following templates * Please do not insert photos * You are encouraged to answer other's questions * Try your best to use English in Hands-on sections ### 模板 Templates > 教科書是否能團購呢? >> 請同學自己揪團喔 >> [name=Joy][color=#42ffe8] > Can someone illustrate cross validation. We are discussing with our group. We think cross validation partitions the Z training set into n z training sets. Say for Decision trees, does it mean that we generate n trees and we pick the one with best accuracy on the test data? More insight and clarification about cross validation would be appreciated >> Let's use decision trees as an example, where n=5 folds. We would first train 5 separate decision trees, where each fold has a turn being the validation data and the rest of the data is the training set ... >> This video might help: >> [](https://)https://www.youtube.com/watch?v=OwPQHmiJURI. >> [name=TA][color=#42ffe8] >>