Publicação: PL-kNN: A Python-based implementation of a parameterless k-Nearest Neighbors classifier [Formula presented]
Nenhuma Miniatura disponível
Data
2023-03-01
Orientador
Coorientador
Pós-graduação
Curso de graduação
Título da Revista
ISSN da Revista
Título de Volume
Editor
Tipo
Artigo
Direito de acesso
Resumo
This paper presents an open-source implementation of PL-kNN, a parameterless version of the k-Nearest Neighbors algorithm. The proposed model, developed in Python 3.6, was designed to avoid the choice of the k parameter required by the standard k-Nearest Neighbors technique. Essentially, the model computes the number of nearest neighbors of a target sample using the data distribution of the training set. The source code provides functions resembling the Scikit-learn methods for fitting the model and predicting the classes of the new samples. The source code is available in the GitHub repository with instructions for installation and examples for usage.
Descrição
Palavras-chave
Idioma
Inglês
Como citar
Software Impacts, v. 15.
