Date of Submission

Spring 2024

Academic Program

Computer Science

Project Advisor 1

Robert McGrail

Abstract/Artist's Statement

With the rapid invention of technology and the increasing accessibility of technology, multi-dimensional database has grown rapidly throughout the world. Due to the impact of the multi-dimensional databases, the efficiency of the nearest neighbor algorithms holds critical importance for diverse applications ranging from information retrieval to machine learning. In this research paper, we will conduct an experiment in order to evaluate the contemporary nearest neighbor search algorithms juxtaposed against the tau- monotonic search algorithm - a recently proposed technique claiming better performance based on the Monotonic Relative Neighborhood Graph (MRNG) in Cong Fu's research. We developed a diligent experimental framework including standardized multidimensional datasets and evaluation metrics that captures accuracy, efficiency, and scalability attributes. Leveraging this methodology, we benchmark the legacy algorithms viz. k-d trees, ball-tree, locality sensitive hashing, etc. vis-à-vis tau-monotonic search across pertinent performance axes. Through reproducible experiments and statistical tests, the findings reveal distinct insights into the comparative strengths and shortcomings of these algorithms across pertinent search quality and speed trade-offs. the study intends to provide an impartial and insightful perspective on how to optimize nearest neighbor search performance in the context of multi-dimensional data. Beyond substantiating the functional advantage fortau-monotonic search, we highlight open challenges to inform future advancements in this integral subfield within database systems research.

Open Access Agreement

On-Campus only

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Download

This work is protected by a Creative Commons license. Any use not permitted under that license is prohibited.

Bard Off-campus Download

Bard College faculty, staff, and students can login from off-campus by clicking on the Off-campus Download button and entering their Bard username and password.

Share

COinS