Exact KNN

This backend uses sklearn's NearestNeighbors to perform exact nearest neighbour search. It is the simplest backend and requires scikit-learn.

---

When to use

• The validation set is small and speed is not a concern
• It performs best when exact results must be guaranteed and cannot afford any recall loss
• It performs worst on large datasets where the linear scan becomes a bottleneck

---

How it works

At fit time, sklearn's NearestNeighbors is fitted on the validation features. At query time, it performs a brute-force linear scan over all validation samples to find the exact k nearest neighbours.

---

Parameters

Parameter	Type	Default	Description
k	int	10	Number of neighbours

---

Dependencies

scikit-learn>=1.0

---

Example

from deskit.des.knorau import KNORAU

router = KNORAU(task="classification", metric="accuracy", mode="max", k=20, preset="exact")
router.fit(X_val, y_val, val_preds)
weights = router.predict(x)

---