Hi, I don’t understand why while using n_neighbor=1, our score isn’t 100% when using kNeighbor on our training data, since it should find the data point itself ?
It happened when we have duplicated samples with different classes. For instance, 2 persons answered the same way to the survey questions but one earns more than 50k$ and one earned below 50k$. Since our classifier can only predict a single class then it will be wrong for one of the two samples for sure.