Supervised Learning - Classification Week 8 Challenge

Ans 1)

Kindly refer attachments.

Ans 2)

Pros of k-Nearest Neighbors (KNN): K-Nearest Neighbors (KNN) boasts simplicity and intuitiveness, making it an accessible algorithm for various applications. Its non-parametric nature allows it to adapt flexibly to different data distributions, while its lazy learning approach enables effective handling of dynamic datasets. KNN excels in low-dimensional spaces, and its lack of a training phase minimizes computational costs during training. Moreover, it is particularly useful when dealing with straightforward classification or regression tasks, where the decision boundary is not complex.

Cons of k-Nearest Neighbors (KNN): Despite its strengths, KNN has limitations that should be considered. One significant drawback is its computational intensity during the prediction phase, especially with growing datasets, as it requires calculating distances for each query point. KNN is sensitive to irrelevant or redundant features, necessitating careful preprocessing. The algorithm can be influenced by outliers, impacting distance calculations and majority voting. Memory usage can become a concern with large datasets, and the optimal choice of the 'k' parameter is crucial for balancing overfitting and underfitting. In high-dimensional spaces, KNN may face challenges due to the "curse of dimensionality," making it less effective in such scenarios. Understanding these limitations is essential for making informed decisions when applying KNN in various machine learning contexts.