thatipamula-jashwanth/smart-knn
smartKNN - A feature-weighted KNN algorithm with automatic preprocessing, normalization, and learned feature importance.
Implements learned feature importance via MSE relevance, mutual information, or random forest metrics to automatically suppress noisy dimensions. Supports both brute-force and approximate nearest-neighbor (ANN) backends with optional GPU acceleration, exposing a scikit-learn–compatible API for regression and classification tasks. Built on vectorized NumPy with Numba JIT compilation for inference latency optimization across variable dataset scales.
Available on PyPI.
Stars
31
Forks
—
Language
Python
License
MIT
Category
Last pushed
Mar 11, 2026
Monthly downloads
203
Commits (30d)
0
Dependencies
5
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/thatipamula-jashwanth/smart-knn"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related frameworks
stabgan/K-Nearest-Neighbours
I applied the K-NN model for my classification problem in both R and Python
ksdkamesh99/KNN-Visualiser
It is a best Visualiser for implementing K-Nearest Neighbours Algorithm with 3 different classes...
lettier/interactiveknn
Interactive K-Nearest Neighbors machine learning algorithm in JavaScript.
ullayne02/Natural-Neighbor
Knn implementation without K parameter
dstarrago/fuzzy_rough_classifiers
Framework for developing fuzzy and fuzzy-rough classification models for multiple instance data