> For the complete documentation index, see [llms.txt](https://www.mlcompendium.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://www.mlcompendium.com/predictive-ml/nearest-neighbors.md). # Nearest Neighbors Nearest neighbours classify a point by the points closest to it, which is simple until the data is large and high-dimensional. This page answers how to do that search at scale: first a library for approximate search, then the question of the optimal K, then the benchmarks that compare libraries up to billion-scale search. ### KNN The method needs a distance, so it depends on how features are scaled, and it is a common interview topic. The same notes are in [Interview questions](/ai-product/data-science-management.md#interview-questions) and [Normalization & Scaling](/data/processing-intro/normalization-and-scaling.md). The first problem is speed. [Nearpy](https://github.com/pixelogik/NearPy) is a Python framework on github for fast (approximated) nearest neighbour search in large, high-dimensional data sets using different locality-sensitive hashes: knn in scale! The second problem is finding the optimal K, how many neighbours should vote. Once several libraries can do the search, they have to be compared: the [Benchmark of nearest neighbours libraries](https://github.com/erikbern/ann-benchmarks/) is erikbern's ann-benchmarks, benchmarks of approximate nearest neighbor libraries in Python. At the far end of scale, [billion scale aprox nearest neighbour search](https://big-ann-benchmarks.com/) is the NeurIPS'23 Competition Track: Big-ANN. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://www.mlcompendium.com/predictive-ml/nearest-neighbors.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.