Nearest Trajectory Strategy for Time Series Prediction about timeseriesprediction.jl HOT 2 OPEN

This has little to do with the interface. Your suggestion disregards the inner workings of the search algorithm.

The only way to incrementally add more nearest neighbors would be to restart
the search with a larger k and consider the additional neighbors found.

It is likely much faster to just use a large k from the beginning so that researching becomes very rare.

from timeseriesprediction.jl.

I do believe that it is not hard to implement a new neighborhood type that does this. It will be slow of course but it can work by finding some "test" neighbors given a point, and then for each of the neighbors check if they are close to each other in data indices. If yes then add more nearest neighbors incrementally.

Maybe the interface for this in NearestNeighbors is really poor though...

from timeseriesprediction.jl.