AI Engineering Degree Practice Exam 2025 - Free AI Engineering Practice Questions and Study Guide

The kNN algorithm, or k-Nearest Neighbors, is a versatile non-parametric method primarily used for both classification and regression tasks. When it comes to continuous targets, kNN is particularly relevant in regression contexts, where it estimates the value of a target variable based on the average (or weighted average) of the values of the k nearest neighbors.

In simple terms, when predicting a continuous outcome, kNN looks at the 'k' closest instances in the feature space, assesses their associated output values, and returns an estimated value based on those neighbors. This ability to compute averages makes kNN effective for regression problems, thereby reinforcing that it can indeed estimate values for a continuous target.

The other options are limited in scope: the algorithm is not exclusively for classification purposes, as it is equally adept at regression; it is capable of estimating values for continuous outcomes; and it does not require a linear relationship between features, making it flexible and applicable in a variety of scenarios where relationships can be non-linear. Thus, the correct answer emphasizes kNN's dual functionality in dealing with continuous targets, confirming that it can serve effectively in regression problems.