Classic Top-$k$ Nearest Neighbor (NN) Querying Algorithms in University Libraries: A Literature Review

Top-$k$ nearest neighbor (NN) querying is a significant scientific problem both in theory and in real life. When researchers want to find related knowledge points in physical and digital libraries or readers want to find the nearest books in a library, the existence of top-$k$ NN querying methods can be imperative. Apart from querying methods based on artificial intelligence, the classic querying algorithms distinguish themselves from a variety of methods by their high efficiency, low consumption, and satisfactory explainability. In this review, we survey the top-$k$ NN querying algorithms to comprehensively structure the top-$k$ NN querying methodologies and techniques in large and small digital libraries based on classic querying algorithms. Our review offers a structured development of top-$k$ NN querying by categorizing algorithms based on mined objects and target quantities. We introduce several approaches, including the classical grid method, the tree method, and the sequential accessing method across various specific top-$k$ NN querying tasks. Their advantages and limitations are compared and discussed. Additionally, potential future research directions are also proposed in this review.