Are decision trees able to be used with time-related data?
I've read that decision trees are based on matrices and that ARRAYS of input matrices can be used to factor in time however I can't find an example of this.
Say for example, I'm monitoring the progress of students taking exams. Each day I ask them questions related to their mental state (fatigued, positivity, ability to concentrate, expectations for coming exam, confidence, etc). I have twenty days worth of questions. Day 1 for student A may see them studying for an exam the following day, while Day 1 for student B may see them actually doing the exam. There will be a relation between student's fatigue (for example) and the results they give the following day.
The examples when provided as input to a matrix will be used to show that IF on any given day, the student has an exam, and has breakfast, and does x,y,z THAT day then the outcome will be y.
However, short of encoding "had exam previous day" and "had exam two days ago" for each day, I can't see how I can include time dependency in decision trees.