make definition of k-neighborhood more rigorous
Opened this issue · 6 comments
in particular, we don't state how the
x can be any point in the feature space. the x_i are features of the data points in the data set. perhaps we should change the part before the | to x_i \in foo where foo is the feature-part of \D (I guess we do not have a symbol for that, yet)?
Ne, lass uns das so machen: Replace \xi by x^star (twice). That way it cannot be confused with an element from the training data set.
Perhaps it would also be better to not call the k-th nearest neighbour x^k because this is reserved for the feature vector in the k-th row of the training data set? We would need a different symbol in that case, perhaps just x_k? Could then be confused with the k-th column, I know :-)
@lisa-wm I set a milestone with due date for this so that we do this next week 🤝 (Not sure what happens after the due date, so I will set it on next Monday to investigate it :) )
well, I did not get any reminder, did you @lisa-wm ?
