Date of Award
Fall 2024
Degree Type
Open Access Dissertation
Degree Name
Mathematics, PhD
Program
Institute of Mathematical Sciences
Advisor/Supervisor/Committee Chair
Marina Chugunova
Dissertation or Thesis Committee Member
Ali Nadim
Dissertation or Thesis Committee Member
Qidi Peng
Terms of Use & License Information
This work is licensed under a Creative Commons Attribution 4.0 License.
Rights Information
© 2024 An Ly
Keywords
Max kCut Optimization, Text Classification
Subject Categories
Mathematics
Abstract
I introduce a novel recursive modification to the classical Goemans-Williamson MaxCut algorithm, offering improved performance in vectorized data clustering tasks. Focusing on the clustering of medical publications, I suggest to employ recursive iterations in conjunction with a dimension relaxation method to enhance density of clustering results. Furthermore, I propose a new vectorization technique for articles, leveraging conditional probabilities for more effective clustering. I believe that these methods will provide advantages in both computational efficiency and clustering accuracy. I will analyze the effectiveness of recursive iterations and higher-dimensional generalizations of the GWA in the hopes of achieving more accurate dissimilarity-based clustering. I think these methods combined with dimensionality reduction have the potential to further enhance clustering results. In addition, the introduction of the vectorization method based on conditional probabilities will provide an additional tool for unsupervised document classification. While GWA shows promise in accurately clustering articles, there are some challenges that will need to be researched and refined on other collected or computer-generated datasets before being applied. Future development of techniques to handle outliers and to fine-tune the parameters will contribute to a more precise and robust method.
ISBN
9798346863373
Recommended Citation
Ly, An. (2024). Applications and Analysis of NLP Deep Learning Models for Antibiotic's Side Effects Classifications. CGU Theses & Dissertations, 900. https://scholarship.claremont.edu/cgu_etd/900.
