- Its a very simple implementation of K-Nearest Neighbor algorithm for Supervide Learning (user labeled data)
- Version 1.0
- The following similarities metrics are presents:
- Euclidian Distance
- Jaccard Distance
- Pearson Correlation
- Cosine Distance
- Minkowski Distance (to be done)
- Manhattan Distance (to be done)
- Mahalanobis Distance (to be done)
- A naive knn implementation with (or without) k-fold cross-validation.
- Just clone the project
- Setup your project
- Init a
SimpleKNNClassifier - Inject the
SimilarityCalculatorof your choice intoSimpleKNNClassifierinstance - Fit with some labeled data
- Train the classifier (you can choose if you want to train using k-fold cross validation or not)
- Pass some data to
predict()method and see the label predicted