I attepted to start off with a linear model and continuously adjust the model. That is start with a modle that guesses the same event for all inputs and continuously refine it. Event 14 was the most prominent event in the training data, so i used this as the starting linear model. However, I obtained a test score of 0. I moved on to the second most frequent event, yet still obtained the same result. I eventually ended up trying all 25 events with the same result.

This would imply that none of the events in the Training data appear in the Test data. This would be quite odd.

I had interpreted the target variable to be the 'event' colomn an this to be categorial in nature (although the categories are represented by numbers). Is this correct?

Is the above interpretation correct. I'm unsure what I am missing here.

Thanks.