KNN Model

Example Usage

1from mltrain.supervised.KNN import KNN
2
3# Initialize the model
4model = KNN(k=3,distance='euclidean',purpose='classification')
5
6# Train the model
7model.train(X_train, y_train)
8
9# Make predictions
10predictions = model.predict(X_test)
11

Overview

The KNN class implements the k-Nearest Neighbors algorithm for both classification and regression tasks. It supports two distance metrics: Euclidean and Manhattan. The class provides methods to train the model, make predictions, and evaluate performance using accuracy or mean squared error.

Hyperparameters

• k (default=3): The number of nearest neighbors to consider.
• distance_metric (default='euclidean'): The distance metric to use. Options are 'euclidean' and 'manhattan'.
• purpose (default='classification'): The purpose of the model. Options are 'classification' and 'regression'.

Methods

__init__(self, k=3, distance_metric='euclidean', purpose='classification')

Initializes the KNN model with specified hyperparameters.

predict_single(self, x)

Predicts the label or value for a single input sample.

predict(self, X)

Predicts labels or values for multiple input samples.

mean_squared_error(self, y_true, y_pred)

Calculates the Mean Squared Error (MSE) between true and predicted values.

accuracy(self, y_true, y_pred)

Calculates the accuracy of predictions for classification tasks.

confusion_matrix(self, y_true, y_pred)

Computes the confusion matrix for classification tasks.

Additional Notes

- The predict_single method calculates distances between the input sample and all training samples, then selects the k nearest neighbors to make a prediction.
- For classification, the most common label among the k nearest neighbors is returned. For regression, the mean value of the k nearest neighbors is returned.
- The mean_squared_error and accuracy methods are only applicable for regression and classification tasks, respectively.