I’m a teacher and I’m working on a pet project to help streamline some of my assessment workflows for my students. One of those workflows is gathering data on student progress in the form of a rubric like the one below:

The rows are particular outcomes we are covering (in this case, reading clocks and working with units of time), and the columns are the kids of questions/tasks that student is able to complete (easy, intermediate, and challenging tasks). Throughout a particular unit, I mark down every time a student attempts a question/task and how they did (a checkmark if they got it right, G if they got it right with a group, A if they needed help, X if they got it wrong, etc). At the end of the unit, I look at all the rows and select the highest level of question they were able to do, and that translates to their grade. In other words, for each row I select a column based on the data in each cell. Data later in the sequence has higher priority, so a bunch of incorrect answers early on do not necessarily outweigh correct answers later in the unit.

I want to using some kind of ML model to predict which column would be selected for each row based on the data present in each column. The rows are evaluated independently from each other. I am using Swift to develop this app on iOS and macOS, but I am very new to the ML world. I wasn’t able to find a way to get Create ML to do what I wanted, but any ideas to point me in the right direction would be much appreciated! I’m not married to Swift, so if I need to use python or another language to create the model that’s fine, as long as it can be integrated into the swift app. My training data is a set of these rows from many rubrics, with each cell from the row having a letter to correlate with the symbols I use on the rubric, as well as a column for what the correct cell would be based on the data.