Considering I have a dataset of which a part is shown in the image below:
Basically it has some info about the football (soccer) matches from the past. Now based on the info let's say I want to predict a upcoming match. For example: for new data I have HomeTeam as Southampton and Away team as Newcastle and I want to predict FTR (Full time result) value which is either D A H (draw, away win and home win). Or may be all the labels except HomeTeam and AwayTeam. For a start we can just consider predicting FTR.
I could do train test split and train various classifier and check which one gives best result and choose best result. But then as far as what I have learned (Decision tree, SVM, Boosting, Random Forest) to predict new data I need to feed all the labels except FTR.
What should I do so that I can feed only few (2 labels in my case) and predict a label(FTR)? I don't think giving appropriate HomeTeam and AwayTeam value and setting other values as nan values would work.
Any suggestion how this kind of problems could be tackled.